November 25, 2025 • 44 mins
Bloomberg’s Caroline Hyde discusses the fall in Nvidia shares following reports that Meta is in talks to spend billions on Google’s AI chips. Plus, Chinese leader Xi Jinping revived the topic of China's sovereignty over Taiwan in a phone call with President Donald Trump. And Apple eliminates sales roles in a rare layoff to streamline the way it offers products to businesses and governments.
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 (00:00):
Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live
from coast to coast with Caroline Hide in New York
and Eva Loow in San Francisco.
Speaker 2 (00:22):
This is Bloomberg Tech coming up, Meta Iron, Google's AI chips.
According to reports, how the market is questioning in Vidia's
long term dominance. Prashujinping brings the issue of sovereignty over
chiphub Taiwan back on the agenda with Donald Trump during
their phone call. And Apple eliminates sales roles in a
rare layoff to streamline the way it offers products to
(00:42):
businesses and governments.
Speaker 3 (00:44):
We have the details.
Speaker 2 (00:45):
The first, let's check in on these markets that are
trying to digest a wall of delayed data at the
same time as baking in maybe an eighty percent chance
of a FED rate cut. That's what the market price
is in but all eyes on tussle at the top
in AI dominance, looking at a five ten percent drop
on the Nasdaq one hundred on the upside, some key
players on the drownd side, we're seeing a key drag
(01:07):
from some of the key chip names.
Speaker 3 (01:08):
In particular.
Speaker 2 (01:09):
Let's delve into what's moving this particular index, which does
lag the SMP and the DOW. Today we're seeing in
video off by five percent. No wonder then, ASAK one
hundred is underwater when the world's most dominant player from
a market cap perspective is currently off by five percent.
Big moves for a four trillion dollar company, and in
large part it's because of Alphabet that actually is trimming
its games but still at a new record high. This
(01:30):
is we understand meta eyeing tpused. This is all about
the dominance of where we get the chips of the
future to train our models as well as use our models.
Meta is the key.
Speaker 3 (01:41):
Line of questioning.
Speaker 2 (01:42):
Here is meta eyeing tpused from Google to put in
its own data centers. They're the reports coming from the
information today, let's talk about the market reaction. Equity reporter
around Nostellica is here with us Ryan. It's monumental move
that Alphabet has had since mid October. It's added a
trillion in market capitalization. Were starting to become aware of
its prowess in AI and in chips.
Speaker 4 (02:05):
Yeah, absolutely, thanks for having me so. Yeah, certainly sentiment
has really reversed on Alphabet, and I think people are
really appreciating how dominant. It is across every layer of
the AI stack. Of course, it recently released the latest
update to Gemini, which was seen as very strong across
all the major benchmarks that people used to evaluate AI models.
(02:26):
The chip business is getting a lot of attention lately,
between the report with Meta and the deal with Anthropic.
People really see this as a really significant potential competitor
to Nvidia, as you were discussing earlier, and beyond that,
it has a huge cloud business that is seeing accelerating growth.
It has so much data and users and distribution and talent.
(02:48):
It really has all of the pieces, and that has
really helped the stock surge, not just recently but over
the past several months. It is by far the biggest
performer or the best performer of the mag seven this year.
Speaker 2 (03:00):
I think Counterpoint Research analyst and co founder Nail shar
and a story on Blueberg saying it's a sleeping giant
in the AI race and is fully awoken. But many
have been trading for the last few days ever since
Gemini three release.
Speaker 3 (03:13):
Maybe going short open.
Speaker 2 (03:15):
AI suppliers and long alphabet suppliers. Is that still bearing out.
I'm looking at Oracle down once again.
Speaker 4 (03:22):
Ryan Oracle has really been under a lot of pressure lately.
I think it's on track for its biggest one month
dropped since two thousand and one, so certainly a real
reversal there. It does seem like people are moving towards
the Google and Alphabet suppliers, which is companies like Broadcom,
while at the same time the companies that are more
connected to open Ai. You mentioned Oracle, also AMD, Microsoft
(03:44):
to a certain extent, these companies have really been under
a lot of pressure lately.
Speaker 2 (03:48):
It's worth reminding though that yes, Alphabet's been on a tear,
and actually its valuations have started to trade way higher
than we're used to I thinking about twenty six twenty
seven times future earnings, but still leap from just mysoft.
Speaker 3 (04:00):
Ahead of it in terms of market cap is Apple.
Speaker 2 (04:03):
I think it's got another seven percent to go to
hit that, but fifteen percent to go to in video.
We're not questioning longer term at this exact moment. In
Vidia's dominance in AI and semiconductors.
Speaker 4 (04:14):
Well, NBA right now has I think ninety percent or
so share of the data center market. If there is
potential for Alphabet to start eating into that, I think
that would change a lot of calculations for people in
terms of how much can Alphabet scale this business. I
have talked to someone who thinks that this could potentially
be worth more than Alphabet's cloud business, potentially up to
nine hundred billion dollars. So that is a huge potential
(04:36):
market there. And if that means that Nvidia starts losing
market share, I think people will start reassessing how to
value that company and its growth and its valuation. Now,
I will just simply say that the AI market is
growing so rapidly that people do see room for a
lot of big players, and even Alphabet remains a major
customer to in Vidia just because there is so much
(04:58):
demand for compute right now, people don't want to be
behold into any single supplier. It does suggest that there's
a lot of room for growth to go around, even
if we start seeing some erosion and market share at Navidia.
Speaker 2 (05:10):
And I'm pretty sure Jenson Wang will be responding to
any concerns about erosion and market share round a SELCA.
Speaker 3 (05:16):
We so appreciate you joining.
Speaker 2 (05:18):
Let's dig in to further analysis here with Stephanie Anagai
to the conversation which has global market strategist Japing Morgan
Asset Management and has four trillion dollars in assets under management,
four trillion, rather similar to the market capitalizations in some
of these companies, And I'm interested, Stephanie, does it matter
to you from a market sentiment perspective if there's scribbling
(05:38):
at the top of the US domiciled companies.
Speaker 3 (05:42):
Is it a.
Speaker 2 (05:42):
Worry that we'll see perhaps in Vidia being questioned in
terms of its dominance.
Speaker 3 (05:46):
I think it's quite healthy.
Speaker 5 (05:48):
We've seen the AI trade has delivered enormous returns for
markets over the last two years, and we're I think
all kind of experiencing the sire of relief with this exhale.
I guess in a way, we've moved from a rise
tied lifting all boats or more choppier waters, and investors
are being far more scrutinizing when it comes to how
much is being spent, the quality of those investments. We're
(06:10):
seeing as this AI trade continues to grow in its enormity,
the investment being made, and the moat in such big
questions and shifting. So I think it's quite healthy that
we're focusing on selectivity. I mean, this is what you
want to see to prevent a dot com bubble, and
we're looking at valuations today, we don't really see a
big risk of that happening amongst the big tech firms.
Speaker 2 (06:32):
So even these worries about debt in particular, and you
are coming to us with a viewpoint that is cross asset.
In many ways, we have seen a real desire to
get into AI related debts. Some of the bond sales
that come from the likes of Alphabet and the likes
of Meta and an Oracle.
Speaker 3 (06:46):
Have been scooped up.
Speaker 2 (06:48):
But there's been this worry that in the longer term
it might start to maybe pull back on overall demand.
We might see some of these big hyper scales coming
to market so often that it drives up prices for others.
Speaker 3 (07:00):
It's possible.
Speaker 5 (07:00):
I mean, I think first just looking at the magnitude
of how much AI investment is needed or being spent
already by any way that you cut it, the amount
of spending right now is enormous. But just like Cardia,
isn't that expensive for a billionaire. When looking at these
capex relative to the sales from these companies, relative to
their current revenue growth, which has also grown significantly, it's
(07:22):
actually not that extreme. So we're seeing this increasing move to.
Speaker 3 (07:25):
Tap into debt markets.
Speaker 5 (07:27):
But for us, it's not so much these companies getting overextended,
but actually more so a reflection of a better sorry,
a better capital structure. You know, there's some investments like
these data centers that are going to be invested over
multiple years. It might make more sense to tap debt
markets for some of these deals or an off balancing structure.
Speaker 2 (07:47):
So for example, we're going to be looking not only
at AI investment surging, but at credit default swaps of
Oracle surging. Has that just been acting as a bell
weather and the necessary bell weather to stop reflecting some
of the risks that maybe people been ignoring for the.
Speaker 3 (08:00):
Past few months.
Speaker 5 (08:01):
I think it's it's very apt, and not all of
these companies have the establishments, you know, they differ in
many different ways and also in their sources of revenue.
And I think it makes sense that you know, Oracle
is one of the more you know, risk your companies
that is tapping these bum markets and you're seeing that
being reflective in CDs spreads. But that also can't be
extrapolated to the entire shift right now towards debt markets
(08:25):
to help finance these data center bills, and I'll also add, look,
when it comes to cloud services, that business model is
one of the most cash generative business models.
Speaker 3 (08:35):
In the world.
Speaker 5 (08:35):
So at the end of the day, these bonds are
also being tied to services business operations that have tended
to do quite well for these companies.
Speaker 2 (08:43):
It's interesting, of course, deciding where the margin a cruise.
We're going to have l HP after the bell, Many
feeding that margin is being eroded because of the cost
of memory. Meanwhile, we'll getting Micron next week with its
Sennings and Many anticipating they're strong because.
Speaker 3 (08:55):
Of the memory demand there.
Speaker 2 (08:57):
From your perspective, is there still room to run in
just the tech trade more broadly, or has that shift
into more value names and certainly with the context of
the FED change things longer term into the end of
the year.
Speaker 5 (09:09):
We still think we're quite early in this AI wave,
but we've seen a chapter or two, and moving forward,
I think the focus is not going to only be
on compute needs and capacity needs, but also on AI
utilization and what companies are really critical for that, whether
it's in software, what companies are leading the way in
financials and entertainment, in adopting AI, and then also how
(09:33):
once we learn more about the end user demand for
AI and the pricing power of these AI services, that's
going to give us a lot of clarity around the
ROI around these AI investments.
Speaker 3 (09:43):
So is that what we need?
Speaker 2 (09:44):
Is it ultimately the revenues of companies outside of the
world of tech to vindicate that? What pushes us higher
in terms of real contextview? Is it December when we
get the FED decision? What is the catalyst do you
think for us to reassess where we are in valuation?
Speaker 5 (10:01):
I'd say it's less of the kind of macro back
job early here and much more of the proof point
around the monetization of AI. I think the more that
you see businesses ramping up their IT budgets, the stickiness
that you see in those investment spending and then also
AI delivering and we've seen some proof cases of that
so far. Coding has been a huge factor of that.
(10:23):
But once you see more companies, particularly outside of tech
maybe tech adjacent, coming at their earnings calls and talking
about their AI generated savings, I think that's going to
be a really important next lever for the AI trade.
Speaker 2 (10:36):
And then one about the Leva FERRATUALI the companies that
are adulting the clients that are cooling you on a
daily basis, are they saying do I double down more
in tech? Are they saying I need to double out
outside of the world of tech.
