September 17, 2025 • 43 mins
Bloomberg’s Caroline Hyde and Ed Ludlow discuss reports that China's internet watchdog has told Alibaba, ByteDance and others to terminate orders for an Nvidia AI chip. Plus, Microsoft, OpenAI and other US tech giants announce plans to spend tens of billions on AI infrastructure in the UK, as President Trump conducts a state visit. And StubHub CEO Eric Baker joins the show as shares of the ticket reseller begin trading publicly.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live
from coast to coast with Caroline Hyde in New York
and v Lovelow in Sent Francisco.
Speaker 2 (00:23):
This is Bloomberg Tech. Coming up.
Speaker 3 (00:25):
China's Internet watchdog Taels, Ali Barber, Byte Dance and others
to terminate orders for an Nvidia AI chip, plus.
Speaker 4 (00:32):
Microsoft OpenAI among the tech giants announcing plans to spend
tens of billions on AI infrastructure in the UK and
in President Trump's visit this week.
Speaker 3 (00:41):
And ticket reseller stub hub begins trading publicly today, will
be joined by its CEO, Eric Baker, live from the NYC.
Speaker 4 (00:49):
Meanwhile, let's get to the rest of the public markets. Ed,
we're in wait and see mode, maybe a bit of
profit taking after what's been a significant run up in
the nasback of late. We're down about five ten percent
on the NASBAK one hun anticipate a fed rate cut
that is being priced in by the market, But what
does it signal for the rest of the year. That's
a key question. On the macroside, what's you've.
Speaker 3 (01:07):
Got on the micro Well, a big part of the
drag at the index level is in video. We are
down almost three percent on a four trillion dollar market
cap company, down for a third straight session, two hundred
billion dollars in market cap loss. The headline's really simple,
the ft, reporting that China's Internet regulator is saying to
big Chinese tech companies cancel orders for the latest depowered.
Speaker 2 (01:29):
Variant of Nvidia's AI chip for that market.
Speaker 3 (01:31):
At the same time, Jensen Wong's in London with the
President of the United States in the UK on that
state visit. There's a lot of focus on this name
right now.
Speaker 5 (01:40):
There certainly is. Let's get to it over in Washington.
Speaker 4 (01:42):
Our tech editor is with us Mike Shephard.
Speaker 5 (01:44):
Look, this is going.
Speaker 4 (01:45):
To be a key conversation that Jensen Wang is likely
to be having with the President of the United States
because once again another chip designed for China been pushed
back by the country itself.
Speaker 6 (01:57):
Well, that's Ray Carol and Jensen Wang has been having
those kinds of conversations with President Donald Trump since he
took office in January. Jensen has been trying to reshape
the picture when it comes to export controls in a
way that's a little bit more advantageous for Nvidia. Over
the past several years, we've seen the US Titan, China's
(02:18):
access to Nvidia's most advanced AI chips, and they just
wont back permission to sell the Age twenty in China,
although now we are seeing China balk at that this
latest turn of events so signals and escalating pressure campaign
by China, and they feel like they have some leverage
there because the local companies in China making semiconductors can
(02:39):
now compete with some of those lower grade, lower performance
AI chips from Nvidia, and now Nvidia would like to
leapfrog that and sell something perhaps a little more advanced,
but they would need President Donald Trump to get a
board with that. So we'll have to be watching not
only for whatever conversation takes place between Jensen Wong andsent
(03:00):
than Trump, but the more important conversation on Friday care
between Trump and his Chinese counterpart Shijianping to resolve a
host of issues. We know about TikTok, but now in
VideA is really emerging on the screen as a topic
that they will discuss. Chines has expressed a lot of
chagrin over export controls and this would be a moment
to settle it all, including for video.
Speaker 3 (03:22):
And to be clear, the chipping question here is RTX
pro six thousand and d, not h twenty, which is
largely Blackwell architecture related and doesn't use HBM, and we're
trying to get more on that story. Bloombergs Mike Shepherd,
thank you very much. Microsoft, Open Ai, and other American
companies announced plans to spend tens of billions of dollars
on tech infrastructure in the UK, from AI to quantum computing.
(03:45):
Other announcements include investments from companies like in Nvidia, Core Weave, Salesforce,
and Google. This comes as President Trump is in the
UK this week. Bloomberg Surveillance is a Marie horder and
joins US from Windsor. The headlines are really clear capital
commitments for the biggest technology names, but they are like
attending this state visit with Trump almost in parallel.
Speaker 2 (04:06):
AMH.
Speaker 3 (04:07):
What do we need to know what happens in the
next couple of days.
Speaker 7 (04:13):
Well, it's tens of billions of dollars, as you've said,
when it comes to a massive dump of capital from
these UK US tech firms into the United Kingdom, massive
win for the Prime Minister Keir Starmer, especially at a
time when he's dealing with a lot of political fallout
here in the United Kingdom, really drowning in the polls
and has been on thin ice.
Speaker 5 (04:35):
We've seen a.
Speaker 7 (04:36):
Number of individuals resigned or ousted from his government, most
notably Lord Mandelssohn after revelations of his close relationship with
Jeffrey Epstein in Bloomberg News reporting.
Speaker 5 (04:47):
So this idea that UK tech firms.
Speaker 7 (04:49):
Want to make sure that they are in the United
Kingdom and building the United Kingdom is definitely a win
for the Prime Minister. You have to think though, that
given all of this money that is coming from the
United States regarding artificial intelligence, regarding clown computing and quantum
computing into the United Kingdom, that might give the President
(05:10):
some wiggle room and leverage to feel emboldened to bring
up once again at the negotiating table the Digital Services tax,
which more than twenty Republicans in Washington are really urging
the White House to do on this trip.
Speaker 5 (05:24):
So there's a lot of tech ticals to.
Speaker 7 (05:26):
Be had, and this evening, the President at a banquet
is going to be flanked by Jensen Wang, who was
of course with him as well on the Middle East trip,
Sam Altman of Altman, Ai Sati Nadella of Microsoft. And
also I think the name to watch, of course is
Apple's Tim Cook, because notably, that was the individual that
(05:46):
was missing in the Middle East that the President continuously
wanted to remind everyone he was absent. He's joining him
on this state visit to the United Kingdom.
Speaker 5 (05:55):
Blomberg Zamorie Horden.
Speaker 4 (05:57):
We thank you so much from Windsor Castle. Talk more
about TikTok now, because it's US operations would be acquired
by consulting that includes Oracle Andrewson Horowitz and a private
equity firm Silver Lake, under a deal President Trump is
set to discuss with Chinese President Hugime Ping this week.
This framework agreement would create a US based version of
the popular social media app that's bringing in Greenberg. Social
(06:18):
media reporter Alexandra Levine. I mean, this whole segment is
about how technology is at the heart of geopolitics.
Speaker 5 (06:25):
TikTok has been throughout many a year.
Speaker 4 (06:28):
Does it feel like we're getting some semblance of that
framework that still might not actually occur.
Speaker 8 (06:33):
Well, Look, the issue has never been can the US
come up with buyers that China likes the issue has
really been the can China part with the algorithm, which
is the secret sauce of TikTok, which is what makes
it's so popular with half of the country. And so
even though we now know that Oracle and Silver Lake
Injury Center are sort of potentially at the finish line
(06:53):
with this, there's very little detail about how Bydowance will
be involved in this potential US app that US users
might be asked to migrate to, and so without knowing
the details around the algorithm and also by Dance's involvement,
we really don't know if what we will end up
seeing with this framework will even be compliant with the
law exactly.
Speaker 3 (07:11):
So everything we're discussing is according to Bloomberg reporting and
Bloomberg sources, and when we did speak to a White
House official, they basically said that any details of this
framework a speculation until such time that the White House
confirms them. However, based on the reporting, Alex, something Caroline
and I were discussing this morning is that the consumer,
the TikTok user, will face a choice go to a
(07:33):
new app, stay or not, you know, and just explain
the kind of no man's land and confusion there is
out there right now for TikTok as a platform.
Speaker 8 (07:43):
Well, look, I think that TikTok users have proven that
they are an extremely loyal base. We saw that earlier
this year when TikTok was on the verge or last
year when TikTok was on the verge of being banned
and TikTokers called their members of Congress and deluged them
with messages saying keep the app, don't be on the app.
But I think one thing that'll be really important to
watch is just the limbo that the people working at
(08:06):
the company will be in. And also, a lot can
change between now and Friday. As we know, there's a
lot going on within video right now, there's sort of
other players that are percolating in the background that could
completely change things between now and Friday, And then a
lot could change between Friday and the middle of December,
when this next phase of the extension will expire.
Speaker 3 (08:24):
Bloomberg's Alexandra Levine, thank you very much. Now coming up
on Bloomberg Tech, shares of the ticket reseller stub Hub
begin trading today. Stick with us for a front row
seat to a conversation with CEO Eric Baker.
Speaker 2 (08:37):
No ticket required. As snacks. This is Bloomberg Tech.
Speaker 4 (09:34):
Shares of stub hub will start trading today with the
shares haven't been priced at twenty three dollars fifty. That
values a company in around eight point six million dollars.
The ticket reseller is the latest company to go public
in a USIPO.
Speaker 5 (09:45):
Market experience a bit of a hot streak.
Speaker 4 (09:47):
Let's speak with StubHub CEO Eric Baker, who joins us
now from the NYSC Eric, the history. You set up
this company in the year two thousand, sold it to eBay,
bought it back, amalgamated with a European giant. You had
already built a Now it goes public. How does it feel.
Speaker 9 (10:03):
It's tremendous, A lot of gratitude to reach the milestone
for the company. It's certainly been a long journey, as
you say, from two thousand, but now we're really focused
on the next twenty five years and the bright future ahead.
Speaker 4 (10:14):
Take us to the next twenty five years. Because there
are regulatory headwinds, there are people worrying about bots.
Speaker 5 (10:19):
There are a way in which you.
Speaker 4 (10:21):
Can beef up this business to add more things in
secondary market ticket sales.
Speaker 5 (10:25):
Where do you go? Where do you focus.
Speaker 10 (10:28):
Yeah, So our.
Speaker 9 (10:28):
Focus really is we've always wanted to make it easier
for people to access live events, and so what we
started with is, as you noted many years ago, was
to make the resale market safe and secure and easier
for people to access tickets. What we're focused on now
is we realize consumers don't care whether it's a primary
ticket or a secondary ticket. They just want to access
the live events that they want to go to and
(10:49):
discover easily, safely, and so forth. And that system has
been pretty broken for a long time how those tickets
are originally distributed. So we're really working on creating a
better experience. We've already had exciting partners like the Yankees
and Major League Baseball. We started selling tickets directly in
this open distribution network.
Speaker 3 (11:08):
You look at going public in a time where Taylor
Swift had gone through the Ears tour, there was a
lot of hype and mentum, and you hit pause. You're
doing it now, And I think a big question everyone's
got for you, Eric, whether they're a consumer or not,
is what happens next with a regulatory crackdown, the practices
of your industry are very much under the microscope at
(11:29):
the moment.
Speaker 9 (11:31):
Well, I think, first of all, we're always excited to
be at the intersection of technology and live events because
we're very bullish on the future live events.
Speaker 10 (11:38):
And we think that doesn't change.
Speaker 9 (11:40):
To your question about regulatory focus, I think two of
the things really people have looked at. One was transparency
and pricing that we at subbub have always advocated for,
and that all in pricing that we lobbied for a
pass at the federal level in May, and that's just
a better experience for consumers, and the key is it's
now a level playing field for everyone in the market.
The second real regulatory focus, quite frankly, has been what
(12:03):
the Justice Department is looking at with the primary issuance
of tickets and making sure that networks are open and
fan friendly, and that is something we certainly applaud because
you know, we're on the side of the fan.
Speaker 3 (12:15):
Eric he is also Fisher Intelligence a headwind or a
tailwind for the secondary ticketing market.
Speaker 9 (12:22):
Yeah, we see it as a tremendous opportunity to tremendous tailwind.
And I think if you think about it again, we're
a technology and product company, and I think Alex Karp
at Palentier.
Speaker 10 (12:32):
I don't want to miss a.
Speaker 9 (12:33):
Tribute and put it really well that he said, the
artificial intalence technology for companies that have tremendous data allows
you to provide a better product while increasing your revenue
tremendously and reducing costs so.
Speaker 10 (12:45):
That everyone wins.
Speaker 9 (12:46):
And we're seeing that across the board, how we deal
with customer service, how we're releasing new features in the
product to make it easier for people to locate the
tickets they want.
Speaker 10 (12:54):
So we're very excited about it.
Speaker 9 (12:56):
And again with the leadership position we have and the
skilled propriety proprietary data we have, we think this guy's
the limit.
Speaker 4 (13:04):
We are speaking with Eric Baker, found a CEO of
stub Hub.
Speaker 5 (13:06):
To our radio and TV audiences.
Speaker 4 (13:08):
You say AI can help drive revenue tremendously. Let's talk
about your revenue thus far, because in the first half
of the year is about three percent growth. That is
significantly slower than the previous Now we know that maybe
we haven't had Taylor Swift on tour this year, but Eric,
are you seeing a slow down in demand?
Speaker 10 (13:26):
Yeah? So, thank you.
Speaker 9 (13:27):
Note we're really seeing people continuing to buy tickets and growth.
I think you note it's important to note two of
the things that happened that are unique this year. One
is Taylor Swift and her tours behind us, and there
is only one Taylor Swift, so.
Speaker 10 (13:40):
I think she's so tremendous.
Speaker 9 (13:41):
We even broke her out in the s one, so
it's great to be that popular. Second is, as I
mentioned all in pricing past in May, and when we
lobbied for it, we knew that creates a one time
hit to the market as consumers just adjust. We've seen
that already in states like New York that had passed
it before. So when you look through those headwinds, the
business is growing extremely nicely, north of twenty percent.
Speaker 4 (14:03):
I'm interested, though, on your profitability or lack thereof, Eric,
because it's interesting you've gone to public without that story
to tell. When will it become the story you can tell?
Speaker 10 (14:15):
Yeah?
Speaker 9 (14:15):
No, I think what you're referring to is by the
accounting metrics like net income that it shows a loss.
But I think when we went on the road show
and when investors look, they really look at EBITDA, which
is your earnings before interest, tax, depreciation amorization, as well
as your cash flow and on that dimension, we've been
profitable for a long time and generating hundreds of millions
of dollars. That's the way we've been able to service
(14:37):
the debt that we've had, and obviously now that we're
deleveraging and taking the debt down, our cash flow will
even go up further. So I think that's one of
the many reasons that our story was very well received when.
Speaker 10 (14:48):
We went out on the road show and buy our
new partners.
Speaker 3 (14:52):
Eric Bloomberg did some deep reporting on the FTC looking
at ticket master arrival on the issue of bots who
ring up sale tickets for resale and they're inaction. Have
you discussed that issue with the regulator and how are
you addressing the problem.
Speaker 9 (15:08):
Yes, we've always and bots, and it's interesting, as you note,
is really about how these tickets are distributed in the
first place, and so one of the things we've always
lobbied for is enforcement of the Bots Act, and that's
really because we believe that these tickets need to be
distributed in a way that's safe, secure, democratic, through open channels.
As you mentioned, they're looking at I believe ticket Master
(15:29):
and others who are handling the primary issues of the
tickets and they really need to make sure that their
technology and safeguards are in place to make sure this
doesn't happen. But again it's something that we're very pro consumer.
That's how we win. So we hope that all that
stuff gets cleaned up, Eric.
Speaker 2 (15:46):
Very quickly.
Speaker 3 (15:47):
Next year's soccer or football World Cup, how big an
opportunity is that for sub Hub.
Speaker 10 (15:53):
Look, we love the World Cup.
Speaker 9 (15:56):
It's been a phenomenal event because think about it, we
run the largest global platform in the world for ticket
exchange for all kinds of events. And if you had
to say what is the largest global event in the world,
I think you can honymously say it's the World Cup.
It's people from all over the world traveling to come
see events.
Speaker 10 (16:12):
It's going to be in North America.
Speaker 9 (16:14):
It's extremely exciting and we're excited to hopefully help as
many fans as possible make their dreams come true to
access getting to the World Cup.
Speaker 10 (16:23):
So we're we're all football fans, so to speak. In
twenty twenty six.
Speaker 2 (16:28):
We're football fans here too.
Speaker 3 (16:29):
Eric Baker, founder and CEO stub Upholding's great to have
you on the program.
Speaker 2 (16:33):
Thank you. Compute provider Gros just closed.
Speaker 3 (16:42):
A new seven hundred and fifty million dollar funding round,
valuing the company as six point nine billion dollars, highlighting
investor interesting companies seeking to alleviate a shortage of chips
and computing.
Speaker 2 (16:52):
Power for AI workloads.
Speaker 3 (16:54):
ROX CEO Jonathan Ross is with us, and I always
like to go back to basics with Jonathan, Like, Rock
stacks up against Nvidia and Google's TPU, and we're largely
focused on inference. Why did you raise this round in
that context? What is it going to allow you to
do to be more competitive against those two giants.
Speaker 11 (17:14):
Well, the bottom line is that the demand for inference
is insatiable. The total amount of capacity that people are
trying to deploy is mind boggling, the numbers that people
are putting up, and that's only growing. And in our case,
we don't feel that we're competitive. We feel that we're differentiated.
If you're not differentiated, there's no point. And in our case,
what we're able to do is we're able to provide
(17:36):
price performance. So when you're training these models, you don't
need the chips to run particularly fast, You just need
them to have a lot of throughput. And in our case,
we're able to provide both speed and throughput, lowering the cost.
And when you're spending money on every single query, the
difference in that cost can be the difference between profitability
and losing money.
Speaker 3 (17:55):
Inference meaning running the models as opposed to training. Have
you done the math on that, Jonathan? How much inference
does the world need.
Speaker 11 (18:04):
It's insatiable for the reason that, unlike with the industrial age,
where as you're producing oil and compute is the new oil.
Let's be clear about that, you can only consume as
much oil as you had hardware built that it could
consume it. In the generative age, every model that you train,
you can scale it up to the amount of compute
that you have available. As such, people are scaling as
(18:27):
much as they can possibly get compute.
Speaker 4 (18:29):
And so that compute needs data center capacity. Where do
you add that? How much is it the US or
internationally focused?
Speaker 2 (18:37):
Jonathan?
Speaker 11 (18:38):
Actually, that's one of our primary advantages. So the GPUs
that are being deployed today often need liquid cooling, and
those are new data centers. In our case, we're deployed
in North America, so the United States, Canada, we've deployed
in Europe, We've deployed in the Middle East. Our first
deployment in Europe was actually an air cool data center
that one of the hyperscalers was moving out of because
(18:59):
they couldn't with GPUs into it. Because we use so
much less energy for our chips, because we produce so
much less heat, we're able to take that data center
on and so that gives us this ability to get
access to data center capacity in a very competitive way.
Speaker 4 (19:12):
So the language processing unit, the LPU that you offer,
I mean you're taking on the TPU that you helped
design over at Google. We understand and indeed, in video
technically speaking, how is it achieving this the run rate
that you can get at a lower temperature point.
Speaker 11 (19:30):
Yeah, So the architecture is very different, and I took
a page out of what I learned at Google. With
respect to distributed systems. Everyone knows that Moore's law means
that the number of transistors doubles roughly every eighteen to
twenty four months. What I think not so many people
have noticed is that the number of chips is more
than doubling every eighteen to twenty four months.
Speaker 10 (19:50):
So we asked.
Speaker 11 (19:50):
Ourselves if we were going to be increasing the number
of chips. Was there an architectural change that we could do,
and the answer was yes, and so, without getting into
the technical detail, what we do is we take the
models AI models and we spread them out across a
large number of chips, and that allows us to operate
much like an automotive factory. And the more chips that
(20:12):
we use, the cheaper it gets, the faster it gets,
and our advantage is how many chips we can bring
to bear on an inference problem. Now, typically when you're
talking about a training cluster, you'll talk about sixteen thirty
two thousand, sixty four thousand GPUs, and the more GPUs
that you can.
Speaker 12 (20:28):
Bring to bear, the better.
Speaker 11 (20:30):
That hasn't been the case in inference. Typically you'll see
eight GPUs, and it's just a technological limitation that prevents
you from bringing more. In our case, we're running some
models for inference with four thousand or more chips, and
the more chips that we can bring to bear, the
faster and the cheaper it gets.
Speaker 3 (20:48):
Jonathan has Grock signed any deals that would see you
benefit from the commitments we've had in the last twenty
four hours in the United Kingdom at the Hyperscala level
and also some of the AI names that have committed
tens of billions of capital to that market.
Speaker 13 (21:03):
Yeah.
Speaker 11 (21:04):
So Gronk has deployed some large clusters around the world,
as an example in the Middle East, and so we
worked with Commerce to be able to get that deployment
done and they were very helpful and we appreciate that.
And also in Finland. Of course, we're looking at the
UK very closely and I'll actually be there in less
(21:25):
than a week.
Speaker 3 (21:27):
The world right now is GPUs and XPUS or in
your case LPUS.
Speaker 2 (21:33):
Why do we need both?
Speaker 11 (21:36):
The architectures are quite different and they have different bottlenecks.
No architecture is perfect, right. This is why we don't
use CPUs anymore.
Speaker 9 (21:43):
For AI.