Speaker 5 (10:47):
I think it's it's diversifying that tech exposure. You know,
after a long run in these AI names, you don't
want all your ex in one basket because it is.
Speaker 3 (10:57):
Probably not.
Speaker 5 (10:57):
At least you want to right size some of that
exposure bill to top of all of those games that
we've experienced in position for how this AI wave is
going to evolve, there will undoubtedly be losers and winners,
but we also don't want to be out of the market.
Speaker 3 (11:10):
And that's another thing that we're trying.
Speaker 5 (11:11):
To talk to clients about because even when you call
a bubble correctly, if you weren't in the market from
nineteen ninety five to nineteen ninety nine, you would have
missed out on over four hundred percent in total return
in the Nasdaq. You were right, but you've locked in
years of underperformance. So when it comes to us f
Wing Marcus today, we don't see that real risk of
a systemic bubble, but we do see a real opportunity
(11:32):
to just make sure that portfolios are built for resiliency
and they're also built to take advantage of how this
ai wave continues.
Speaker 2 (11:39):
To evolve, well, hopefully keep having you on a zimbubble
or indeed the narrative does evolve. We so appreciate Stephanie
Aliaga JPM Wollngan asset management can cross tech for us.
Mean while coming out with China, Shijin paying revives talks
to the sovereignty of a Taiwan and a phone call
with President Trump.
Speaker 3 (11:56):
More on that next, This is boom Bag Tech.
Speaker 2 (12:10):
Chinese President Xijingping well has revived the topic of China's
sovereignty over Taiwan in a phone call with President Trump yesterday,
discussion that didn't come up during their face to face
meeting last month in Beijing. From Magxinia, tech editor Mike
Sheppard joined us for the latest and Mike remind us
from a tech perspective, Taiwan, we know its dominance and chips,
(12:31):
We understand its integral nature to the tech ecosystem. What
is happening between Xijiping and Trump on.
Speaker 6 (12:37):
This Well, what was interesting yesterday, Carr is that we
get two very different versions of this phone call between
the leaders of the world's two largest economies. The first
version came from Beijing. The official state news agency SHINOA
put out its version of the conversation, presenting it really
as one centered on the question of Taiwan. And then
(12:59):
a few hours later we heard from President Donald Trump
himself on truth social posting that they had had a
great conversation about issues including soybeans and other matters, rare earths,
and other key topics that were deared to the US president,
but he made no mention of Taiwan there Now, while
he did not bring it up, several hours later, he
(13:20):
did call the new Prime Minister of Japan, Sanai Takaichi,
who had enraged Beijing with comments a few weeks ago.
You'll remember Caro saying that Japan would consider leaping to
Taiwan's defense in the event that China.
Speaker 7 (13:35):
Were to try to take it.
Speaker 6 (13:36):
Now, all of this is huge implications for the supply
chain of semiconductors, as you noted, especially Taiwan semiconductor. It's
one of the world's largest producers of AI chips, and
they are moving some of their production, as we know,
to the US they've pledged one hundred and sixty five
billion dollars in investment implants here in the United States,
(13:57):
but they would still retain a significant amount of their
capacity on the island. Therefore, any question of Taiwan really
does bring up tech issues for US.
Speaker 2 (14:06):
And talking of tech issues, the administration once again trying
to signal its commitment seeing the ai infrastructure build out
akin to the Apollo Mission or to the Manhattan Project, Mike.
Speaker 6 (14:17):
Now they are talking about a Manhattan Project like effort
in this Executive Order to call Genesis that President Donald
Trump signed yesterday, But it was really more a call
to action for various agencies to start working together more closely,
and that includes the Department of Energy and its National
Research laboratories. But when we talk about Manhattan Project, though,
(14:39):
we do need to remember that that effort took thirty
six billion dollars in real dollars today from back then
in the nineteen forties, as the United States was in
the race to develop a nuclear weapon ahead of the
Access Powers. This is a very different time. We are
not seeing new money being pledged towards this effort. It's
(15:00):
important to remember, Caro, though, that the US already has
put a significant investment in production of chips that would
be needed for artificial intelligence, and that is the Chips
and Science Act of twenty twenty two that put tens
of billions of dollars in loans and grants and other support,
including tax incentives, to support the development of a chip
(15:22):
industry domestically that would help artificial intelligence take hold and
gain ground and lead the world, as President Biden and
President Donald Trump now say they would like the US
to do.
Speaker 7 (15:34):
Now, we are.
Speaker 6 (15:34):
Seeing companies like open ai push for further investment to
support data centers, and that would include extending some of
those tax credits to data centers. So it'll be interesting
to see how this develops and whether more of those
tax credits will go to some of those AI projects.
Just Caro, as we are wrestling with those questions of
whether we are seeing too much money going into this space.
Speaker 2 (15:56):
Max make Shepard all the context from Washington, We appreciate it.
Speaker 3 (16:00):
Meanwhile, it's time for talking tech now.
Speaker 2 (16:02):
First up, Ali Baba reported thirty four percent growth in
its cloud unit during the September quarter. Despite the gain,
though spending on consumer subsidies data centers that's eaten into
its profits the company's ADRs, as you can currently see,
just training.
Speaker 3 (16:15):
Off by some two percent day.
Speaker 2 (16:17):
Meanwhile, Anthropic it's got a new model, claud Opens four
point five that the company says it's better at coding
and office tasks such as financial analysis and creating presentations
or spreadsheets.
Speaker 3 (16:27):
It's part of anthropics efforts.
Speaker 2 (16:29):
To compete with open Ai with Google business customers. And Meanwhile,
open ais a new tool to generate personalized shopping guides.
The company trained a version of GBT five Mini.
Speaker 3 (16:40):
Model to ask follow up questions.
Speaker 2 (16:42):
Draw answers from reviews published on what the company considers
higher quality websites. Now, let's talk Apple, because, in a
rare mood for the company, the tech giant has eliminated
dozens of sales roles in an effort to streamline the
way it offers its products to businesses, schools, and governments.
For more on this breaking story, Mark German joins us,
(17:03):
it's not one hundreds, We're.
Speaker 3 (17:04):
Talking tens of people.
Speaker 2 (17:06):
But still it's notable because we don't often see layofs
of that one.
Speaker 8 (17:10):
Yeah, to your point, it was several dozen people across
Apple sales division, and this sales division they partner with
carriers across the world to sell iPhones, but they also
partner with enterprises, large scale businesses, government organizations, schools, educational institutions,
major universities across the world to sell products like iPhones, iPads,
(17:33):
max and you name it. And over the course of
this month, there was a big streamlining rounds of layoffs, including,
like I said, several dozen people. There were account managers,
they are called account executives for specific government agencies, for
specific university systems, people who partner in pitch companies on
(17:53):
buying Apple products. There are these tiny Apple store like
fixtures called briefing centers at Apple offices in California and Texas,
and the people managing that, for the most part, we're
laid off as well. And so this is going to
change how Apple sells products to these different organizations. The
majority of products are bought through what's called the channel,
(18:15):
so third party retailers, and so those products are still
going to sell, but quite a bit of a shake
up here for Apple sells products and delivers these devices
to the major customers. And of course, as you said,
a rare layoff for Apple.
Speaker 2 (18:30):
Apple did respond to your reporting and saying we're continuing
to hire, and those employees can.
Speaker 3 (18:35):
Apply for new roles.
Speaker 2 (18:36):
Mark, But what do you think this signifies more broadly
about how Apple is trying to streamline, trying to become
more efficient, trying to ensure that it doesn't seem like
it's a lagged in this age of AI.
Speaker 8 (18:48):
Yeah, you know, I don't think this has much to
do with AI. I think this has to do with
cutting roles internally to lower costs because they realize most
of these sales are happening from the channel and there's
a lot of duplicate efforts internally with the channel, so
the third party retailers. So I think it's just one
of your classic layoffs to create more efficiency and cost cutting,
(19:10):
rather than having much to do with artificial intelligence. In
terms of layoffs related to AI at Apple, I guess
the only thing you've seen related to AI from Apple
that has to touch a layoff was the self Drenking
Car project job cuts of one thousand people in the
beginning of twenty twenty four, and that was actually to
do more AI rather than because of AI. So they
(19:32):
moved a lot of those folks over to the Generative
AI division. But I haven't seen any job cuts at
Apple to date because of AI. That doesn't mean they're
not going to happen, but so far they haven't.
Speaker 3 (19:43):
Well that's far.
Speaker 2 (19:44):
We started to still see some job cuts across technology
and Mark Gurman, you've been at the front of that reporting.
Speaker 3 (19:49):
We really appreciate it.
Speaker 2 (19:51):
Now coming up, Michael Barry stands by his in video
criticisms as after calling out the company for stop by
gap facts for compensation dilusion, but in video itself responded
to analysts.
Speaker 3 (20:02):
More on that next. As a Bloomberg Tech.
Speaker 2 (20:15):
As we've been reporting in video shares, they are under
pressure today, Competition fierce when it.
Speaker 3 (20:19):
Comes from Alphabet and TPUs.
Speaker 2 (20:21):
We understand reporting that Meta is eyeing potentially turning to
Google for its chips in its data centers in the future.
But there's also Michael Burry there isn't there standing by
his criticism of the company after a video pushed back
on his earlier analysis of stock based compensation of share
buybacks for more Bluemberg equity supporter com and Ryanikey reminds
us of what the Cassandra as he dubs himself, has
(20:44):
been saying. Michael Burry laid on issues about the circularity
of tech deals, worries about the interoperability of what big
tech are currently doing in the world of generator AI.
What did he take issue though, within video when it
comes to share buybacks and stock compensation.
Speaker 9 (21:00):
Yeah, so, I think most basically his argument is that
the amount of stock based compensation is diluting you know,
owner's power, that if you hold the stock, it's being
diluted by the stock based compensation. And so it really
is just another thing and sort of a myriad of
things that he has called out within Vidia in recent weeks.
And yeah, we saw the company, you know, push back.
(21:22):
There was a memo that they sent to some Wall
Street analysis is according to a Baron's report that said,
you know, we think his math is wrong. Kind of
explained the situation a little bit better and also very
latantly stated, you know, we're not Enron. We're not you know,
there's no fraud here. But but Bury really said, you know,
I stand by my analysis that you know about the
(21:45):
stock based compensation, dilution share buybacks. And he also said,
you know, I'm not comparing Nvidia to Enron, I'm comparing
it too Cisco, which I thought was really interesting thinking
about Cisco, you know, in the dot com era, it
had this huge run up, but it was really associated
with the over build in fiber optic cable. So he's
you know, comparing that to what's happening now, I guess
(22:07):
with the video with data centers, and these are really
some of the biggest you know, concerns or pain points
that we're seeing in this debate over if AI is
a bubble, and you know, in video shares are down.
I think they were down as much as six percent today.
We're seeing you know, more than two hundred billion dollars
in market value just wiped off, and we're also watching
the sort of twenty percent level and videas nearing twenty
(22:28):
percent draw down from it's high at the end of October,
which was a significant level.