Speaker 11 (21:44):
We use predominantly GPUs and now LPUS. The difference between
the two is that GPUs are very good at parallel processing,
that is, if there's no sequential component to it. But
when you're talking about language or other reasoning or logic problems,
you can't produce what's called a token. You can't produce
the hundredth token until you've produced the ninety ninth, or
(22:06):
you can't produce the hundredth word until you produce the
ninety ninth and that brings in a sequential component. Now,
CPUs have been good at sequential, but these problems also
have a parallel component at the same time.
Speaker 10 (22:16):
What the LPU.
Speaker 11 (22:17):
Does is it's really good when you have both a
lot of parallel and a lot of sequential components to
the problem, and that allows us to output tokens at
a very fast rate. And the experience difference. It's the
difference between broadband and dial up, but at the much
lower cost than you would expect given the speed.
Speaker 4 (22:38):
Jonathan Ross, CEO and founder of Grock on the latest fundraise,
We thank you for joining us today.
Speaker 3 (22:50):
Welcome back to Bloomberg Tech. I'm going back to semiconductors.
We talked in the show about how Nvidia shares are
under pressure. Well at the index level, they're putting a
drag on the sock. So Philadelphia se Conductor Index. Over
the course of ten sessions, you're up eight point two percent.
But actually, as of last night's close, the socks had
its best streak of gain since twenty seventeen. As it
(23:11):
stands that street's over, we're down four ten toeven percent
in the session. So we've gone from a story where
we had AI fueled growth on chips. Now we're again
concerned about China in particular, and over the next few
days we should hear more because Jensen one, the CEO
of Nvidia, is in the United Kingdom with President Trump
and in video is at the heart of these negotiations
(23:31):
between the US and China. It's an index to watch
going forward, but it has seen a lot of momentum
of late character.
Speaker 4 (23:37):
It has someone who's not in the UK, is Mark Zuckerberg.
Why because Metaconnect is kicking off later today, Analysts they're
anxiously watching for any groundbreaking announcement like well, maybe some
smart glasses project. Brad Erickson, internet analyst at RBC Capital Markets,
joins us. Now smart glasses has been where it's at
from a hardware perspective. For meta how can they push
(23:59):
that story forward with?
Speaker 14 (24:01):
Yeah, I think it's just integrating more of the capability
with Marcu's vision on Meta AI, right, Like they've given
some applications in terms of some of the translation capabilities,
being able to have some of the augmented reality capabilities,
and I think you'll just see, you know, we expect
to see more of that tonight.
Speaker 4 (24:22):
So really the vision of generative AI coming in the
heart and why Meta AI is really amplified with the
hardware story. But the competition's coming. We know Google's got
them in the works. We know that, of course Apple
might or might not in the future. But Brad, therefore,
what sets their deal with ray ban apart? How can
they continue to offer a little bit more to the
experience than an augmented reality vibe.
Speaker 14 (24:43):
Yeah, it's always going to come down to the same
things that arguably made Apples so successful over the last
nearly twenty years, which is tight integration of hardware and software.
Speaker 6 (24:53):
Right.
Speaker 14 (24:54):
Meta has the platform that people already know it for
for the various family of apps they've created AI have
more than a billion users. The utility that that creates
and the stickiness that that creates for their customers, that's
going to be the key thing.
Speaker 12 (25:09):
The hardware everybody's going to be able to.
Speaker 14 (25:11):
Do right, just like they've always been with any piece
of consumer electronics hardware. It's the software that's going to
keep people coming back.
Speaker 3 (25:17):
If they're going to come back, Rad, thank you for
bringing up Apple, and you also reference translation. Last week
at the iPhone event, Apple shares swung from a decline
to positive territory, and Duolingo shares plunged when Apple demonstrated
simultaneous translation through the airport's Pro three a year ago,
Meta demonstrated simultaneous translation through that generation of rayband metas.
(25:43):
Why did Meta not get the same credit for that
technology that Apple did a week ago?
Speaker 14 (25:49):
Yeah, I think part of it is just it was
so new. Secondarily just relative to the stock. You've also
got a giant core business right that you're compying, and
that's clearly not going to be the focus of it,
or is it necessarily clear how you would how you
would yield an.
Speaker 12 (26:05):
Advertising business out of that.
Speaker 14 (26:06):
So I think that's just mechanically that's kind of the
primary reason. But clearly this is going to be, you know,
a capability for a lot of companies here going forward,
and certainly could be a super cool compliment as you know, again,
Meta looks to build in all these capabilities to these
new glasses.
Speaker 3 (26:23):
A lot of people still don't understand the Meta AI story.
With Chat GPT, I just have an app on my
phone or on my MacBook. I use it every single day.
Right With Meta, I have a standalone app. It's integrated
into Instagram, WhatsApp. I am an owner of the RayBan
metas a different interface or medium to use it. But
(26:46):
do you see kind of what the end thing is
going to be for meta how consumers actually interact with
the work they've done on the model side.
Speaker 14 (26:55):
I don't think anybody does candidly. I think one of
the things that I'm always struck by is we all
stay and learn AI every day is that, you know,
week by week, month by month, we kind of all
have to expand our minds on what the capability is.
Vibe coding is a great example something we've been spending
a lot of time in lately with the web design companies.
Speaker 12 (27:14):
People are using these things to.
Speaker 14 (27:16):
Develop applications, not to build businesses necessarily in the traditional
sense that's where everybody's brain would have gone.
Speaker 12 (27:22):
But they're creating daily productivity.
Speaker 14 (27:24):
Tools enterprises, they're using them to create actual internal work
processes right unrelated to their core software. So there's all
sorts of new things. I think METAAI is very similar
in that, you know, we know that it's going to
be able to provide like you said, chat GBT functionality
with the app. What we don't know is how else
(27:44):
might people build that into their day to day lives,
whether it's for.
Speaker 12 (27:48):
Work, personal or productivity or what have you.
Speaker 14 (27:51):
That's going to be value add and get to a
point where you're going to say, hey, I can't live
without this new capability in my life.
Speaker 12 (27:57):
And we don't exactly know what that is yet.
Speaker 14 (27:59):
We're learning about it, we're trying to sort of expand
our minds, but we haven't gotten there yet.
Speaker 4 (28:04):
Well, I mean, Mark Zuckerberg's been expanding the amount that
he wants to spend on it and the appetite and
ambition that's for sure brand and he's been given a pass.
And you think the stock could climb even higher up
to eight p fifty of another to nine percent gain.
And where we currently are, how much does the TikTok
story bear into its bread and butter right now?
Speaker 14 (28:23):
It matters, But TikTok is a very small business compared
to the size of metas overall ad business.
Speaker 12 (28:31):
I will say that, you know, for a quarter or.
Speaker 14 (28:33):
Two now in our channel checks talking with advertisers, what
we have heard was, you know, the stalled or the
uncertain future of TikTok and really caused advertisers to pause
any spending growth. They're still spending to be clear on
ads on TikTok, but there wasn't nearly as much growth
to the degree that we get a deal here at
(28:55):
some point maybe that provides clarity becomes a little bit
of a headwind. But I wouldn't I don't want to
like overstate it because it is TikTok's still you know,
somewhere in the call it ten percent of the business
relative to the size of metas ad business.
Speaker 3 (29:10):
Brad Ericson of RBC Capital Markets, thank you very much.
Lift and Weimo are partnering to bring self driving taxis
to Nashville.
Speaker 2 (29:24):
Customers will be able to.
Speaker 3 (29:25):
Request driver US rides on both companies platforms, and Lift
will provide fleet management for Waymo's national operations. Let's get
to more with Lift CEO David Rischer. This is something
you've been working on for some time. That's right, and
the state of play is very clear. Weimo also has
a partnership with Uber in different markets. You now have
a partnership in Nashville. Just explain the rationale for you
(29:49):
on why this is a great go to market for ROBOTAXI.
Speaker 13 (29:52):
Well, as you know, we've been saying for a while,
we think self driving cars are going to be hugely
market expanding for ride share. It's a really good product.
Speaker 2 (30:00):
And then when we looked at.
Speaker 13 (30:01):
Potential partners of course Weemo has to be at the
top of the list. The great thing about this partnership
is it does two things. It supplies demand, So if
you're in the self driving car business, you got to
have demand, and we got lots of demand every single day,
and we do it in a very integrated way. But
the other thing is, as you just mentioned, it supplies
fleet management. And that sounds kind of geeky, but it's
really important that every one of those cars be ready
(30:22):
to drive twenty four to seven if needed, and that's
something we're really good at. We've been doing that for
a long time.
Speaker 3 (30:27):
So you're emphasizing the sort of capital light business model
for you, right. The counter argument is the Tesla or
the zooks one, which is control the full stack hardware, software, operationally,
the app the consumer in faces with. Would you argue
that you can scale this faster, going this route, go
to more cities quicker?
Speaker 2 (30:47):
I would?
Speaker 3 (30:47):
Yeah.
Speaker 13 (30:48):
I mean this allows every company to specialize, right, so
Wimo can really focus on providing great self driving technology.
The way Moore Driver is obviously a great product. They
partner with OEMs, the car manufacturers that can choose the
best car, and then we can provide the fleet management
as they say, that keeps the cars clean and charged
and ready to go, as well as the demand so
that anytime a person wants a car, they can get
(31:09):
one or they can get a driver car, right, I mean,
they can get both. So it's I think it's a
great division of labor.
Speaker 4 (31:14):
Let's talk about that fleet management flex drive, because you're
going to be building, you're going to be putting some
capal expenditure to work David, Why does that make sense
and when will that be finished?
Speaker 2 (31:24):
So we've been in the flex drive world. Let me
explain for a second.
Speaker 13 (31:27):
So about a decade ago, Lift decided that for a
certain type of driver, it really makes sense to have
cars that they can rent right on the platform and
they can do we can take care of the maintenance
and making sure that those cars are always ready to drive.
So that's what we've been doing for about a decade now.
So when we look at the autonomous vehicle world, you know, look,
(31:49):
when you see a car driving around the city, it
looks like magic. It looks like no one had to
touch it, but no someone had to charge it, someone
had to clean it, someone has to maintain it. Some
probably has to reboot it from time to time, and
we're really good at that. So, as you say, we're
building a new depot there, we've got a small one.
We're building a bigger one for this partnership. And what
it does is it allows the cars to be ready
to drive.
Speaker 2 (32:07):
Ready to drive.
Speaker 13 (32:08):
Think about so when you travel on an airline and
you know that of course they want the airplane in
the air, but they also needed to be ready to
fly at all times. They want that turnaround to be
super fast. And that's what flex Drive does. It means
that people like Waymo can get the most out of
our asset, which is super good for them.
Speaker 4 (32:23):
Look today, investors love this announcement and the shares a rising.
Shares have been doing well all year because you've had
three straight quarters of profitability. David, how are you going
to finance what is a little less asset like is
building these facilities to help keep all our auto cars clean.
Speaker 2 (32:39):
Yeah.
Speaker 13 (32:40):
So, as I say, because we've been in the flex
drive business, the subsidiary business for a long time, we've
actually had capital deployed to that for a long time.
We have about twenty four locations in about fifteen different
states across the United States, which I think gives you
a sense of how this kind of partnership could scale
over time. And we own some of those cars ourselves,
others of those cars or financed by other groups. So
(33:01):
the good news is this won't be a significant change
to the current strategy. It's just sort of a revectoring,
let's say, of the money we spend more towards.
Speaker 2 (33:08):
Autonomous okay, scale over time.
Speaker 3 (33:10):
This morning, Bloomberg Tech producer Marguerite Gallery, and he put
this table together. You know the table, right, It explains
the roadmap from here, So let's look at it together.
This is twenty twenty six and it focuses on WEIMO principally.
But there's you there, right Nashville. You are going to
come up against suber in negotiating with WEIMO for cities
that are not yet claimed for what are the advantages
(33:32):
that you're going to present to win that business with
WEIMO against Duba And how much you're gonna have to
give up, you know, to get a deal done.
Speaker 13 (33:41):
Yeah, So I think the way I would look at
That is one of the reasons we were so patient
in working through this deal because it took a while.
It was we wanted to come up with something that
is good for Weimo, good for Lyft, and good for riders,
and that's hard to do. It's hard to have, you know,
kind of a three away win. We feel really good
about it and feel like this is a great model
for us to expand to other cities in the future.
(34:02):
But look, city by city, you know some places we're
actually going to compete with Weimo itself, right because for
example here in San Francisco, they've got their own appreciate
that admission, Yeah, of course, So this is something that
we have to kind of recognize and say, how do
we come up with a construct that is really durable
and frankly can benefit all the companies and then the
other guys will do what they do and you know,
we'll see who does a better job.
Speaker 3 (34:21):
Look, people are going to accuse me of wanting to
create a headline here, but it's a really important question.
Tesla and Zeke's are committed to their own go to
market model. Right, we talked about they control the full stack,
But have you held discussions with both Tesla and Zeke's
to say like, look, this is the model we're doing
with Weimo, we can do something similar.
Speaker 2 (34:41):
Do you see that as being realistic?
Speaker 13 (34:43):
So of course I don't want to come in on,
you know, conversations we might be having. What I can
tell you is we're talking to everyone who finds the
partnership model interesting. You know, there are, of course always companies.
Whenever you see a new industry, you have some companies
that say I.
Speaker 2 (34:58):
Want to do it all.
Speaker 13 (34:58):
And frankly, if you look back a couple years ago,
just about every company in the self driving space was
in that zone. They wanted to do everything, you know,
top to bottom. I think over time most companies find,
particularly when it's capital intensive and quite complex, like ride
share is and self driving is that a division of
labor where everyone does what they're good at, is the
is the best model. But we'll see, you know, it's
still such early days.
Speaker 4 (35:20):
But you struck deals with interesting players in Europe. I do,
what's it like working with a Chinese giant? What were
those negotiations like? You know that?
Speaker 13 (35:28):
Actually, I'll tell you what was very interesting. It came
together fairly quickly and in part I think is you know, frankly,
Chinese companies are kind of built for speed. You know,
they for those of you who don't know, you know,
by do you can sort of think of them as
kind of the Google or Alphabet of China. They have
many businesses, but self driving cars are a very important
one for them, and we committed that we would roll
out some of their technology in one to two markets.
(35:50):
We would expect in Europe over the next twelve months
or so.
Speaker 10 (35:53):
It's still very.
Speaker 13 (35:54):
Early days there, but I can tell you they've got
really high standards for their technology. We're super excited about
working with those guys in the wall to have more
to say about that next year once we start to
roll out in Europe.
Speaker 5 (36:03):
One minute, how hard is it to get to New York?
Speaker 13 (36:09):
I mean I can take a flight there pretty easily.
Speaker 4 (36:11):
But also the ROBOTAXI how hot? When am I going
to have it here?
Speaker 10 (36:16):
I love that?
Speaker 13 (36:16):
You know, it's going to take a little while. New
York is a complex place to do business. The streets
are complicated, you know, there's a lot of interest there. Look,
you know this Caroline, I mean lift has been there
for many years. It's super strong in New York. We
also run the city bike program. As you know, super
excited bring self driving there someday, but that was probably
gonna take a little while.
Speaker 2 (36:32):
Sorry about that.
Speaker 12 (36:32):
You can have to work for all.
Speaker 4 (36:33):
My addiction with the bikes will just have to continue.
Speaker 5 (36:35):
Lift to see David so as a joy to have
you on the show. Take Care.
Speaker 3 (36:46):
Mentioned earlier in the show that Microsoft will invest thirty
billion dollars in the UK. It's part of a push
to expand AI infrastructure. We also heard from GROC about
its data center plans. Of course, tech companies are pledging
to build more data center steer around the world. Let's
bringing Maria Davidson, CEO of Kojo. It's the largest construction
procurement platform in the US, giving her a unique insight
(37:09):
into the data center build out. I think we can
get to where you fit into the process of what
is just massive infrastructure as a project, but that your
reaction first to what's happening in the UK. It seems
to be the next market in this AI cycle.
Speaker 15 (37:25):
Oh absolutely. I mean when you think about how much
potential there is in the UK, I think it is
extremely exciting and very very untapped. So I love to
see companies like Microsoft and Google coming in and investing
in the UK, and the US and the UK actually
working closer together.
Speaker 2 (37:40):
You say a lot of potential in the UK.
Speaker 15 (37:42):
Why, I think because one, there's extremely talented workforce in
the UK, and you know, you have world class universities.
You have incredible research like deep Mind coming out years
before anyone would have expected an outcome like that in
the UK, and so I think that means that in
the UK right now, if we keep on investing, I
(38:04):
think we're going to see a lot more deep Mind
like outcomes.
Speaker 4 (38:07):
I've got to say, You've got three Brits around a
table talking about the opportunity in the UK. So I'll
diverse us a little bit because we're three Brits that
sit in the US right now. Maria, you're building businesses
to try and make supply chains clearer for contractors for
builders out there, and I'm interested as to what the
scale is like here in the United States and what
the key bottlenecks are that you're trying to fix.
Speaker 15 (38:26):
I mean, the scale is enormous. So just this year,
we have forty billion dollars going into the construction of
data centers in the US, that's expected to grow more
than ten percent a year, so by twenty thirty that's
going to be sixty five billion dollars. That's unprecedented. And
the biggest issue with building these it's not getting GPUs,
it's getting the materials and finding the electricians to be
(38:48):
able to actually meet this demand.
Speaker 4 (38:50):
Okay, so your role is ensuring that when the contractor,
which is very hard to come by and labor is short,
and we've seen immigration policies hit that, but when you
get the on site, they've actually got the pieces in
the parts to work with what has been some of
the most complex areas when it comes to particularly the
hyperscaler's need and the data center build out.
Speaker 15 (39:09):
It's a lot when you think about the construction industry
in the US. Even though they're building some of the
most sophisticated infrastructure out there right they're building projects that
involve huge amounts of technology and AI infrastructure specific needs,
they still largely have to rely on pen and paper
and Excel to track where things are. It's hard to
(39:30):
know what's arriving when these are all huge challenges. And
so when you couple that with the fact that there
are five hundred thousand workers missing from the US construction workforce.
And you couple that with the fact that materials on
average of forty four percent more expensive than they were
in February twenty twenty before COVID, that starts really creating
(39:50):
a perfect storm.
Speaker 3 (39:52):
I actually think it's worth going a step earlier in
the process than what Caroline just asked about. Right, We've
kind of learned over the course of a lot two
or three years about how it works in the United States.
You have a financing partner, you have someone like Cruso
putting it all together.
Speaker 2 (40:06):
You need to select a site.
Speaker 3 (40:08):
Then you've got to talk to the utility because you
need to power that site. The utility has to think
about where they get the electricity from. Is that going
smoothly and well that that initial process it depends.
Speaker 15 (40:20):
It's definitely gotten a lot faster. If you think about
your average size data centers right as data centers on
the smaller side maybe ten to thirty megawatts. Those previously
would have taken that's say, twenty four months to thirty
six months to build. Now you're seeing data centers build
probably twice as fast, in part because things like site readiness,
inter connection, they're all being dealt with in advance.
Speaker 2 (40:42):
Now.
Speaker 15 (40:42):
I think prefab is a huge subject here because the
more that we can lean into modular construction, the more
we don't have to rely on the sequential building, but
we can actually build in parallel.
Speaker 3 (40:52):
There's also like a self fulfilling prophecy in the sense
that all of the AI names who actually just end
up leasing the capacity, right, they don't even own the thing,
but they would say, Oh, all the work we're doing
in AI is actually going to benefit the planning and
execution of these buildouts. I guess that's what you'd say.
You also help.
Speaker 15 (41:09):
With Yeah, absolutely, I mean a lot of it is.
The better planned that you can be in advance, the
better the project is actually going to go. And so
if you know all of the components that you're going
to need, if you have the mapped out schema, if
you have all of those interconnection blocks ready to go,
then of course contractors can be much better prepared.
Speaker 4 (41:29):
I want to talk about you about Kojo, about what
the ninety four million dollars you've thus far raised six
hundred construction companies that you're working with. You've just had
an add on to your CERC I think I believe
how much are.
Speaker 5 (41:42):
You seeing inbound?
Speaker 2 (41:43):
How much are you.
Speaker 4 (41:44):
Fighting off people wanting to put money into your company
because they're realizing it's about bricks and mortar, it's about
the very underlying infrastructure.
Speaker 15 (41:50):
Right now, well, we're seeing a huge amount of inbound.
The focus for us has always been how can we
empower contractors, how can we take the contract is that
for a really long time were completely ignored by the
tech industry, and specifically there I'm talking about the trades, right,
the electricians, the folks working in mechanical and HVAC and drywall,
(42:12):
They for a very long time had very few solutions
that were built specifically for them. And so for us,
it's about saying, well, how do we build solutions that
are specifically targeted towards you and that solve your biggest
pain points like knowing where your materials are and having
the visibility of how you're doing against your estimates in
real time.
Speaker 4 (42:31):
Well, you're just completing a ten million dollar Series C extension.
Speaker 5 (42:35):
Come back when you're raising yet more.
Speaker 4 (42:37):
I'm sure Maria Davidson's CEO, Kojo, it's great to have
you on the Showwow, that does it for this edition
at Bloomberg Tech, we've done our whole host of things zed,
but we're going to count down to Metaconnect tonight. Of course,
tune in tomorrow for an interview with NEA's Chief product officer,
Chris Cox live on the show tomorrow, which ed.
Speaker 5 (42:54):
You'll be doing.
Speaker 2 (42:56):
Yeah, looking forward to that.
Speaker 3 (42:57):
I have a lot of questions about the relationship between
software and hardware.
Speaker 2 (43:00):
After all, that is technology. Check out the pod.