Speaker 3 (22:33):
For the shares technical band market. Extraordinary. Thank you very much,
comment and Rhanick.
Speaker 2 (22:37):
He always has some of the most read stories across
all of the technology moves.
Speaker 3 (22:41):
You've got to keep them up to date with it.
Speaker 2 (22:42):
Meanwhile, coming up Google's potential chip deal with Meta's raising
questions about in Vidia's dominance and the erase for AI leadership.
Speaker 3 (22:49):
More on that take next. This is boom Bag Tech
Welcome back.
Speaker 2 (23:03):
To Bloomberg Tech, let's take a check on these markets,
because we have seen some sell off continuing.
Speaker 3 (23:08):
In the world of technology.
Speaker 2 (23:09):
Unlike SMP, unlike the Dow, the Moon music remains resolutely
in the red.
Speaker 3 (23:13):
We're off by four tenths of percent.
Speaker 2 (23:15):
We're seeing some of the big tech names, namely in video.
Speaker 3 (23:18):
On the downside.
Speaker 2 (23:19):
We're still questioning valuations as we get that myriad of
data that comes late to the party when it comes
to certainly our own consumer sentiment seems to be on
the low side. But we're seeing retail sales maybe pointing
towards whether or not we've got some resiliency in the
overall macroeconomic picture. But in Vidia is more a story
of its resiliency versus competition.
Speaker 3 (23:37):
We're down by four percent.
Speaker 2 (23:38):
Once again, we're wondering if other chips will be created
by other players, like Alphabet for example. It's TPUs maybe
being eyed by Meta. That story we're going to delve into.
We're seeing both shares trade higher. Oracle and any name
really in the open Ai ecosystem has been a typical
short for the last week or so, as people question
open AI's dominance versus Gemini. Three for example, so Oracle
(24:00):
once again off by another two percent. Let's really dig
in them into the story of the day of alphabet
really giving m video a run for its money, certainly
a market capitalization front at least Mandy saying, you're here,
senior techanalyst a roomag intelligence.
Speaker 3 (24:13):
You have for months, if not years, been reminding me and.
Speaker 2 (24:17):
Our viewers of the power of the ecosystem of Google
and TPUs.
Speaker 3 (24:21):
Why now are we only just getting it?
Speaker 7 (24:24):
Wow?
Speaker 10 (24:25):
Because Gemini three showed that you could use the TPUs
both for training the model and for inferencing. I mean,
so far the story was all these secondary providers could
be used for inferencing. The fact that TPUs were used
for training Gemini three and most likely for Entropics Frontier
model as well. So two of your three frontier models
(24:47):
are using TPUs, And I think there is an acknowledgement
now from the market that TPUs are comparable to in
video in terms of functionality, and they are a lot cheaper,
which is why I'm not surprised to see Meta do that.
I mean, they will raise CAPEX and you know, going
to the secondary provider who is a much cheaper option
(25:09):
makes sense. And I think a market will like it
when they raise the capex and say that we'll be
using Google as a secondary provider.
Speaker 2 (25:17):
These are thus just reports as it stands, Mandy.
Speaker 3 (25:21):
But what's interesting has.
Speaker 2 (25:22):
Been looking at Jensen Wang's reaction when Alphabet or others
have made inroads into some of their key clients, Anthropic
for example, getting a load of gtpus from Google, and
then we see more of a deal done between in
Video and Anthropic.
Speaker 3 (25:38):
We know that in Videos double down on open AI with.
Speaker 2 (25:41):
One hundred billion dollars being offered in return for GPUs
being considered in their future training. So what do you
think the response mechanism could be of Nvidia.
Speaker 10 (25:50):
I mean, right now, in Video's problem is no one
wants to pay the you know, the high costs they
have for their chips, and that gets reflected in in
Video's margin. Seventy five percent gross margin is something we
have never seen with a semiconductor company. So from that
perspective that you know, the providers who are doing inferencing
(26:10):
are offering their products below costs. Even an open AI
when it's deploying you know it's chatbot at scale, I
would argue, you know, their gross margins are negative because
they're offering their product at below their costs. In the
case of Google, they're deploying you know, generative AI at
scale on search and across their family of apps, and
(26:32):
they're able to do it without really hurting their margins
because their cost phase is much lower. They're running their
infrastructure a lot more efficiently, and so that is where
the problem lies is you can't just keep subsidizing the
inferencing because the cost of your chips is so high.
With Nvidia, and you've got to find a way to
bring down the cost. Everyone wants more inferencing. You just
(26:55):
have to bring down the cost so that you know
they can do it profitaboud.
Speaker 2 (26:59):
Well, see how the response does indeed turn out, and
how they compare the software offerings as well as the hardware.
Man keep seeing a BlueBag intelligence always across the story.
Speaker 3 (27:08):
We so appreciate it.
Speaker 2 (27:09):
More insight into AI and indeed margins is set to
come after the Bell, Dell HP they report Bluebag's DNA.
Speaker 3 (27:16):
Bass gives us the preview.
Speaker 2 (27:17):
We're just hearing about the very healthy margins that Nvidia has,
and in many ways it's because the margins of Dell
and HP and server offerers are much thinner.
Speaker 3 (27:27):
Sure, Yeah, Dell.
Speaker 11 (27:29):
The most watched part of Dell's business for the last
couple of quarters has been its AI server.
Speaker 3 (27:33):
Business, many most all running in media.
Speaker 11 (27:37):
GPUs, and they have some a really marque list of customers.
Speaker 3 (27:41):
There's corewe there's XCAI.
Speaker 11 (27:43):
We reported the other week there they just got a
deal for the first Armenian AI data center.
Speaker 3 (27:48):
The problem is, in order.
Speaker 11 (27:49):
To win some of those deals and to execute on them,
Dell has.
Speaker 3 (27:53):
Had to put up with some pretty.
Speaker 11 (27:55):
Narrow margins over at eachp. The margin pressure is now
coming from memory, so HP, the memory chips that they
need to use for their personal computers are also rising
in price, and so there's a concern for the future
numbers from HP about how they're going to handle that
march an impact on the PC side.
Speaker 2 (28:14):
I mean it as well for Micron, who's earning has
come the week after. But what's interesting more broadly is
have they from a stock perspective, from an investor perspective,
ridden the AI wave? How much have people been looking
to HP more broadly for AI to be the real
winning staff for it?
Speaker 11 (28:30):
So there for them, it's more on the aipcs, so
an increasing percentage of their personal computers are these aipcs
which have a different special chip, not an in video one,
in order to run AI functions natively in the personal computer.
They're also really riding an upgrade wave with Windows ten
(28:52):
going out of support, people are needing to upgrade to
Windows eleven.
Speaker 3 (28:55):
So that's been helping them. But having that comme.
Speaker 11 (28:59):
At a time where they're going to have to potentially
incur higher cost for the memory going into those machines
is a concern.
Speaker 2 (29:06):
Well, you're going to be busy after the earning spell
tonight in most then of us across all things l NHP,
we keep an eye out. Meanwhile, coming up, we go
to the private markets. Sequoia Capital partner Brian Halligan is
going to be with us and how you.
Speaker 3 (29:18):
Model for a desirable founder and CEO that has changed
for venture investors. That's sex the su Blomberg Tech Look.
Speaker 2 (29:32):
We've spent the show, perhaps the last week of shows,
with ourselves and investors questioning in videas valuation and the
prospect of a so called AI bubble in their public markets.
Our next guest says, it's actually a private market issue
if you think about AI bubbles. Brian Halligan is with US,
partner at Sequoia Capital, professor at.
Speaker 3 (29:49):
MIT, co founder of HubSpot.
Speaker 2 (29:51):
That just goes on, Brian, what are you seeing in
the private markets in certain isolated instances that maybe reflect
some anxiety that we're getting in the public markets.
Speaker 12 (30:00):
Well, I'm old enough that I lived through the last bubble, Caroline,
and history doesn't repeat itself. But at rhymes, yeah, and
there's definitely some rhyming going on. Like man, the evaluations
are high, and they're high early. The thing that's different is,
my goodness, is their galactic level growth in these startups.
(30:21):
The demand is amazing and it sort of started at
the model level and then went to infrastructure. The app
level companies are absolutely ripping now and so it's an
interesting time.
Speaker 7 (30:35):
It's different than ninety nine.
Speaker 2 (30:37):
How are you therefore setting up as you're helping CEOs become.
Speaker 3 (30:40):
From startup to scale up.
Speaker 2 (30:41):
Mindset and eleven Labs, for example, which I heard time
and time again getting adopted, even Jensen Wang saying how
much he's loving that particular product. That's a CEO you're
helping navigate. What do you say to them in these moments?
Speaker 12 (30:53):
So if I were a founder and I were warried
it were a bubble, I would do a couple of things.
First thing I would do is in my next round,
I would take a little bit of money off the table.
Speaker 6 (31:03):
Oh.
Speaker 12 (31:04):
The second thing I would do is that would raise
a lot more than a planned because if it is
a bubble and it dips and it eventually comes back,
you want enough to last through.
Speaker 7 (31:12):
Those would be the two places.
Speaker 3 (31:13):
Twenty twenty one mindset A little bit, A little bit.
Speaker 12 (31:15):
Yeah, A lot of those companies twenty twenty on, a
lot of great companies just didn't kind of make it out.
They didn't raise enough, they didn't make it through. And
if you look at ninety nine or the bubble era,
you know, some good companies came out of there. Google
came out of there, Amazon came out of their sales,
short dot Com came out of there. And so even
if the valuations are really inflated, if you can find
some amazing founders, I mean, there's there's going to be
(31:38):
a lot more than three that come out of this one.
Speaker 2 (31:41):
There's a lot more than three companies trying to get
it on each other's space as well. And this is
where kind of marketing comes in. And I want your
brain space as someone who's helped led HubSpot founded it.
But also you've got this great new book out. I'm
no deadhead, but I know many people are. You are
a quintessential dead head and loved all things about the
Grateful Dead. But you think we should look at the
Great Dead as a marketing model as well.
Speaker 3 (32:01):
What is it about community?
Speaker 2 (32:03):
What is it about I'm sure it's not about bootlegging
music that you think is the thing to repeat.
Speaker 12 (32:08):
Absolutely, there's so much founders can learn from Jerry Garcia
in the Grave of Dead. First of all, Garcia was
like the ultimate and original Silicon Valley founder founded in
Palo Alto, built an amazing company. He did a lot
of interesting marketing things that all the founders I coach
are trying to do.
Speaker 7 (32:26):
First thing he.
Speaker 12 (32:26):
Did was he kind of created a whole category around
this thing called jam bands that lots of people following around.
Hard to do. Second thing he did is he didn't
use traditional ways to market his product, like radio stations
and albums. He'd let people come in with all their
equipment and record the concerts in trade tapes.
Speaker 7 (32:42):
He went.
Speaker 12 (32:43):
He was like the first viral marketer in Silicon Valley.
The third thing he did that I think is quite
remarkable is he didn't like that ticket Master in the
scalpers made all the money and inflated the prices for
his customers. So he disintimated those two layers. And so
we're going to start a ticketing company and we're going
to sell tickets directly to customers. We're going to cut
(33:04):
out Ticketmaster and all this scalfer. So he fought in
a very original way. He was He was a radically
first Principles founder. He rhymes a lot with Jensen Hwang,
rhymes a lot with Sam Almon, rams a lot with
Steve Jobs.
Speaker 2 (33:16):
What's interesting, And I'm going to keep going with this
grateful dead analogy because I love it. Disbanded in nineteen
ninety five after the passing of Jerry Garcia, and there's
been different combinations since that, and co we see some
artistic differences. Dare I say there's a few artistic differences
at Sequoia at the moment and invention Will Broady, You've
(33:37):
been at Sequoia for a year, There's been a lot
of change at the top. How are you seeing the
venture community set up for this moment? What can they
learn from grateful dead and from entrepreneurialism at this moment.
Speaker 12 (33:47):
I think Sequoia is particularly well set up at the moment.
The two new leaders were fantastic that have been there
a long time. They have amazing track records, and like
I think of the stack as like the hardware, the labs, infrastrate,
the apps is well positioned with amazing investments across all
of them, particularly at the app level, and particularly you're
in New York City. Last night I had dinner with
(34:07):
the CEO of profound terrific company that does not SEO
but like SEO for Chetchubt and Gemini in the founder
of Rogo, fantastic CEO. Rogo is like AI for investment bankers.
I think this is emble manic of what's going on.
The app layer is starting to pop and Sequoya is
in a lot of these things. Other great companies in
New York basis is selling to accountants. You've got Crosby
(34:31):
in Harvey's selling to lawyers. New York is actually having
a moment in AI, and it's kind of at that
app level.
Speaker 2 (34:37):
The app level is where perhaps the productivity really starts
to rain in. That is what the proof point is
needed many would say for the market, when actually you
and I are not just using it for our own personal life.
But see productivity go up into the right and companies
start doubling down on the purchase.
Speaker 3 (34:52):
Of these applications.
Speaker 2 (34:53):
What does that show up when don't we stop even
talking about an AI bubble because we see the.
Speaker 12 (34:58):
Productivity, Well, there's just giant demand and galactic growth. One
of the interesting things about all these founders is I'm like, well,
you're certainly going to grow with less people, right, and
they say, well, actually no, we're hiring, really hiring aggressively,
and so like this. Some people are like AI is
going to make humans unnecessary. They're like, now, we're going
(35:19):
to make users unstoppable. And that's sort of the mindset
across most of these AI startups. So they're pressing hard,
hiring hard, growing hard, and I think you'll start seeing
over the next couple of years big productivity advantages like
HubSpot uses AI across the enterprise and customers support in
R and D, massive productivity benefits across a couple of
big parts.
Speaker 7 (35:39):
Of the enterprise.
Speaker 2 (35:40):
I mean, we're using juice box, which helps with hiring
in the world of AI as well.
Speaker 12 (35:43):
One of my favorite founders of companies on fire.
Speaker 3 (35:45):
Yes, it's been wonderful having you here. You're on far too.
Speaker 2 (35:49):
Enjoy the rest of your thanksgiving. Grian Halligan in the
house Aquoia Capital partner that.
Speaker 3 (35:53):
We thank him. Meanwhile, coming up, robots housekeepers? Are they
close to a reality?
Speaker 2 (36:00):
To the CEO behind memo the robot trained on and
for your housework, bring bag text. Let's return to our
key story of the day. Shares Alphabet another record high.
They're rising as the company is said to be in
talks potentially with Meta over a deal to provide AI
(36:21):
chips TPUs Google Zone in house chips to Meta for
the future. It's all according to a report by The Information.
Let's go more on this and other trends in AI.
Want not to be using generative haarheng models for this holiday,
Davey Alba, you're with us and just first to the
bread and buster of Alphabet. How much has caught you
off guard? People working at Alphabet off guard that finally
we get the understanding of the vertical model integration.
Speaker 13 (36:45):
You know, I don't know that I was necessarily caught
off guard by that. I think that this has been
creeping up for a while, but it does seem like
the rest of the industry is catching up to this
idea that GPUs have enormous value and are really could
be a really valuable part of you know, people's AI
mixes that you know in video is not the only
(37:08):
game in town when it comes to chips.
Speaker 2 (37:10):
Certainly, we've heard Da Davidson, you've heard Bernstein, You've had
a lot of analysts saying this could be a really
individual way of selling it. And interestingly, now maybe not
just for Google Cloud, but TPUs in and of themselves,
but aside from Alphabet, what they're doing in terms of
the chip stack, their models, how are people going to
be using them this holiday You've got a great story
out about the anxiety perhaps this is going to create
(37:32):
in the kitchen.
Speaker 13 (37:34):
Yeah, we published a story this morning about how food
bloggers are warning consumers about AI recipe slop ahead of Thanksgiving.
We talked to twenty two food bloggers ahead of the
holiday season, and all of them report, you know, traffic
declines and AI Frankenstein recipes that remix their recipes and
(38:00):
pull in bits and pieces of other recipes to create
content that is not accurate. That where if you follow
the actual recipe steps that are generated by these AI models.
You could come out with literal slab, you know, inedible food.
And it's really confusing people these days sort of where
(38:23):
to find quality content on food recipes this this holiday season.
Speaker 2 (38:29):
Maybe stick to the source for now at least. Davey
Alba punning all puns. I thank you. Meanwhile, let's talk
about what else you need helping you in your kitchen.
Maybe it's robots. Well, they've been busy dancing, they've been boxing,
they've been running marathons, so why are they not doing
more of your chores? This is memo from AI startup
Sunday Robotics. The company says it's robot is purpose built
(38:50):
for housework, trained on millions of episodes of everyday household routines.
Sunday's co founder and CEO, Tony Jao joins us, now you're.
Speaker 3 (39:00):
Robot.
Speaker 2 (39:01):
It's kind of humanoid like a little but not totally.
How does it differ from other robotics.
Speaker 14 (39:08):
Yeah, I think we just think about safety as a
really high priority item, and we define it as being
like passively safe. And what it means is that you
can put the robot into any configuration and you can
cut power and the robot will still be stable. So
this is why we build this whole mobile base as
opposed to lex.
Speaker 2 (39:28):
Where did you ultimately come to decide that this was
the way in which you should think about robotics, Maybe
not in a humanoid manner, maybe just with real safety. First,
you've got a stellar background, you're a Google Deep Mind,
Tesla Auto Power at Google X. You're also, of course
just coming out of stealth with a call thirty million
to put to work.
Speaker 7 (39:48):
Yeah.
Speaker 14 (39:48):
I think the the biggest way we think very differently
is actually on how to train these robots, not just
the design, but.
Speaker 7 (39:55):
How it obtained its intelligence.
Speaker 14 (39:58):
So normally people train their robot through tally operation, which
essentially means that you kind of lock into the robot and.
Speaker 7 (40:04):
Control how it moves.
Speaker 14 (40:05):
But the way we learn is actually very very different
that we learn from humans directly. That's essentially we design
these device a glove that captures how human do their chores,
and we're able to transfer those data directly into the robots,
and that's how the robot is able to learn from
like hundreds of humans simultaneously.
Speaker 2 (40:25):
These robots don't come cheap, but interestingly, Tony you're not
looking to sell immediately, you're.
Speaker 3 (40:30):
Looking to beta test.
Speaker 2 (40:32):
Now, how are you finding the right people to bring
these robots into their home?
Speaker 14 (40:38):
Yeah, so if you look at our website, we actually
have a huge signup doc for people who are interested,
and we're ready to get more than a few thousand
of these applications.
Speaker 7 (40:48):
So what we're going to work on.
Speaker 14 (40:49):
Next is to like very carefully sit through all these
applications and find people who what we call like founding families,
who'll be there to give us feedback, will be there
to kind of shape what a product will look like
in the future.
Speaker 3 (41:04):
What do you think the hardest element for these robots.
Speaker 7 (41:06):
Is the hardest element?
Speaker 14 (41:11):
I think it will be how people will react to
this like big robot in their homes. And again, this
is the first time that anyone has put a mobile
manipulator like a robot with arms into real living homes,
and this is something that we're incredibly excited about. I
think will be people will be pleasantly surprised by how
useful it is.
Speaker 3 (41:31):
Why do you.
Speaker 2 (41:31):
Think so many tech companies do end up turning to humanoids,
to turning to the physical form of a human rather
than a more stable basis you have.
Speaker 14 (41:43):
Yeah, I think if you're working on environments that with
a lot of stairs, or you're working on environments with
like you know, like hills, I think having lux will
be helpful in that case and for us in our
first we decide to go for a veal base just
for the simplicity, for lowering costs and to allow us
(42:06):
to move faster.
Speaker 3 (42:07):
Talk to us about costs.
Speaker 2 (42:08):
You have managed to raise seed funding from Sarah Gao
a conviction. You've now got money coming in led by Benchmark.
Speaker 3 (42:15):
What is the key cost for you? Is it the talent?
Is it the hardware? What is it?
Speaker 7 (42:21):
Yeah?
Speaker 14 (42:22):
So our hardware is actually quite differentiated from a lot
of humanoids. Even at quantity zero when we prototype it
these days, it costs around twenty five K to make,
and at quantity around like five thousand, we're able to
get a cost to below ten K. So I think
we'll be ending up selling it around to five to
ten K in the final price. And this is we're thinking.
Speaker 7 (42:44):
About robots, not like another car.
Speaker 14 (42:46):
Like purchase, but more like a fancy smartphone or a laptop.
Speaker 2 (42:51):
How does American ingenuity when it comes to robotics stack
up to that of Asia in China and how you
seeing your own supply chain develop Yeah.
Speaker 14 (43:00):
This is a great question. I think American has incredible
mechanical engineers, software engineers, but we are lacking in terms
of some of the supply chain infrastructures. So I think
we're at the point that we need to leverage some
of the growing supply chains the humanoids in China and
(43:22):
we actually share a lot of components with them so
that we can have the economy of scale before us
shipping like millions of robots.
Speaker 2 (43:30):
Any Ja, CEO of robotic startup Sunday, fascinating to have
you on.
Speaker 3 (43:35):
Thank you very much.
Speaker 2 (43:35):
Indeed, that does it for this edition at Blueberg Tech,
we do want to remind you of the market moves
today in video under significant pressure off of its lows,
still down by four percent. As we question is dominance
in the world of chips to train as.
Speaker 3 (43:49):
Well as use your models.
Speaker 2 (43:50):
That competition coming maybe from Alphabet TPUs, maybe Meta Eyeing
buying some for its data centers of the future as
Information is currently reporting Oracle once again off by one
point nine percent from New York.
Speaker 3 (44:02):
This is Bloomberg Tech. Don't forget to check out the
podcast
Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live
from coast to coast with Caroline Hide in New York
and Eva Loow in San Francisco.
Speaker 2 (00:22):
This is Bloomberg Tech coming up, Meta Iron, Google's AI chips.