Speaker 3 (43:03):
You know where to find it on the Bloomberg terminal,
as well as online on Apple, Spotify, iHeart, and all
kinds of other places on the internet.
Speaker 2 (43:12):
Thank you for watching. This is Bloomberg Tech.
Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live
from coast to coast with Caroline Hyde in New York
and v Lovelow in Sent Francisco.
Speaker 2 (00:23):
This is Bloomberg Tech. Coming up.
Speaker 3 (00:25):
China's Internet watchdog Taels, Ali Barber, Byte Dance and others
to terminate orders for an Nvidia AI chip, plus.
Speaker 4 (00:32):
Microsoft OpenAI among the tech giants announcing plans to spend
tens of billions on AI infrastructure in the UK and
in President Trump's visit this week.
Speaker 3 (00:41):
And ticket reseller stub hub begins trading publicly today, will
be joined by its CEO, Eric Baker, live from the NYC.
Speaker 4 (00:49):
Meanwhile, let's get to the rest of the public markets. Ed,
we're in wait and see mode, maybe a bit of
profit taking after what's been a significant run up in
the nasback of late. We're down about five ten percent
on the NASBAK one hun anticipate a fed rate cut
that is being priced in by the market, But what
does it signal for the rest of the year. That's
a key question. On the macroside, what's you've.
Speaker 3 (01:07):
Got on the micro Well, a big part of the
drag at the index level is in video. We are
down almost three percent on a four trillion dollar market
cap company, down for a third straight session, two hundred
billion dollars in market cap loss. The headline's really simple,
the ft, reporting that China's Internet regulator is saying to
big Chinese tech companies cancel orders for the latest depowered.
Speaker 2 (01:29):
Variant of Nvidia's AI chip for that market.
Speaker 3 (01:31):
At the same time, Jensen Wong's in London with the
President of the United States in the UK on that
state visit. There's a lot of focus on this name
right now.
Speaker 5 (01:40):
There certainly is. Let's get to it over in Washington.
Speaker 4 (01:42):
Our tech editor is with us Mike Shephard.
Speaker 5 (01:44):
Look, this is going.
Speaker 4 (01:45):
To be a key conversation that Jensen Wang is likely
to be having with the President of the United States
because once again another chip designed for China been pushed
back by the country itself.
Speaker 6 (01:57):
Well, that's Ray Carol and Jensen Wang has been having
those kinds of conversations with President Donald Trump since he
took office in January. Jensen has been trying to reshape
the picture when it comes to export controls in a
way that's a little bit more advantageous for Nvidia. Over
the past several years, we've seen the US Titan, China's
(02:18):
access to Nvidia's most advanced AI chips, and they just
wont back permission to sell the Age twenty in China,
although now we are seeing China balk at that this
latest turn of events so signals and escalating pressure campaign
by China, and they feel like they have some leverage
there because the local companies in China making semiconductors can
(02:39):
now compete with some of those lower grade, lower performance
AI chips from Nvidia, and now Nvidia would like to
leapfrog that and sell something perhaps a little more advanced,
but they would need President Donald Trump to get a
board with that. So we'll have to be watching not
only for whatever conversation takes place between Jensen Wong andsent
(03:00):
than Trump, but the more important conversation on Friday care
between Trump and his Chinese counterpart Shijianping to resolve a
host of issues. We know about TikTok, but now in
VideA is really emerging on the screen as a topic
that they will discuss. Chines has expressed a lot of
chagrin over export controls and this would be a moment
to settle it all, including for video.
Speaker 3 (03:22):
And to be clear, the chipping question here is RTX
pro six thousand and d, not h twenty, which is
largely Blackwell architecture related and doesn't use HBM, and we're
trying to get more on that story. Bloombergs Mike Shepherd,
thank you very much. Microsoft, Open Ai, and other American
companies announced plans to spend tens of billions of dollars
on tech infrastructure in the UK, from AI to quantum computing.
(03:45):
Other announcements include investments from companies like in Nvidia, Core Weave, Salesforce,
and Google. This comes as President Trump is in the
UK this week. Bloomberg Surveillance is a Marie horder and
joins US from Windsor. The headlines are really clear capital
commitments for the biggest technology names, but they are like
attending this state visit with Trump almost in parallel.
Speaker 2 (04:06):
AMH.
Speaker 3 (04:07):
What do we need to know what happens in the
next couple of days.
Speaker 7 (04:13):
Well, it's tens of billions of dollars, as you've said,
when it comes to a massive dump of capital from
these UK US tech firms into the United Kingdom, massive
win for the Prime Minister Keir Starmer, especially at a
time when he's dealing with a lot of political fallout
here in the United Kingdom, really drowning in the polls
and has been on thin ice.
Speaker 5 (04:35):
We've seen a.
Speaker 7 (04:36):
Number of individuals resigned or ousted from his government, most
notably Lord Mandelssohn after revelations of his close relationship with
Jeffrey Epstein in Bloomberg News reporting.
Speaker 5 (04:47):
So this idea that UK tech firms.
Speaker 7 (04:49):
Want to make sure that they are in the United
Kingdom and building the United Kingdom is definitely a win
for the Prime Minister. You have to think though, that
given all of this money that is coming from the
United States regarding artificial intelligence, regarding clown computing and quantum
computing into the United Kingdom, that might give the President
(05:10):
some wiggle room and leverage to feel emboldened to bring
up once again at the negotiating table the Digital Services tax,
which more than twenty Republicans in Washington are really urging
the White House to do on this trip.
Speaker 5 (05:24):
So there's a lot of tech ticals to.
Speaker 7 (05:26):
Be had, and this evening, the President at a banquet
is going to be flanked by Jensen Wang, who was
of course with him as well on the Middle East trip,
Sam Altman of Altman, Ai Sati Nadella of Microsoft. And
also I think the name to watch, of course is
Apple's Tim Cook, because notably, that was the individual that
(05:46):
was missing in the Middle East that the President continuously
wanted to remind everyone he was absent. He's joining him
on this state visit to the United Kingdom.
Speaker 5 (05:55):
Blomberg Zamorie Horden.
Speaker 4 (05:57):
We thank you so much from Windsor Castle. Talk more
about TikTok now, because it's US operations would be acquired
by consulting that includes Oracle Andrewson Horowitz and a private
equity firm Silver Lake, under a deal President Trump is
set to discuss with Chinese President Hugime Ping this week.
This framework agreement would create a US based version of
the popular social media app that's bringing in Greenberg. Social
(06:18):
media reporter Alexandra Levine. I mean, this whole segment is
about how technology is at the heart of geopolitics.
Speaker 5 (06:25):
TikTok has been throughout many a year.
Speaker 4 (06:28):
Does it feel like we're getting some semblance of that
framework that still might not actually occur.
Speaker 8 (06:33):
Well, Look, the issue has never been can the US
come up with buyers that China likes the issue has
really been the can China part with the algorithm, which
is the secret sauce of TikTok, which is what makes
it's so popular with half of the country. And so
even though we now know that Oracle and Silver Lake
Injury Center are sort of potentially at the finish line
(06:53):
with this, there's very little detail about how Bydowance will
be involved in this potential US app that US users
might be asked to migrate to, and so without knowing
the details around the algorithm and also by Dance's involvement,
we really don't know if what we will end up
seeing with this framework will even be compliant with the
law exactly.
Speaker 3 (07:11):
So everything we're discussing is according to Bloomberg reporting and
Bloomberg sources, and when we did speak to a White
House official, they basically said that any details of this
framework a speculation until such time that the White House
confirms them. However, based on the reporting, Alex, something Caroline
and I were discussing this morning is that the consumer,
the TikTok user, will face a choice go to a
(07:33):
new app, stay or not, you know, and just explain
the kind of no man's land and confusion there is
out there right now for TikTok as a platform.
Speaker 8 (07:43):
Well, look, I think that TikTok users have proven that
they are an extremely loyal base. We saw that earlier
this year when TikTok was on the verge or last
year when TikTok was on the verge of being banned
and TikTokers called their members of Congress and deluged them
with messages saying keep the app, don't be on the app.
But I think one thing that'll be really important to
watch is just the limbo that the people working at
(08:06):
the company will be in. And also, a lot can
change between now and Friday. As we know, there's a
lot going on within video right now, there's sort of
other players that are percolating in the background that could
completely change things between now and Friday, And then a
lot could change between Friday and the middle of December,
when this next phase of the extension will expire.
Speaker 3 (08:24):
Bloomberg's Alexandra Levine, thank you very much. Now coming up
on Bloomberg Tech, shares of the ticket reseller stub Hub
begin trading today. Stick with us for a front row
seat to a conversation with CEO Eric Baker.
Speaker 2 (08:37):
No ticket required. As snacks. This is Bloomberg Tech.
Speaker 4 (09:34):
Shares of stub hub will start trading today with the
shares haven't been priced at twenty three dollars fifty. That
values a company in around eight point six million dollars.
The ticket reseller is the latest company to go public
in a USIPO.
Speaker 5 (09:45):
Market experience a bit of a hot streak.
Speaker 4 (09:47):
Let's speak with StubHub CEO Eric Baker, who joins us
now from the NYSC Eric, the history. You set up
this company in the year two thousand, sold it to eBay,
bought it back, amalgamated with a European giant. You had
already built a Now it goes public. How does it feel.
Speaker 9 (10:03):
It's tremendous, A lot of gratitude to reach the milestone
for the company. It's certainly been a long journey, as
you say, from two thousand, but now we're really focused
on the next twenty five years and the bright future ahead.
Speaker 4 (10:14):
Take us to the next twenty five years. Because there
are regulatory headwinds, there are people worrying about bots.
Speaker 5 (10:19):
There are a way in which you.
Speaker 4 (10:21):
Can beef up this business to add more things in
secondary market ticket sales.
Speaker 5 (10:25):
Where do you go? Where do you focus.
Speaker 10 (10:28):
Yeah, So our.
Speaker 9 (10:28):
Focus really is we've always wanted to make it easier
for people to access live events, and so what we
started with is, as you noted many years ago, was
to make the resale market safe and secure and easier
for people to access tickets. What we're focused on now
is we realize consumers don't care whether it's a primary
ticket or a secondary ticket. They just want to access
the live events that they want to go to and
(10:49):
discover easily, safely, and so forth. And that system has
been pretty broken for a long time how those tickets
are originally distributed. So we're really working on creating a
better experience. We've already had exciting partners like the Yankees
and Major League Baseball. We started selling tickets directly in
this open distribution network.
Speaker 3 (11:08):
You look at going public in a time where Taylor
Swift had gone through the Ears tour, there was a
lot of hype and mentum, and you hit pause. You're
doing it now, And I think a big question everyone's
got for you, Eric, whether they're a consumer or not,
is what happens next with a regulatory crackdown, the practices
of your industry are very much under the microscope at
(11:29):
the moment.
Speaker 9 (11:31):
Well, I think, first of all, we're always excited to
be at the intersection of technology and live events because
we're very bullish on the future live events.