According to reports, how the market is questioning in Vidia's
long term dominance. Prashujinping brings the issue of sovereignty over
chiphub Taiwan back on the agenda with Donald Trump during
their phone call. And Apple eliminates sales roles in a
rare layoff to streamline the way it offers products to
(00:42):
businesses and governments.
Speaker 3 (00:44):
We have the details.
Speaker 2 (00:45):
The first, let's check in on these markets that are
trying to digest a wall of delayed data at the
same time as baking in maybe an eighty percent chance
of a FED rate cut. That's what the market price
is in but all eyes on tussle at the top
in AI dominance, looking at a five ten percent drop
on the Nasdaq one hundred on the upside, some key
players on the drownd side, we're seeing a key drag
(01:07):
from some of the key chip names.
Speaker 3 (01:08):
In particular.
Speaker 2 (01:09):
Let's delve into what's moving this particular index, which does
lag the SMP and the DOW. Today we're seeing in
video off by five percent. No wonder then, ASAK one
hundred is underwater when the world's most dominant player from
a market cap perspective is currently off by five percent.
Big moves for a four trillion dollar company, and in
large part it's because of Alphabet that actually is trimming
its games but still at a new record high. This
(01:30):
is we understand meta eyeing tpused. This is all about
the dominance of where we get the chips of the
future to train our models as well as use our models.
Meta is the key.
Speaker 3 (01:41):
Line of questioning.
Speaker 2 (01:42):
Here is meta eyeing tpused from Google to put in
its own data centers. They're the reports coming from the
information today, let's talk about the market reaction. Equity reporter
around Nostellica is here with us Ryan. It's monumental move
that Alphabet has had since mid October. It's added a
trillion in market capitalization. Were starting to become aware of
its prowess in AI and in chips.
Speaker 4 (02:05):
Yeah, absolutely, thanks for having me so. Yeah, certainly sentiment
has really reversed on Alphabet, and I think people are
really appreciating how dominant. It is across every layer of
the AI stack. Of course, it recently released the latest
update to Gemini, which was seen as very strong across
all the major benchmarks that people used to evaluate AI models.
(02:26):
The chip business is getting a lot of attention lately,
between the report with Meta and the deal with Anthropic.
People really see this as a really significant potential competitor
to Nvidia, as you were discussing earlier, and beyond that,
it has a huge cloud business that is seeing accelerating growth.
It has so much data and users and distribution and talent.
(02:48):
It really has all of the pieces, and that has
really helped the stock surge, not just recently but over
the past several months. It is by far the biggest
performer or the best performer of the mag seven this year.
Speaker 2 (03:00):
I think Counterpoint Research analyst and co founder Nail shar
and a story on Blueberg saying it's a sleeping giant
in the AI race and is fully awoken. But many
have been trading for the last few days ever since
Gemini three release.
Speaker 3 (03:13):
Maybe going short open.
Speaker 2 (03:15):
AI suppliers and long alphabet suppliers. Is that still bearing out.
I'm looking at Oracle down once again.
Speaker 4 (03:22):
Ryan Oracle has really been under a lot of pressure lately.
I think it's on track for its biggest one month
dropped since two thousand and one, so certainly a real
reversal there. It does seem like people are moving towards
the Google and Alphabet suppliers, which is companies like Broadcom,
while at the same time the companies that are more
connected to open Ai. You mentioned Oracle, also AMD, Microsoft
(03:44):
to a certain extent, these companies have really been under
a lot of pressure lately.
Speaker 2 (03:48):
It's worth reminding though that yes, Alphabet's been on a tear,
and actually its valuations have started to trade way higher
than we're used to I thinking about twenty six twenty
seven times future earnings, but still leap from just mysoft.
Speaker 3 (04:00):
Ahead of it in terms of market cap is Apple.
Speaker 2 (04:03):
I think it's got another seven percent to go to
hit that, but fifteen percent to go to in video.
We're not questioning longer term at this exact moment. In
Vidia's dominance in AI and semiconductors.
Speaker 4 (04:14):
Well, NBA right now has I think ninety percent or
so share of the data center market. If there is
potential for Alphabet to start eating into that, I think
that would change a lot of calculations for people in
terms of how much can Alphabet scale this business. I
have talked to someone who thinks that this could potentially
be worth more than Alphabet's cloud business, potentially up to
nine hundred billion dollars. So that is a huge potential
(04:36):
market there. And if that means that Nvidia starts losing
market share, I think people will start reassessing how to
value that company and its growth and its valuation. Now,
I will just simply say that the AI market is
growing so rapidly that people do see room for a
lot of big players, and even Alphabet remains a major
customer to in Vidia just because there is so much
(04:58):
demand for compute right now, people don't want to be
behold into any single supplier. It does suggest that there's
a lot of room for growth to go around, even
if we start seeing some erosion and market share at Navidia.
Speaker 2 (05:10):
And I'm pretty sure Jenson Wang will be responding to
any concerns about erosion and market share round a SELCA.
Speaker 3 (05:16):
We so appreciate you joining.
Speaker 2 (05:18):
Let's dig in to further analysis here with Stephanie Anagai
to the conversation which has global market strategist Japing Morgan
Asset Management and has four trillion dollars in assets under management,
four trillion, rather similar to the market capitalizations in some
of these companies, And I'm interested, Stephanie, does it matter
to you from a market sentiment perspective if there's scribbling
(05:38):
at the top of the US domiciled companies.
Speaker 3 (05:42):
Is it a.
Speaker 2 (05:42):
Worry that we'll see perhaps in Vidia being questioned in
terms of its dominance.
Speaker 3 (05:46):
I think it's quite healthy.
Speaker 5 (05:48):
We've seen the AI trade has delivered enormous returns for
markets over the last two years, and we're I think
all kind of experiencing the sire of relief with this exhale.
I guess in a way, we've moved from a rise
tied lifting all boats or more choppier waters, and investors
are being far more scrutinizing when it comes to how
much is being spent, the quality of those investments. We're
(06:10):
seeing as this AI trade continues to grow in its enormity,
the investment being made, and the moat in such big
questions and shifting. So I think it's quite healthy that
we're focusing on selectivity. I mean, this is what you
want to see to prevent a dot com bubble, and
we're looking at valuations today, we don't really see a
big risk of that happening amongst the big tech firms.
Speaker 2 (06:32):
So even these worries about debt in particular, and you
are coming to us with a viewpoint that is cross asset.
In many ways, we have seen a real desire to
get into AI related debts. Some of the bond sales
that come from the likes of Alphabet and the likes
of Meta and an Oracle.
Speaker 3 (06:46):
Have been scooped up.
Speaker 2 (06:48):
But there's been this worry that in the longer term
it might start to maybe pull back on overall demand.
We might see some of these big hyper scales coming
to market so often that it drives up prices for others.
Speaker 3 (07:00):
It's possible.
Speaker 5 (07:00):
I mean, I think first just looking at the magnitude
of how much AI investment is needed or being spent
already by any way that you cut it, the amount
of spending right now is enormous. But just like Cardia,
isn't that expensive for a billionaire. When looking at these
capex relative to the sales from these companies, relative to
their current revenue growth, which has also grown significantly, it's
(07:22):
actually not that extreme. So we're seeing this increasing move to.
Speaker 3 (07:25):
Tap into debt markets.
Speaker 5 (07:27):
But for us, it's not so much these companies getting overextended,
but actually more so a reflection of a better sorry,
a better capital structure. You know, there's some investments like
these data centers that are going to be invested over
multiple years. It might make more sense to tap debt
markets for some of these deals or an off balancing structure.
Speaker 2 (07:47):
So for example, we're going to be looking not only
at AI investment surging, but at credit default swaps of
Oracle surging. Has that just been acting as a bell
weather and the necessary bell weather to stop reflecting some
of the risks that maybe people been ignoring for the.
Speaker 3 (08:00):
Past few months.
Speaker 5 (08:01):
I think it's it's very apt, and not all of
these companies have the establishments, you know, they differ in
many different ways and also in their sources of revenue.
And I think it makes sense that you know, Oracle
is one of the more you know, risk your companies
that is tapping these bum markets and you're seeing that
being reflective in CDs spreads. But that also can't be
extrapolated to the entire shift right now towards debt markets
(08:25):
to help finance these data center bills, and I'll also add, look,
when it comes to cloud services, that business model is
one of the most cash generative business models.
Speaker 3 (08:35):
In the world.
Speaker 5 (08:35):
So at the end of the day, these bonds are
also being tied to services business operations that have tended
to do quite well for these companies.
Speaker 2 (08:43):
It's interesting, of course, deciding where the margin a cruise.
We're going to have l HP after the bell, Many
feeding that margin is being eroded because of the cost
of memory. Meanwhile, we'll getting Micron next week with its
Sennings and Many anticipating they're strong because.
Speaker 3 (08:55):
Of the memory demand there.
Speaker 2 (08:57):
From your perspective, is there still room to run in
just the tech trade more broadly, or has that shift
into more value names and certainly with the context of
the FED change things longer term into the end of
the year.
Speaker 5 (09:09):
We still think we're quite early in this AI wave,
but we've seen a chapter or two, and moving forward,
I think the focus is not going to only be
on compute needs and capacity needs, but also on AI
utilization and what companies are really critical for that, whether
it's in software, what companies are leading the way in
financials and entertainment, in adopting AI, and then also how
(09:33):
once we learn more about the end user demand for
AI and the pricing power of these AI services, that's
going to give us a lot of clarity around the
ROI around these AI investments.
Speaker 3 (09:43):
So is that what we need?
Speaker 2 (09:44):
Is it ultimately the revenues of companies outside of the
world of tech to vindicate that? What pushes us higher
in terms of real contextview? Is it December when we
get the FED decision? What is the catalyst do you
think for us to reassess where we are in valuation?
Speaker 5 (10:01):
I'd say it's less of the kind of macro back
job early here and much more of the proof point
around the monetization of AI. I think the more that
you see businesses ramping up their IT budgets, the stickiness
that you see in those investment spending and then also
AI delivering and we've seen some proof cases of that
so far. Coding has been a huge factor of that.
(10:23):
But once you see more companies, particularly outside of tech
maybe tech adjacent, coming at their earnings calls and talking
about their AI generated savings, I think that's going to
be a really important next lever for the AI trade.
Speaker 2 (10:36):
And then one about the Leva FERRATUALI the companies that
are adulting the clients that are cooling you on a
daily basis, are they saying do I double down more
in tech? Are they saying I need to double out
outside of the world of tech.
Speaker 5 (10:47):
I think it's it's diversifying that tech exposure. You know,
after a long run in these AI names, you don't
want all your ex in one basket because it is.
Speaker 3 (10:57):
Probably not.
Speaker 5 (10:57):
At least you want to right size some of that
exposure bill to top of all of those games that
we've experienced in position for how this AI wave is
going to evolve, there will undoubtedly be losers and winners,
but we also don't want to be out of the market.
Speaker 3 (11:10):
And that's another thing that we're trying.
Speaker 5 (11:11):
To talk to clients about because even when you call
a bubble correctly, if you weren't in the market from
nineteen ninety five to nineteen ninety nine, you would have
missed out on over four hundred percent in total return
in the Nasdaq. You were right, but you've locked in
years of underperformance. So when it comes to us f
Wing Marcus today, we don't see that real risk of
a systemic bubble, but we do see a real opportunity
(11:32):
to just make sure that portfolios are built for resiliency
and they're also built to take advantage of how this
ai wave continues.
Speaker 2 (11:39):
To evolve, well, hopefully keep having you on a zimbubble
or indeed the narrative does evolve. We so appreciate Stephanie
Aliaga JPM Wollngan asset management can cross tech for us.
Mean while coming out with China, Shijin paying revives talks
to the sovereignty of a Taiwan and a phone call
with President Trump.
Speaker 3 (11:56):
More on that next, This is boom Bag Tech.
Speaker 2 (12:10):
Chinese President Xijingping well has revived the topic of China's
sovereignty over Taiwan in a phone call with President Trump yesterday,
discussion that didn't come up during their face to face
meeting last month in Beijing. From Magxinia, tech editor Mike
Sheppard joined us for the latest and Mike remind us
from a tech perspective, Taiwan, we know its dominance and chips,
(12:31):
We understand its integral nature to the tech ecosystem. What
is happening between Xijiping and Trump on.
Speaker 6 (12:37):
This Well, what was interesting yesterday, Carr is that we
get two very different versions of this phone call between
the leaders of the world's two largest economies. The first
version came from Beijing. The official state news agency SHINOA
put out its version of the conversation, presenting it really
as one centered on the question of Taiwan. And then
(12:59):
a few hours later we heard from President Donald Trump
himself on truth social posting that they had had a
great conversation about issues including soybeans and other matters, rare earths,
and other key topics that were deared to the US president,
but he made no mention of Taiwan there Now, while
he did not bring it up, several hours later, he
(13:20):
did call the new Prime Minister of Japan, Sanai Takaichi,
who had enraged Beijing with comments a few weeks ago.
You'll remember Caro saying that Japan would consider leaping to
Taiwan's defense in the event that China.
Speaker 7 (13:35):
Were to try to take it.
Speaker 6 (13:36):
Now, all of this is huge implications for the supply
chain of semiconductors, as you noted, especially Taiwan semiconductor. It's
one of the world's largest producers of AI chips, and
they are moving some of their production, as we know,
to the US they've pledged one hundred and sixty five
billion dollars in investment implants here in the United States,
(13:57):
but they would still retain a significant amount of their
capacity on the island. Therefore, any question of Taiwan really
does bring up tech issues for US.
Speaker 2 (14:06):
And talking of tech issues, the administration once again trying
to signal its commitment seeing the ai infrastructure build out
akin to the Apollo Mission or to the Manhattan Project, Mike.
Speaker 6 (14:17):
Now they are talking about a Manhattan Project like effort
in this Executive Order to call Genesis that President Donald
Trump signed yesterday, But it was really more a call
to action for various agencies to start working together more closely,
and that includes the Department of Energy and its National
Research laboratories. But when we talk about Manhattan Project, though,
(14:39):
we do need to remember that that effort took thirty
six billion dollars in real dollars today from back then
in the nineteen forties, as the United States was in
the race to develop a nuclear weapon ahead of the
Access Powers. This is a very different time. We are
not seeing new money being pledged towards this effort. It's
(15:00):
important to remember, Caro, though, that the US already has
put a significant investment in production of chips that would
be needed for artificial intelligence, and that is the Chips
and Science Act of twenty twenty two that put tens
of billions of dollars in loans and grants and other support,
including tax incentives, to support the development of a chip
(15:22):
industry domestically that would help artificial intelligence take hold and
gain ground and lead the world, as President Biden and
President Donald Trump now say they would like the US
to do.
Speaker 7 (15:34):
Now, we are.
Speaker 6 (15:34):
Seeing companies like open ai push for further investment to
support data centers, and that would include extending some of
those tax credits to data centers. So it'll be interesting
to see how this develops and whether more of those
tax credits will go to some of those AI projects.
Just Caro, as we are wrestling with those questions of
whether we are seeing too much money going into this space.
Speaker 2 (15:56):
Max make Shepard all the context from Washington, We appreciate it.
Speaker 3 (16:00):
Meanwhile, it's time for talking tech now.
Speaker 2 (16:02):
First up, Ali Baba reported thirty four percent growth in
its cloud unit during the September quarter. Despite the gain,
though spending on consumer subsidies data centers that's eaten into
its profits the company's ADRs, as you can currently see,
just training.
Speaker 3 (16:15):
Off by some two percent day.
Speaker 2 (16:17):
Meanwhile, Anthropic it's got a new model, claud Opens four
point five that the company says it's better at coding
and office tasks such as financial analysis and creating presentations
or spreadsheets.
Speaker 3 (16:27):
It's part of anthropics efforts.
Speaker 2 (16:29):
To compete with open Ai with Google business customers. And Meanwhile,
open ais a new tool to generate personalized shopping guides.
The company trained a version of GBT five Mini.
Speaker 3 (16:40):
Model to ask follow up questions.
Speaker 2 (16:42):
Draw answers from reviews published on what the company considers
higher quality websites. Now, let's talk Apple, because, in a
rare mood for the company, the tech giant has eliminated
dozens of sales roles in an effort to streamline the
way it offers its products to businesses, schools, and governments.
For more on this breaking story, Mark German joins us,
(17:03):
it's not one hundreds, We're.
Speaker 3 (17:04):
Talking tens of people.
Speaker 2 (17:06):
But still it's notable because we don't often see layofs
of that one.
Speaker 8 (17:10):
Yeah, to your point, it was several dozen people across
Apple sales division, and this sales division they partner with
carriers across the world to sell iPhones, but they also
partner with enterprises, large scale businesses, government organizations, schools, educational institutions,
major universities across the world to sell products like iPhones, iPads,
(17:33):
max and you name it. And over the course of
this month, there was a big streamlining rounds of layoffs, including,
like I said, several dozen people. There were account managers,
they are called account executives for specific government agencies, for
specific university systems, people who partner in pitch companies on
(17:53):
buying Apple products. There are these tiny Apple store like
fixtures called briefing centers at Apple offices in California and Texas,
and the people managing that, for the most part, we're
laid off as well. And so this is going to
change how Apple sells products to these different organizations. The
majority of products are bought through what's called the channel,
(18:15):
so third party retailers, and so those products are still
going to sell, but quite a bit of a shake
up here for Apple sells products and delivers these devices
to the major customers. And of course, as you said,
a rare layoff for Apple.
Speaker 2 (18:30):
Apple did respond to your reporting and saying we're continuing
to hire, and those employees can.
Speaker 3 (18:35):
Apply for new roles.
Speaker 2 (18:36):
Mark, But what do you think this signifies more broadly
about how Apple is trying to streamline, trying to become
more efficient, trying to ensure that it doesn't seem like
it's a lagged in this age of AI.
Speaker 8 (18:48):
Yeah, you know, I don't think this has much to
do with AI. I think this has to do with
cutting roles internally to lower costs because they realize most
of these sales are happening from the channel and there's
a lot of duplicate efforts internally with the channel, so
the third party retailers. So I think it's just one
of your classic layoffs to create more efficiency and cost cutting,
(19:10):
rather than having much to do with artificial intelligence. In
terms of layoffs related to AI at Apple, I guess
the only thing you've seen related to AI from Apple
that has to touch a layoff was the self Drenking
Car project job cuts of one thousand people in the
beginning of twenty twenty four, and that was actually to
do more AI rather than because of AI. So they
(19:32):
moved a lot of those folks over to the Generative
AI division. But I haven't seen any job cuts at
Apple to date because of AI. That doesn't mean they're
not going to happen, but so far they haven't.
Speaker 3 (19:43):
Well that's far.
Speaker 2 (19:44):
We started to still see some job cuts across technology
and Mark Gurman, you've been at the front of that reporting.
Speaker 3 (19:49):
We really appreciate it.
Speaker 2 (19:51):
Now coming up, Michael Barry stands by his in video
criticisms as after calling out the company for stop by
gap facts for compensation dilusion, but in video itself responded
to analysts.
Speaker 3 (20:02):
More on that next. As a Bloomberg Tech.
Speaker 2 (20:15):
As we've been reporting in video shares, they are under
pressure today, Competition fierce when it.
Speaker 3 (20:19):
Comes from Alphabet and TPUs.
Speaker 2 (20:21):
We understand reporting that Meta is eyeing potentially turning to
Google for its chips in its data centers in the future.
But there's also Michael Burry there isn't there standing by
his criticism of the company after a video pushed back
on his earlier analysis of stock based compensation of share
buybacks for more Bluemberg equity supporter com and Ryanikey reminds
us of what the Cassandra as he dubs himself, has
(20:44):
been saying. Michael Burry laid on issues about the circularity
of tech deals, worries about the interoperability of what big
tech are currently doing in the world of generator AI.
What did he take issue though, within video when it
comes to share buybacks and stock compensation.
Speaker 9 (21:00):
Yeah, so, I think most basically his argument is that
the amount of stock based compensation is diluting you know,
owner's power, that if you hold the stock, it's being
diluted by the stock based compensation. And so it really
is just another thing and sort of a myriad of
things that he has called out within Vidia in recent weeks.
And yeah, we saw the company, you know, push back.
(21:22):
There was a memo that they sent to some Wall
Street analysis is according to a Baron's report that said,
you know, we think his math is wrong. Kind of
explained the situation a little bit better and also very
latantly stated, you know, we're not Enron. We're not you know,
there's no fraud here. But but Bury really said, you know,
I stand by my analysis that you know about the
(21:45):
stock based compensation, dilution share buybacks. And he also said,
you know, I'm not comparing Nvidia to Enron, I'm comparing
it too Cisco, which I thought was really interesting thinking
about Cisco, you know, in the dot com era, it
had this huge run up, but it was really associated
with the over build in fiber optic cable. So he's
you know, comparing that to what's happening now, I guess
(22:07):
with the video with data centers, and these are really
some of the biggest you know, concerns or pain points
that we're seeing in this debate over if AI is
a bubble, and you know, in video shares are down.
I think they were down as much as six percent today.
We're seeing you know, more than two hundred billion dollars
in market value just wiped off, and we're also watching
the sort of twenty percent level and videas nearing twenty
(22:28):
percent draw down from it's high at the end of October,
which was a significant level.
Speaker 3 (22:33):
For the shares technical band market. Extraordinary. Thank you very much,
comment and Rhanick.
Speaker 2 (22:37):
He always has some of the most read stories across
all of the technology moves.
Speaker 3 (22:41):
You've got to keep them up to date with it.
Speaker 2 (22:42):
Meanwhile, coming up Google's potential chip deal with Meta's raising
questions about in Vidia's dominance and the erase for AI leadership.
Speaker 3 (22:49):
More on that take next. This is boom Bag Tech
Welcome back.