Speaker 10 (11:38):
And we think that doesn't change.
Speaker 9 (11:40):
To your question about regulatory focus, I think two of
the things really people have looked at. One was transparency
and pricing that we at subbub have always advocated for,
and that all in pricing that we lobbied for a
pass at the federal level in May, and that's just
a better experience for consumers, and the key is it's
now a level playing field for everyone in the market.
The second real regulatory focus, quite frankly, has been what
(12:03):
the Justice Department is looking at with the primary issuance
of tickets and making sure that networks are open and
fan friendly, and that is something we certainly applaud because
you know, we're on the side of the fan.
Speaker 3 (12:15):
Eric he is also Fisher Intelligence a headwind or a
tailwind for the secondary ticketing market.
Speaker 9 (12:22):
Yeah, we see it as a tremendous opportunity to tremendous tailwind.
And I think if you think about it again, we're
a technology and product company, and I think Alex Karp
at Palentier.
Speaker 10 (12:32):
I don't want to miss a.
Speaker 9 (12:33):
Tribute and put it really well that he said, the
artificial intalence technology for companies that have tremendous data allows
you to provide a better product while increasing your revenue
tremendously and reducing costs so.
Speaker 10 (12:45):
That everyone wins.
Speaker 9 (12:46):
And we're seeing that across the board, how we deal
with customer service, how we're releasing new features in the
product to make it easier for people to locate the
tickets they want.
Speaker 10 (12:54):
So we're very excited about it.
Speaker 9 (12:56):
And again with the leadership position we have and the
skilled propriety proprietary data we have, we think this guy's
the limit.
Speaker 4 (13:04):
We are speaking with Eric Baker, found a CEO of
stub Hub.
Speaker 5 (13:06):
To our radio and TV audiences.
Speaker 4 (13:08):
You say AI can help drive revenue tremendously. Let's talk
about your revenue thus far, because in the first half
of the year is about three percent growth. That is
significantly slower than the previous Now we know that maybe
we haven't had Taylor Swift on tour this year, but Eric,
are you seeing a slow down in demand?
Speaker 10 (13:26):
Yeah? So, thank you.
Speaker 9 (13:27):
Note we're really seeing people continuing to buy tickets and growth.
I think you note it's important to note two of
the things that happened that are unique this year. One
is Taylor Swift and her tours behind us, and there
is only one Taylor Swift, so.
Speaker 10 (13:40):
I think she's so tremendous.
Speaker 9 (13:41):
We even broke her out in the s one, so
it's great to be that popular. Second is, as I
mentioned all in pricing past in May, and when we
lobbied for it, we knew that creates a one time
hit to the market as consumers just adjust. We've seen
that already in states like New York that had passed
it before. So when you look through those headwinds, the
business is growing extremely nicely, north of twenty percent.
Speaker 4 (14:03):
I'm interested, though, on your profitability or lack thereof, Eric,
because it's interesting you've gone to public without that story
to tell. When will it become the story you can tell?
Speaker 10 (14:15):
Yeah?
Speaker 9 (14:15):
No, I think what you're referring to is by the
accounting metrics like net income that it shows a loss.
But I think when we went on the road show
and when investors look, they really look at EBITDA, which
is your earnings before interest, tax, depreciation amorization, as well
as your cash flow and on that dimension, we've been
profitable for a long time and generating hundreds of millions
of dollars. That's the way we've been able to service
(14:37):
the debt that we've had, and obviously now that we're
deleveraging and taking the debt down, our cash flow will
even go up further. So I think that's one of
the many reasons that our story was very well received when.
Speaker 10 (14:48):
We went out on the road show and buy our
new partners.
Speaker 3 (14:52):
Eric Bloomberg did some deep reporting on the FTC looking
at ticket master arrival on the issue of bots who
ring up sale tickets for resale and they're inaction. Have
you discussed that issue with the regulator and how are
you addressing the problem.
Speaker 9 (15:08):
Yes, we've always and bots, and it's interesting, as you note,
is really about how these tickets are distributed in the
first place, and so one of the things we've always
lobbied for is enforcement of the Bots Act, and that's
really because we believe that these tickets need to be
distributed in a way that's safe, secure, democratic, through open channels.
As you mentioned, they're looking at I believe ticket Master
(15:29):
and others who are handling the primary issues of the
tickets and they really need to make sure that their
technology and safeguards are in place to make sure this
doesn't happen. But again it's something that we're very pro consumer.
That's how we win. So we hope that all that
stuff gets cleaned up, Eric.
Speaker 2 (15:46):
Very quickly.
Speaker 3 (15:47):
Next year's soccer or football World Cup, how big an
opportunity is that for sub Hub.
Speaker 10 (15:53):
Look, we love the World Cup.
Speaker 9 (15:56):
It's been a phenomenal event because think about it, we
run the largest global platform in the world for ticket
exchange for all kinds of events. And if you had
to say what is the largest global event in the world,
I think you can honymously say it's the World Cup.
It's people from all over the world traveling to come
see events.
Speaker 10 (16:12):
It's going to be in North America.
Speaker 9 (16:14):
It's extremely exciting and we're excited to hopefully help as
many fans as possible make their dreams come true to
access getting to the World Cup.
Speaker 10 (16:23):
So we're we're all football fans, so to speak. In
twenty twenty six.
Speaker 2 (16:28):
We're football fans here too.
Speaker 3 (16:29):
Eric Baker, founder and CEO stub Upholding's great to have
you on the program.
Speaker 2 (16:33):
Thank you. Compute provider Gros just closed.
Speaker 3 (16:42):
A new seven hundred and fifty million dollar funding round,
valuing the company as six point nine billion dollars, highlighting
investor interesting companies seeking to alleviate a shortage of chips
and computing.
Speaker 2 (16:52):
Power for AI workloads.
Speaker 3 (16:54):
ROX CEO Jonathan Ross is with us, and I always
like to go back to basics with Jonathan, Like, Rock
stacks up against Nvidia and Google's TPU, and we're largely
focused on inference. Why did you raise this round in
that context? What is it going to allow you to
do to be more competitive against those two giants.
Speaker 11 (17:14):
Well, the bottom line is that the demand for inference
is insatiable. The total amount of capacity that people are
trying to deploy is mind boggling, the numbers that people
are putting up, and that's only growing. And in our case,
we don't feel that we're competitive. We feel that we're differentiated.
If you're not differentiated, there's no point. And in our case,
what we're able to do is we're able to provide
(17:36):
price performance. So when you're training these models, you don't
need the chips to run particularly fast, You just need
them to have a lot of throughput. And in our case,
we're able to provide both speed and throughput, lowering the cost.
And when you're spending money on every single query, the
difference in that cost can be the difference between profitability
and losing money.
Speaker 3 (17:55):
Inference meaning running the models as opposed to training. Have
you done the math on that, Jonathan? How much inference
does the world need.
Speaker 11 (18:04):
It's insatiable for the reason that, unlike with the industrial age,
where as you're producing oil and compute is the new oil.
Let's be clear about that, you can only consume as
much oil as you had hardware built that it could
consume it. In the generative age, every model that you train,
you can scale it up to the amount of compute
that you have available. As such, people are scaling as
(18:27):
much as they can possibly get compute.
Speaker 4 (18:29):
And so that compute needs data center capacity. Where do
you add that? How much is it the US or
internationally focused?
Speaker 2 (18:37):
Jonathan?
Speaker 11 (18:38):
Actually, that's one of our primary advantages. So the GPUs
that are being deployed today often need liquid cooling, and
those are new data centers. In our case, we're deployed
in North America, so the United States, Canada, we've deployed
in Europe, We've deployed in the Middle East. Our first
deployment in Europe was actually an air cool data center
that one of the hyperscalers was moving out of because
(18:59):
they couldn't with GPUs into it. Because we use so
much less energy for our chips, because we produce so
much less heat, we're able to take that data center
on and so that gives us this ability to get
access to data center capacity in a very competitive way.
Speaker 4 (19:12):
So the language processing unit, the LPU that you offer,
I mean you're taking on the TPU that you helped
design over at Google. We understand and indeed, in video
technically speaking, how is it achieving this the run rate
that you can get at a lower temperature point.
Speaker 11 (19:30):
Yeah, So the architecture is very different, and I took
a page out of what I learned at Google. With
respect to distributed systems. Everyone knows that Moore's law means
that the number of transistors doubles roughly every eighteen to
twenty four months. What I think not so many people
have noticed is that the number of chips is more
than doubling every eighteen to twenty four months.
Speaker 10 (19:50):
So we asked.
Speaker 11 (19:50):
Ourselves if we were going to be increasing the number
of chips. Was there an architectural change that we could do,
and the answer was yes, and so, without getting into
the technical detail, what we do is we take the
models AI models and we spread them out across a
large number of chips, and that allows us to operate
much like an automotive factory. And the more chips that
(20:12):
we use, the cheaper it gets, the faster it gets,
and our advantage is how many chips we can bring
to bear on an inference problem. Now, typically when you're
talking about a training cluster, you'll talk about sixteen thirty
two thousand, sixty four thousand GPUs, and the more GPUs
that you can.
Speaker 12 (20:28):
Bring to bear, the better.
Speaker 11 (20:30):
That hasn't been the case in inference. Typically you'll see
eight GPUs, and it's just a technological limitation that prevents
you from bringing more. In our case, we're running some
models for inference with four thousand or more chips, and
the more chips that we can bring to bear, the
faster and the cheaper it gets.
Speaker 3 (20:48):
Jonathan has Grock signed any deals that would see you
benefit from the commitments we've had in the last twenty
four hours in the United Kingdom at the Hyperscala level
and also some of the AI names that have committed
tens of billions of capital to that market.
Speaker 13 (21:03):
Yeah.
Speaker 11 (21:04):
So Gronk has deployed some large clusters around the world,
as an example in the Middle East, and so we
worked with Commerce to be able to get that deployment
done and they were very helpful and we appreciate that.
And also in Finland. Of course, we're looking at the
UK very closely and I'll actually be there in less
(21:25):
than a week.
Speaker 3 (21:27):
The world right now is GPUs and XPUS or in
your case LPUS.
Speaker 2 (21:33):
Why do we need both?
Speaker 11 (21:36):
The architectures are quite different and they have different bottlenecks.
No architecture is perfect, right. This is why we don't
use CPUs anymore.
Speaker 9 (21:43):
For AI.
Speaker 11 (21:44):
We use predominantly GPUs and now LPUS. The difference between
the two is that GPUs are very good at parallel processing,
that is, if there's no sequential component to it. But
when you're talking about language or other reasoning or logic problems,
you can't produce what's called a token. You can't produce
the hundredth token until you've produced the ninety ninth, or
(22:06):
you can't produce the hundredth word until you produce the
ninety ninth and that brings in a sequential component. Now,
CPUs have been good at sequential, but these problems also
have a parallel component at the same time.