Speaker 2 (23:03):
To Bloomberg Tech, let's take a check on these markets,
because we have seen some sell off continuing.
Speaker 3 (23:08):
In the world of technology.
Speaker 2 (23:09):
Unlike SMP, unlike the Dow, the Moon music remains resolutely
in the red.
Speaker 3 (23:13):
We're off by four tenths of percent.
Speaker 2 (23:15):
We're seeing some of the big tech names, namely in video.
Speaker 3 (23:18):
On the downside.
Speaker 2 (23:19):
We're still questioning valuations as we get that myriad of
data that comes late to the party when it comes
to certainly our own consumer sentiment seems to be on
the low side. But we're seeing retail sales maybe pointing
towards whether or not we've got some resiliency in the
overall macroeconomic picture. But in Vidia is more a story
of its resiliency versus competition.
Speaker 3 (23:37):
We're down by four percent.
Speaker 2 (23:38):
Once again, we're wondering if other chips will be created
by other players, like Alphabet for example. It's TPUs maybe
being eyed by Meta. That story we're going to delve into.
We're seeing both shares trade higher. Oracle and any name
really in the open Ai ecosystem has been a typical
short for the last week or so, as people question
open AI's dominance versus Gemini. Three for example, so Oracle
(24:00):
once again off by another two percent. Let's really dig
in them into the story of the day of alphabet
really giving m video a run for its money, certainly
a market capitalization front at least Mandy saying, you're here,
senior techanalyst a roomag intelligence.
Speaker 3 (24:13):
You have for months, if not years, been reminding me and.
Speaker 2 (24:17):
Our viewers of the power of the ecosystem of Google
and TPUs.
Speaker 3 (24:21):
Why now are we only just getting it?
Speaker 7 (24:24):
Wow?
Speaker 10 (24:25):
Because Gemini three showed that you could use the TPUs
both for training the model and for inferencing. I mean,
so far the story was all these secondary providers could
be used for inferencing. The fact that TPUs were used
for training Gemini three and most likely for Entropics Frontier
model as well. So two of your three frontier models
(24:47):
are using TPUs, And I think there is an acknowledgement
now from the market that TPUs are comparable to in
video in terms of functionality, and they are a lot cheaper,
which is why I'm not surprised to see Meta do that.
I mean, they will raise CAPEX and you know, going
to the secondary provider who is a much cheaper option
(25:09):
makes sense. And I think a market will like it
when they raise the capex and say that we'll be
using Google as a secondary provider.
Speaker 2 (25:17):
These are thus just reports as it stands, Mandy.
Speaker 3 (25:21):
But what's interesting has.
Speaker 2 (25:22):
Been looking at Jensen Wang's reaction when Alphabet or others
have made inroads into some of their key clients, Anthropic
for example, getting a load of gtpus from Google, and
then we see more of a deal done between in
Video and Anthropic.
Speaker 3 (25:38):
We know that in Videos double down on open AI with.
Speaker 2 (25:41):
One hundred billion dollars being offered in return for GPUs
being considered in their future training. So what do you
think the response mechanism could be of Nvidia.
Speaker 10 (25:50):
I mean, right now, in Video's problem is no one
wants to pay the you know, the high costs they
have for their chips, and that gets reflected in in
Video's margin. Seventy five percent gross margin is something we
have never seen with a semiconductor company. So from that
perspective that you know, the providers who are doing inferencing
(26:10):
are offering their products below costs. Even an open AI
when it's deploying you know it's chatbot at scale, I
would argue, you know, their gross margins are negative because
they're offering their product at below their costs. In the
case of Google, they're deploying you know, generative AI at
scale on search and across their family of apps, and
(26:32):
they're able to do it without really hurting their margins
because their cost phase is much lower. They're running their
infrastructure a lot more efficiently, and so that is where
the problem lies is you can't just keep subsidizing the
inferencing because the cost of your chips is so high.
With Nvidia, and you've got to find a way to
bring down the cost. Everyone wants more inferencing. You just
(26:55):
have to bring down the cost so that you know
they can do it profitaboud.
Speaker 2 (26:59):
Well, see how the response does indeed turn out, and
how they compare the software offerings as well as the hardware.
Man keep seeing a BlueBag intelligence always across the story.
Speaker 3 (27:08):
We so appreciate it.
Speaker 2 (27:09):
More insight into AI and indeed margins is set to
come after the Bell, Dell HP they report Bluebag's DNA.
Speaker 3 (27:16):
Bass gives us the preview.
Speaker 2 (27:17):
We're just hearing about the very healthy margins that Nvidia has,
and in many ways it's because the margins of Dell
and HP and server offerers are much thinner.
Speaker 3 (27:27):
Sure, Yeah, Dell.
Speaker 11 (27:29):
The most watched part of Dell's business for the last
couple of quarters has been its AI server.
Speaker 3 (27:33):
Business, many most all running in media.
Speaker 11 (27:37):
GPUs, and they have some a really marque list of customers.
Speaker 3 (27:41):
There's corewe there's XCAI.
Speaker 11 (27:43):
We reported the other week there they just got a
deal for the first Armenian AI data center.
Speaker 3 (27:48):
The problem is, in order.
Speaker 11 (27:49):
To win some of those deals and to execute on them,
Dell has.
Speaker 3 (27:53):
Had to put up with some pretty.
Speaker 11 (27:55):
Narrow margins over at eachp. The margin pressure is now
coming from memory, so HP, the memory chips that they
need to use for their personal computers are also rising
in price, and so there's a concern for the future
numbers from HP about how they're going to handle that
march an impact on the PC side.
Speaker 2 (28:14):
I mean it as well for Micron, who's earning has
come the week after. But what's interesting more broadly is
have they from a stock perspective, from an investor perspective,
ridden the AI wave? How much have people been looking
to HP more broadly for AI to be the real
winning staff for it?
Speaker 11 (28:30):
So there for them, it's more on the aipcs, so
an increasing percentage of their personal computers are these aipcs
which have a different special chip, not an in video one,
in order to run AI functions natively in the personal computer.
They're also really riding an upgrade wave with Windows ten
(28:52):
going out of support, people are needing to upgrade to
Windows eleven.
Speaker 3 (28:55):
So that's been helping them. But having that comme.
Speaker 11 (28:59):
At a time where they're going to have to potentially
incur higher cost for the memory going into those machines
is a concern.
Speaker 2 (29:06):
Well, you're going to be busy after the earning spell
tonight in most then of us across all things l NHP,
we keep an eye out. Meanwhile, coming up, we go
to the private markets. Sequoia Capital partner Brian Halligan is
going to be with us and how you.
Speaker 3 (29:18):
Model for a desirable founder and CEO that has changed
for venture investors. That's sex the su Blomberg Tech Look.
Speaker 2 (29:32):
We've spent the show, perhaps the last week of shows,
with ourselves and investors questioning in videas valuation and the
prospect of a so called AI bubble in their public markets.
Our next guest says, it's actually a private market issue
if you think about AI bubbles. Brian Halligan is with US,
partner at Sequoia Capital, professor at.
Speaker 3 (29:49):
MIT, co founder of HubSpot.
Speaker 2 (29:51):
That just goes on, Brian, what are you seeing in
the private markets in certain isolated instances that maybe reflect
some anxiety that we're getting in the public markets.
Speaker 12 (30:00):
Well, I'm old enough that I lived through the last bubble, Caroline,
and history doesn't repeat itself. But at rhymes, yeah, and
there's definitely some rhyming going on. Like man, the evaluations
are high, and they're high early. The thing that's different is,
my goodness, is their galactic level growth in these startups.
(30:21):
The demand is amazing and it sort of started at
the model level and then went to infrastructure. The app
level companies are absolutely ripping now and so it's an
interesting time.
Speaker 7 (30:35):
It's different than ninety nine.
Speaker 2 (30:37):
How are you therefore setting up as you're helping CEOs become.
Speaker 3 (30:40):
From startup to scale up.
Speaker 2 (30:41):
Mindset and eleven Labs, for example, which I heard time
and time again getting adopted, even Jensen Wang saying how
much he's loving that particular product. That's a CEO you're
helping navigate. What do you say to them in these moments?
Speaker 12 (30:53):
So if I were a founder and I were warried
it were a bubble, I would do a couple of things.
First thing I would do is in my next round,
I would take a little bit of money off the table.
Speaker 6 (31:03):
Oh.
Speaker 12 (31:04):
The second thing I would do is that would raise
a lot more than a planned because if it is
a bubble and it dips and it eventually comes back,
you want enough to last through.
Speaker 7 (31:12):
Those would be the two places.
Speaker 3 (31:13):
Twenty twenty one mindset A little bit, A little bit.
Speaker 12 (31:15):
Yeah, A lot of those companies twenty twenty on, a
lot of great companies just didn't kind of make it out.
They didn't raise enough, they didn't make it through. And
if you look at ninety nine or the bubble era,
you know, some good companies came out of there. Google
came out of there, Amazon came out of their sales,
short dot Com came out of there. And so even
if the valuations are really inflated, if you can find
some amazing founders, I mean, there's there's going to be
(31:38):
a lot more than three that come out of this one.
Speaker 2 (31:41):
There's a lot more than three companies trying to get
it on each other's space as well. And this is
where kind of marketing comes in. And I want your
brain space as someone who's helped led HubSpot founded it.
But also you've got this great new book out. I'm
no deadhead, but I know many people are. You are
a quintessential dead head and loved all things about the
Grateful Dead. But you think we should look at the
Great Dead as a marketing model as well.
Speaker 3 (32:01):
What is it about community?
Speaker 2 (32:03):
What is it about I'm sure it's not about bootlegging
music that you think is the thing to repeat.
Speaker 12 (32:08):
Absolutely, there's so much founders can learn from Jerry Garcia
in the Grave of Dead. First of all, Garcia was
like the ultimate and original Silicon Valley founder founded in
Palo Alto, built an amazing company. He did a lot
of interesting marketing things that all the founders I coach
are trying to do.
Speaker 7 (32:26):
First thing he.
Speaker 12 (32:26):
Did was he kind of created a whole category around
this thing called jam bands that lots of people following around.
Hard to do. Second thing he did is he didn't
use traditional ways to market his product, like radio stations
and albums. He'd let people come in with all their
equipment and record the concerts in trade tapes.
Speaker 7 (32:42):
He went.
Speaker 12 (32:43):
He was like the first viral marketer in Silicon Valley.
The third thing he did that I think is quite
remarkable is he didn't like that ticket Master in the
scalpers made all the money and inflated the prices for
his customers. So he disintimated those two layers. And so
we're going to start a ticketing company and we're going
to sell tickets directly to customers. We're going to cut
(33:04):
out Ticketmaster and all this scalfer. So he fought in
a very original way. He was He was a radically
first Principles founder. He rhymes a lot with Jensen Hwang,
rhymes a lot with Sam Almon, rams a lot with
Steve Jobs.