Speaker 10 (22:16):
What the LPU.
Speaker 11 (22:17):
Does is it's really good when you have both a
lot of parallel and a lot of sequential components to
the problem, and that allows us to output tokens at
a very fast rate. And the experience difference. It's the
difference between broadband and dial up, but at the much
lower cost than you would expect given the speed.
Speaker 4 (22:38):
Jonathan Ross, CEO and founder of Grock on the latest fundraise,
We thank you for joining us today.
Speaker 3 (22:50):
Welcome back to Bloomberg Tech. I'm going back to semiconductors.
We talked in the show about how Nvidia shares are
under pressure. Well at the index level, they're putting a
drag on the sock. So Philadelphia se Conductor Index. Over
the course of ten sessions, you're up eight point two percent.
But actually, as of last night's close, the socks had
its best streak of gain since twenty seventeen. As it
(23:11):
stands that street's over, we're down four ten toeven percent
in the session. So we've gone from a story where
we had AI fueled growth on chips. Now we're again
concerned about China in particular, and over the next few
days we should hear more because Jensen one, the CEO
of Nvidia, is in the United Kingdom with President Trump
and in video is at the heart of these negotiations
(23:31):
between the US and China. It's an index to watch
going forward, but it has seen a lot of momentum
of late character.
Speaker 4 (23:37):
It has someone who's not in the UK, is Mark Zuckerberg.
Why because Metaconnect is kicking off later today, Analysts they're
anxiously watching for any groundbreaking announcement like well, maybe some
smart glasses project. Brad Erickson, internet analyst at RBC Capital Markets,
joins us. Now smart glasses has been where it's at
from a hardware perspective. For meta how can they push
(23:59):
that story forward with?
Speaker 14 (24:01):
Yeah, I think it's just integrating more of the capability
with Marcu's vision on Meta AI, right, Like they've given
some applications in terms of some of the translation capabilities,
being able to have some of the augmented reality capabilities,
and I think you'll just see, you know, we expect
to see more of that tonight.
Speaker 4 (24:22):
So really the vision of generative AI coming in the
heart and why Meta AI is really amplified with the
hardware story. But the competition's coming. We know Google's got
them in the works. We know that, of course Apple
might or might not in the future. But Brad, therefore,
what sets their deal with ray ban apart? How can
they continue to offer a little bit more to the
experience than an augmented reality vibe.
Speaker 14 (24:43):
Yeah, it's always going to come down to the same
things that arguably made Apples so successful over the last
nearly twenty years, which is tight integration of hardware and software.
Speaker 6 (24:53):
Right.
Speaker 14 (24:54):
Meta has the platform that people already know it for
for the various family of apps they've created AI have
more than a billion users. The utility that that creates
and the stickiness that that creates for their customers, that's
going to be the key thing.
Speaker 12 (25:09):
The hardware everybody's going to be able to.
Speaker 14 (25:11):
Do right, just like they've always been with any piece
of consumer electronics hardware. It's the software that's going to
keep people coming back.
Speaker 3 (25:17):
If they're going to come back, Rad, thank you for
bringing up Apple, and you also reference translation. Last week
at the iPhone event, Apple shares swung from a decline
to positive territory, and Duolingo shares plunged when Apple demonstrated
simultaneous translation through the airport's Pro three a year ago,
Meta demonstrated simultaneous translation through that generation of rayband metas.
(25:43):
Why did Meta not get the same credit for that
technology that Apple did a week ago?
Speaker 14 (25:49):
Yeah, I think part of it is just it was
so new. Secondarily just relative to the stock. You've also
got a giant core business right that you're compying, and
that's clearly not going to be the focus of it,
or is it necessarily clear how you would how you
would yield an.
Speaker 12 (26:05):
Advertising business out of that.
Speaker 14 (26:06):
So I think that's just mechanically that's kind of the
primary reason. But clearly this is going to be, you know,
a capability for a lot of companies here going forward,
and certainly could be a super cool compliment as you know, again,
Meta looks to build in all these capabilities to these
new glasses.
Speaker 3 (26:23):
A lot of people still don't understand the Meta AI story.
With Chat GPT, I just have an app on my
phone or on my MacBook. I use it every single day.
Right With Meta, I have a standalone app. It's integrated
into Instagram, WhatsApp. I am an owner of the RayBan
metas a different interface or medium to use it. But
(26:46):
do you see kind of what the end thing is
going to be for meta how consumers actually interact with
the work they've done on the model side.
Speaker 14 (26:55):
I don't think anybody does candidly. I think one of
the things that I'm always struck by is we all
stay and learn AI every day is that, you know,
week by week, month by month, we kind of all
have to expand our minds on what the capability is.
Vibe coding is a great example something we've been spending
a lot of time in lately with the web design companies.
Speaker 12 (27:14):
People are using these things to.
Speaker 14 (27:16):
Develop applications, not to build businesses necessarily in the traditional
sense that's where everybody's brain would have gone.
Speaker 12 (27:22):
But they're creating daily productivity.
Speaker 14 (27:24):
Tools enterprises, they're using them to create actual internal work
processes right unrelated to their core software. So there's all
sorts of new things. I think METAAI is very similar
in that, you know, we know that it's going to
be able to provide like you said, chat GBT functionality
with the app. What we don't know is how else
(27:44):
might people build that into their day to day lives,
whether it's for.
Speaker 12 (27:48):
Work, personal or productivity or what have you.
Speaker 14 (27:51):
That's going to be value add and get to a
point where you're going to say, hey, I can't live
without this new capability in my life.
Speaker 12 (27:57):
And we don't exactly know what that is yet.
Speaker 14 (27:59):
We're learning about it, we're trying to sort of expand
our minds, but we haven't gotten there yet.
Speaker 4 (28:04):
Well, I mean, Mark Zuckerberg's been expanding the amount that
he wants to spend on it and the appetite and
ambition that's for sure brand and he's been given a pass.
And you think the stock could climb even higher up
to eight p fifty of another to nine percent gain.
And where we currently are, how much does the TikTok
story bear into its bread and butter right now?
Speaker 14 (28:23):
It matters, But TikTok is a very small business compared
to the size of metas overall ad business.
Speaker 12 (28:31):
I will say that, you know, for a quarter or.
Speaker 14 (28:33):
Two now in our channel checks talking with advertisers, what
we have heard was, you know, the stalled or the
uncertain future of TikTok and really caused advertisers to pause
any spending growth. They're still spending to be clear on
ads on TikTok, but there wasn't nearly as much growth
to the degree that we get a deal here at
(28:55):
some point maybe that provides clarity becomes a little bit
of a headwind. But I wouldn't I don't want to
like overstate it because it is TikTok's still you know,
somewhere in the call it ten percent of the business
relative to the size of metas ad business.
Speaker 3 (29:10):
Brad Ericson of RBC Capital Markets, thank you very much.
Lift and Weimo are partnering to bring self driving taxis
to Nashville.
Speaker 2 (29:24):
Customers will be able to.
Speaker 3 (29:25):
Request driver US rides on both companies platforms, and Lift
will provide fleet management for Waymo's national operations. Let's get
to more with Lift CEO David Rischer. This is something
you've been working on for some time. That's right, and
the state of play is very clear. Weimo also has
a partnership with Uber in different markets. You now have
a partnership in Nashville. Just explain the rationale for you
(29:49):
on why this is a great go to market for ROBOTAXI.
Speaker 13 (29:52):
Well, as you know, we've been saying for a while,
we think self driving cars are going to be hugely
market expanding for ride share. It's a really good product.
Speaker 2 (30:00):
And then when we looked at.
Speaker 13 (30:01):
Potential partners of course Weemo has to be at the
top of the list. The great thing about this partnership
is it does two things. It supplies demand, So if
you're in the self driving car business, you got to
have demand, and we got lots of demand every single day,
and we do it in a very integrated way. But
the other thing is, as you just mentioned, it supplies
fleet management. And that sounds kind of geeky, but it's
really important that every one of those cars be ready
(30:22):
to drive twenty four to seven if needed, and that's
something we're really good at. We've been doing that for
a long time.
Speaker 3 (30:27):
So you're emphasizing the sort of capital light business model
for you, right. The counter argument is the Tesla or
the zooks one, which is control the full stack hardware, software, operationally,
the app the consumer in faces with. Would you argue
that you can scale this faster, going this route, go
to more cities quicker?
Speaker 2 (30:47):
I would?
Speaker 3 (30:47):
Yeah.
Speaker 13 (30:48):
I mean this allows every company to specialize, right, so
Wimo can really focus on providing great self driving technology.
The way Moore Driver is obviously a great product. They
partner with OEMs, the car manufacturers that can choose the
best car, and then we can provide the fleet management
as they say, that keeps the cars clean and charged
and ready to go, as well as the demand so
that anytime a person wants a car, they can get
(31:09):
one or they can get a driver car, right, I mean,
they can get both. So it's I think it's a
great division of labor.
Speaker 4 (31:14):
Let's talk about that fleet management flex drive, because you're
going to be building, you're going to be putting some
capal expenditure to work David, Why does that make sense
and when will that be finished?
Speaker 2 (31:24):
So we've been in the flex drive world. Let me
explain for a second.
Speaker 13 (31:27):
So about a decade ago, Lift decided that for a
certain type of driver, it really makes sense to have
cars that they can rent right on the platform and
they can do we can take care of the maintenance
and making sure that those cars are always ready to drive.
So that's what we've been doing for about a decade now.
So when we look at the autonomous vehicle world, you know, look,
(31:49):
when you see a car driving around the city, it
looks like magic. It looks like no one had to
touch it, but no someone had to charge it, someone
had to clean it, someone has to maintain it. Some
probably has to reboot it from time to time, and
we're really good at that. So, as you say, we're
building a new depot there, we've got a small one.
We're building a bigger one for this partnership. And what
it does is it allows the cars to be ready
to drive.
Speaker 2 (32:07):
Ready to drive.
Speaker 13 (32:08):
Think about so when you travel on an airline and
you know that of course they want the airplane in
the air, but they also needed to be ready to
fly at all times. They want that turnaround to be
super fast. And that's what flex Drive does. It means
that people like Waymo can get the most out of
our asset, which is super good for them.
Speaker 4 (32:23):
Look today, investors love this announcement and the shares a rising.
Shares have been doing well all year because you've had
three straight quarters of profitability. David, how are you going
to finance what is a little less asset like is
building these facilities to help keep all our auto cars clean.
Speaker 2 (32:39):
Yeah.
Speaker 13 (32:40):
So, as I say, because we've been in the flex
drive business, the subsidiary business for a long time, we've
actually had capital deployed to that for a long time.
We have about twenty four locations in about fifteen different
states across the United States, which I think gives you
a sense of how this kind of partnership could scale
over time. And we own some of those cars ourselves,
others of those cars or financed by other groups. So
(33:01):
the good news is this won't be a significant change
to the current strategy. It's just sort of a revectoring,
let's say, of the money we spend more towards.
Speaker 2 (33:08):
Autonomous okay, scale over time.
Speaker 3 (33:10):
This morning, Bloomberg Tech producer Marguerite Gallery, and he put
this table together. You know the table, right, It explains
the roadmap from here, So let's look at it together.
This is twenty twenty six and it focuses on WEIMO principally.
But there's you there, right Nashville. You are going to
come up against suber in negotiating with WEIMO for cities
that are not yet claimed for what are the advantages
(33:32):
that you're going to present to win that business with
WEIMO against Duba And how much you're gonna have to
give up, you know, to get a deal done.
Speaker 13 (33:41):
Yeah, So I think the way I would look at
That is one of the reasons we were so patient
in working through this deal because it took a while.
It was we wanted to come up with something that
is good for Weimo, good for Lyft, and good for riders,
and that's hard to do. It's hard to have, you know,
kind of a three away win. We feel really good
about it and feel like this is a great model
for us to expand to other cities in the future.
(34:02):
But look, city by city, you know some places we're
actually going to compete with Weimo itself, right because for
example here in San Francisco, they've got their own appreciate
that admission, Yeah, of course, So this is something that
we have to kind of recognize and say, how do
we come up with a construct that is really durable
and frankly can benefit all the companies and then the
other guys will do what they do and you know,
we'll see who does a better job.
Speaker 3 (34:21):
Look, people are going to accuse me of wanting to
create a headline here, but it's a really important question.
Tesla and Zeke's are committed to their own go to
market model. Right, we talked about they control the full stack,
But have you held discussions with both Tesla and Zeke's
to say like, look, this is the model we're doing
with Weimo, we can do something similar.
Speaker 2 (34:41):
Do you see that as being realistic?
Speaker 13 (34:43):
So of course I don't want to come in on,
you know, conversations we might be having. What I can
tell you is we're talking to everyone who finds the
partnership model interesting. You know, there are, of course always companies.
Whenever you see a new industry, you have some companies
that say I.
Speaker 2 (34:58):
Want to do it all.
Speaker 13 (34:58):
And frankly, if you look back a couple years ago,
just about every company in the self driving space was
in that zone. They wanted to do everything, you know,
top to bottom. I think over time most companies find,
particularly when it's capital intensive and quite complex, like ride
share is and self driving is that a division of
labor where everyone does what they're good at, is the
is the best model. But we'll see, you know, it's
still such early days.
Speaker 4 (35:20):
But you struck deals with interesting players in Europe. I do,
what's it like working with a Chinese giant? What were
those negotiations like? You know that?
Speaker 13 (35:28):
Actually, I'll tell you what was very interesting. It came
together fairly quickly and in part I think is you know, frankly,
Chinese companies are kind of built for speed. You know,
they for those of you who don't know, you know,
by do you can sort of think of them as
kind of the Google or Alphabet of China. They have
many businesses, but self driving cars are a very important
one for them, and we committed that we would roll
out some of their technology in one to two markets.
(35:50):
We would expect in Europe over the next twelve months
or so.
Speaker 10 (35:53):
It's still very.
Speaker 13 (35:54):
Early days there, but I can tell you they've got
really high standards for their technology. We're super excited about
working with those guys in the wall to have more
to say about that next year once we start to
roll out in Europe.
Speaker 5 (36:03):
One minute, how hard is it to get to New York?
Speaker 13 (36:09):
I mean I can take a flight there pretty easily.
Speaker 4 (36:11):
But also the ROBOTAXI how hot? When am I going
to have it here?
Speaker 10 (36:16):
I love that?
Speaker 13 (36:16):
You know, it's going to take a little while. New
York is a complex place to do business. The streets
are complicated, you know, there's a lot of interest there. Look,
you know this Caroline, I mean lift has been there
for many years. It's super strong in New York. We
also run the city bike program. As you know, super
excited bring self driving there someday, but that was probably
gonna take a little while.
Speaker 2 (36:32):
Sorry about that.
Speaker 12 (36:32):
You can have to work for all.
Speaker 4 (36:33):
My addiction with the bikes will just have to continue.
Speaker 5 (36:35):
Lift to see David so as a joy to have
you on the show. Take Care.
Speaker 3 (36:46):
Mentioned earlier in the show that Microsoft will invest thirty
billion dollars in the UK. It's part of a push
to expand AI infrastructure. We also heard from GROC about
its data center plans. Of course, tech companies are pledging
to build more data center steer around the world. Let's
bringing Maria Davidson, CEO of Kojo. It's the largest construction
procurement platform in the US, giving her a unique insight
(37:09):
into the data center build out. I think we can
get to where you fit into the process of what
is just massive infrastructure as a project, but that your
reaction first to what's happening in the UK. It seems
to be the next market in this AI cycle.
Speaker 15 (37:25):
Oh absolutely. I mean when you think about how much
potential there is in the UK, I think it is
extremely exciting and very very untapped. So I love to
see companies like Microsoft and Google coming in and investing
in the UK, and the US and the UK actually
working closer together.
Speaker 2 (37:40):
You say a lot of potential in the UK.
Speaker 15 (37:42):
Why, I think because one, there's extremely talented workforce in
the UK, and you know, you have world class universities.
You have incredible research like deep Mind coming out years
before anyone would have expected an outcome like that in
the UK, and so I think that means that in
the UK right now, if we keep on investing, I
(38:04):
think we're going to see a lot more deep Mind
like outcomes.
Speaker 4 (38:07):
I've got to say, You've got three Brits around a
table talking about the opportunity in the UK. So I'll
diverse us a little bit because we're three Brits that
sit in the US right now. Maria, you're building businesses
to try and make supply chains clearer for contractors for
builders out there, and I'm interested as to what the
scale is like here in the United States and what
the key bottlenecks are that you're trying to fix.
Speaker 15 (38:26):
I mean, the scale is enormous. So just this year,
we have forty billion dollars going into the construction of
data centers in the US, that's expected to grow more
than ten percent a year, so by twenty thirty that's
going to be sixty five billion dollars. That's unprecedented. And
the biggest issue with building these it's not getting GPUs,
it's getting the materials and finding the electricians to be
(38:48):
able to actually meet this demand.
Speaker 4 (38:50):
Okay, so your role is ensuring that when the contractor,
which is very hard to come by and labor is short,
and we've seen immigration policies hit that, but when you
get the on site, they've actually got the pieces in
the parts to work with what has been some of
the most complex areas when it comes to particularly the
hyperscaler's need and the data center build out.
Speaker 15 (39:09):
It's a lot when you think about the construction industry
in the US. Even though they're building some of the
most sophisticated infrastructure out there right they're building projects that
involve huge amounts of technology and AI infrastructure specific needs,
they still largely have to rely on pen and paper
and Excel to track where things are. It's hard to
(39:30):
know what's arriving when these are all huge challenges. And
so when you couple that with the fact that there
are five hundred thousand workers missing from the US construction workforce.
And you couple that with the fact that materials on
average of forty four percent more expensive than they were
in February twenty twenty before COVID, that starts really creating
(39:50):
a perfect storm.
Speaker 3 (39:52):
I actually think it's worth going a step earlier in
the process than what Caroline just asked about. Right, We've
kind of learned over the course of a lot two
or three years about how it works in the United States.
You have a financing partner, you have someone like Cruso
putting it all together.
Speaker 2 (40:06):
You need to select a site.
Speaker 3 (40:08):
Then you've got to talk to the utility because you
need to power that site. The utility has to think
about where they get the electricity from. Is that going
smoothly and well that that initial process it depends.
Speaker 15 (40:20):
It's definitely gotten a lot faster. If you think about
your average size data centers right as data centers on
the smaller side maybe ten to thirty megawatts. Those previously
would have taken that's say, twenty four months to thirty
six months to build. Now you're seeing data centers build
probably twice as fast, in part because things like site readiness,
inter connection, they're all being dealt with in advance.
Speaker 2 (40:42):
Now.
Speaker 15 (40:42):
I think prefab is a huge subject here because the
more that we can lean into modular construction, the more
we don't have to rely on the sequential building, but
we can actually build in parallel.
Speaker 3 (40:52):
There's also like a self fulfilling prophecy in the sense
that all of the AI names who actually just end
up leasing the capacity, right, they don't even own the thing,
but they would say, Oh, all the work we're doing
in AI is actually going to benefit the planning and
execution of these buildouts. I guess that's what you'd say.
You also help.
Speaker 15 (41:09):
With Yeah, absolutely, I mean a lot of it is.
The better planned that you can be in advance, the
better the project is actually going to go. And so
if you know all of the components that you're going
to need, if you have the mapped out schema, if
you have all of those interconnection blocks ready to go,
then of course contractors can be much better prepared.
Speaker 4 (41:29):
I want to talk about you about Kojo, about what
the ninety four million dollars you've thus far raised six
hundred construction companies that you're working with. You've just had
an add on to your CERC I think I believe
how much are.
Speaker 5 (41:42):
You seeing inbound?
Speaker 2 (41:43):
How much are you.
Speaker 4 (41:44):
Fighting off people wanting to put money into your company
because they're realizing it's about bricks and mortar, it's about
the very underlying infrastructure.
Speaker 15 (41:50):
Right now, well, we're seeing a huge amount of inbound.
The focus for us has always been how can we
empower contractors, how can we take the contract is that
for a really long time were completely ignored by the
tech industry, and specifically there I'm talking about the trades, right,
the electricians, the folks working in mechanical and HVAC and drywall,
(42:12):
They for a very long time had very few solutions
that were built specifically for them. And so for us,
it's about saying, well, how do we build solutions that
are specifically targeted towards you and that solve your biggest
pain points like knowing where your materials are and having
the visibility of how you're doing against your estimates in
real time.
Speaker 4 (42:31):
Well, you're just completing a ten million dollar Series C extension.
Speaker 5 (42:35):
Come back when you're raising yet more.
Speaker 4 (42:37):
I'm sure Maria Davidson's CEO, Kojo, it's great to have
you on the Showwow, that does it for this edition
at Bloomberg Tech, we've done our whole host of things zed,
but we're going to count down to Metaconnect tonight. Of course,
tune in tomorrow for an interview with NEA's Chief product officer,
Chris Cox live on the show tomorrow, which ed.
Speaker 5 (42:54):
You'll be doing.
Speaker 2 (42:56):
Yeah, looking forward to that.
Speaker 3 (42:57):
I have a lot of questions about the relationship between
software and hardware.
Speaker 2 (43:00):
After all, that is technology. Check out the pod.
Speaker 3 (43:03):
You know where to find it on the Bloomberg terminal,
as well as online on Apple, Spotify, iHeart, and all
kinds of other places on the internet.
Speaker 2 (43:12):
Thank you for watching. This is Bloomberg Tech.
Popular Podcasts
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com