Speaker 2 (33:16):
What's interesting, And I'm going to keep going with this
grateful dead analogy because I love it. Disbanded in nineteen
ninety five after the passing of Jerry Garcia, and there's
been different combinations since that, and co we see some
artistic differences. Dare I say there's a few artistic differences
at Sequoia at the moment and invention Will Broady, You've
(33:37):
been at Sequoia for a year, There's been a lot
of change at the top. How are you seeing the
venture community set up for this moment? What can they
learn from grateful dead and from entrepreneurialism at this moment.
Speaker 12 (33:47):
I think Sequoia is particularly well set up at the moment.
The two new leaders were fantastic that have been there
a long time. They have amazing track records, and like
I think of the stack as like the hardware, the labs, infrastrate,
the apps is well positioned with amazing investments across all
of them, particularly at the app level, and particularly you're
in New York City. Last night I had dinner with
(34:07):
the CEO of profound terrific company that does not SEO
but like SEO for Chetchubt and Gemini in the founder
of Rogo, fantastic CEO. Rogo is like AI for investment bankers.
I think this is emble manic of what's going on.
The app layer is starting to pop and Sequoya is
in a lot of these things. Other great companies in
New York basis is selling to accountants. You've got Crosby
(34:31):
in Harvey's selling to lawyers. New York is actually having
a moment in AI, and it's kind of at that
app level.
Speaker 2 (34:37):
The app level is where perhaps the productivity really starts
to rain in. That is what the proof point is
needed many would say for the market, when actually you
and I are not just using it for our own personal life.
But see productivity go up into the right and companies
start doubling down on the purchase.
Speaker 3 (34:52):
Of these applications.
Speaker 2 (34:53):
What does that show up when don't we stop even
talking about an AI bubble because we see the.
Speaker 12 (34:58):
Productivity, Well, there's just giant demand and galactic growth. One
of the interesting things about all these founders is I'm like, well,
you're certainly going to grow with less people, right, and
they say, well, actually no, we're hiring, really hiring aggressively,
and so like this. Some people are like AI is
going to make humans unnecessary. They're like, now, we're going
(35:19):
to make users unstoppable. And that's sort of the mindset
across most of these AI startups. So they're pressing hard,
hiring hard, growing hard, and I think you'll start seeing
over the next couple of years big productivity advantages like
HubSpot uses AI across the enterprise and customers support in
R and D, massive productivity benefits across a couple of
big parts.
Speaker 7 (35:39):
Of the enterprise.
Speaker 2 (35:40):
I mean, we're using juice box, which helps with hiring
in the world of AI as well.
Speaker 12 (35:43):
One of my favorite founders of companies on fire.
Speaker 3 (35:45):
Yes, it's been wonderful having you here. You're on far too.
Speaker 2 (35:49):
Enjoy the rest of your thanksgiving. Grian Halligan in the
house Aquoia Capital partner that.
Speaker 3 (35:53):
We thank him. Meanwhile, coming up, robots housekeepers? Are they
close to a reality?
Speaker 2 (36:00):
To the CEO behind memo the robot trained on and
for your housework, bring bag text. Let's return to our
key story of the day. Shares Alphabet another record high.
They're rising as the company is said to be in
talks potentially with Meta over a deal to provide AI
(36:21):
chips TPUs Google Zone in house chips to Meta for
the future. It's all according to a report by The Information.
Let's go more on this and other trends in AI.
Want not to be using generative haarheng models for this holiday,
Davey Alba, you're with us and just first to the
bread and buster of Alphabet. How much has caught you
off guard? People working at Alphabet off guard that finally
we get the understanding of the vertical model integration.
Speaker 13 (36:45):
You know, I don't know that I was necessarily caught
off guard by that. I think that this has been
creeping up for a while, but it does seem like
the rest of the industry is catching up to this
idea that GPUs have enormous value and are really could
be a really valuable part of you know, people's AI
mixes that you know in video is not the only
(37:08):
game in town when it comes to chips.
Speaker 2 (37:10):
Certainly, we've heard Da Davidson, you've heard Bernstein, You've had
a lot of analysts saying this could be a really
individual way of selling it. And interestingly, now maybe not
just for Google Cloud, but TPUs in and of themselves,
but aside from Alphabet, what they're doing in terms of
the chip stack, their models, how are people going to
be using them this holiday You've got a great story
out about the anxiety perhaps this is going to create
(37:32):
in the kitchen.
Speaker 13 (37:34):
Yeah, we published a story this morning about how food
bloggers are warning consumers about AI recipe slop ahead of Thanksgiving.
We talked to twenty two food bloggers ahead of the
holiday season, and all of them report, you know, traffic
declines and AI Frankenstein recipes that remix their recipes and
(38:00):
pull in bits and pieces of other recipes to create
content that is not accurate. That where if you follow
the actual recipe steps that are generated by these AI models.
You could come out with literal slab, you know, inedible food.
And it's really confusing people these days sort of where
(38:23):
to find quality content on food recipes this this holiday season.
Speaker 2 (38:29):
Maybe stick to the source for now at least. Davey
Alba punning all puns. I thank you. Meanwhile, let's talk
about what else you need helping you in your kitchen.
Maybe it's robots. Well, they've been busy dancing, they've been boxing,
they've been running marathons, so why are they not doing
more of your chores? This is memo from AI startup
Sunday Robotics. The company says it's robot is purpose built
(38:50):
for housework, trained on millions of episodes of everyday household routines.
Sunday's co founder and CEO, Tony Jao joins us, now you're.
Speaker 3 (39:00):
Robot.
Speaker 2 (39:01):
It's kind of humanoid like a little but not totally.
How does it differ from other robotics.
Speaker 14 (39:08):
Yeah, I think we just think about safety as a
really high priority item, and we define it as being
like passively safe. And what it means is that you
can put the robot into any configuration and you can
cut power and the robot will still be stable. So
this is why we build this whole mobile base as
opposed to lex.
Speaker 2 (39:28):
Where did you ultimately come to decide that this was
the way in which you should think about robotics, Maybe
not in a humanoid manner, maybe just with real safety. First,
you've got a stellar background, you're a Google Deep Mind,
Tesla Auto Power at Google X. You're also, of course
just coming out of stealth with a call thirty million
to put to work.
Speaker 7 (39:48):
Yeah.
Speaker 14 (39:48):
I think the the biggest way we think very differently
is actually on how to train these robots, not just
the design, but.
Speaker 7 (39:55):
How it obtained its intelligence.
Speaker 14 (39:58):
So normally people train their robot through tally operation, which
essentially means that you kind of lock into the robot and.
Speaker 7 (40:04):
Control how it moves.
Speaker 14 (40:05):
But the way we learn is actually very very different
that we learn from humans directly. That's essentially we design
these device a glove that captures how human do their chores,
and we're able to transfer those data directly into the robots,
and that's how the robot is able to learn from
like hundreds of humans simultaneously.
Speaker 2 (40:25):
These robots don't come cheap, but interestingly, Tony you're not
looking to sell immediately, you're.
Speaker 3 (40:30):
Looking to beta test.
Speaker 2 (40:32):
Now, how are you finding the right people to bring
these robots into their home?
Speaker 14 (40:38):
Yeah, so if you look at our website, we actually
have a huge signup doc for people who are interested,
and we're ready to get more than a few thousand
of these applications.
Speaker 7 (40:48):
So what we're going to work on.
Speaker 14 (40:49):
Next is to like very carefully sit through all these
applications and find people who what we call like founding families,
who'll be there to give us feedback, will be there
to kind of shape what a product will look like
in the future.
Speaker 3 (41:04):
What do you think the hardest element for these robots.
Speaker 7 (41:06):
Is the hardest element?
Speaker 14 (41:11):
I think it will be how people will react to
this like big robot in their homes. And again, this
is the first time that anyone has put a mobile
manipulator like a robot with arms into real living homes,
and this is something that we're incredibly excited about. I
think will be people will be pleasantly surprised by how
useful it is.
Speaker 3 (41:31):
Why do you.
Speaker 2 (41:31):
Think so many tech companies do end up turning to humanoids,
to turning to the physical form of a human rather
than a more stable basis you have.
Speaker 14 (41:43):
Yeah, I think if you're working on environments that with
a lot of stairs, or you're working on environments with
like you know, like hills, I think having lux will
be helpful in that case and for us in our
first we decide to go for a veal base just
for the simplicity, for lowering costs and to allow us
(42:06):
to move faster.
Speaker 3 (42:07):
Talk to us about costs.
Speaker 2 (42:08):
You have managed to raise seed funding from Sarah Gao
a conviction. You've now got money coming in led by Benchmark.
Speaker 3 (42:15):
What is the key cost for you? Is it the talent?
Is it the hardware? What is it?
Speaker 7 (42:21):
Yeah?
Speaker 14 (42:22):
So our hardware is actually quite differentiated from a lot
of humanoids. Even at quantity zero when we prototype it
these days, it costs around twenty five K to make,
and at quantity around like five thousand, we're able to
get a cost to below ten K. So I think
we'll be ending up selling it around to five to
ten K in the final price. And this is we're thinking.
Speaker 7 (42:44):
About robots, not like another car.
Speaker 14 (42:46):
Like purchase, but more like a fancy smartphone or a laptop.
Speaker 2 (42:51):
How does American ingenuity when it comes to robotics stack
up to that of Asia in China and how you
seeing your own supply chain develop Yeah.
Speaker 14 (43:00):
This is a great question. I think American has incredible
mechanical engineers, software engineers, but we are lacking in terms
of some of the supply chain infrastructures. So I think
we're at the point that we need to leverage some
of the growing supply chains the humanoids in China and
(43:22):
we actually share a lot of components with them so
that we can have the economy of scale before us
shipping like millions of robots.
Speaker 2 (43:30):
Any Ja, CEO of robotic startup Sunday, fascinating to have
you on.
Speaker 3 (43:35):
Thank you very much.
Speaker 2 (43:35):
Indeed, that does it for this edition at Blueberg Tech,
we do want to remind you of the market moves
today in video under significant pressure off of its lows,
still down by four percent. As we question is dominance
in the world of chips to train as.
Speaker 3 (43:49):
Well as use your models.
Speaker 2 (43:50):
That competition coming maybe from Alphabet TPUs, maybe Meta Eyeing
buying some for its data centers of the future as
Information is currently reporting Oracle once again off by one
point nine percent from New York.
Speaker 3 (44:02):
This is Bloomberg Tech. Don't forget to check out the
podcast
Popular Podcasts
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
Are You A Charlotte?
In 1997, actress Kristin Davis’ life was forever changed when she took on the role of Charlotte York in Sex and the City. As we watched Carrie, Samantha, Miranda and Charlotte navigate relationships in NYC, the show helped push once unacceptable conversation topics out of the shadows and altered the narrative around women and sex. We all saw ourselves in them as they searched for fulfillment in life, sex and friendships. Now, Kristin Davis wants to connect with you, the fans, and share untold stories and all the behind the scenes. Together, with Kristin and special guests, what will begin with Sex and the City will evolve into talks about themes that are still so relevant today. "Are you a Charlotte?" is much more than just rewatching this beloved show, it brings the past and the present together as we talk with heart, humor and of course some optimism.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.