Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is live
from coast to coast with Caroline Hide in New York
and Vla Lolow in San Francisco.
Speaker 2 (00:22):
This is Bloomberg Tech coming up.
Speaker 3 (00:24):
President Trump allows Nvidia to ship its eight two hundred
AI chip to China in exchange for a twenty five
percent surcharge.
Speaker 4 (00:32):
Plus. We'll drill more into the Warner Brothers, Discovery, SAGA
and antitrust concerns rising from the proposed.
Speaker 3 (00:37):
Bids, and Microsoft is committing seventeen point five billion dollars
over four years to help build India's cloud and AI infrastructure.
Speaker 4 (00:46):
We dig into the market's first and foremost so and
ed not much of a movement. We're up ten percent
on the NASAK one hundred jolts data. Basically jobs opening
coming in better than had been anticipated, but scratched beneath
the hood, and maybe less optimism there the market had
hoped for. More Broadly, we're in wait and see mode
for the FED tomorrow, but we're not in wait and
see mode when it comes to some significant moves to
(01:07):
one stop.
Speaker 2 (01:08):
Yep in vidio.
Speaker 3 (01:10):
This is how Invidia is traded over a twenty four
hour period and two sessions, we're down marginally basically flat.
In Tuesday's session, when news broke from the President that
Nvidia would be authorized to ship H two hundred to
China under certain circumstances with a twenty five percent surcharge,
the stock spiked. But then when reports hit that China
(01:33):
would move with its own restrictions on how and why
and which companies could have access with that technology, the
gains fizzled, and the same can be applied to both
AMD and Intel.
Speaker 2 (01:44):
Let's stick with the Nvidia story.
Speaker 3 (01:46):
As you look at that two day chart of the
stock and the big jump that led to a big fate.
Let's bring in Blue Mosey and King, who leads our
coverage of semiconductors.
Speaker 2 (01:54):
That's the news broadly.
Speaker 3 (01:56):
What else do we need to know about the calculus
of America's decision to allow H two hundred in particular
to go to China in certain circumstances, and whether or
not China wants it?
Speaker 5 (02:06):
Yeah, I mean, that's exactly it. We've shown a willingness
to sell. We've effectively said, hey, we were prepared to
give you something better. The question now is whether China
actually wants that and how Beijing will react. We have
the President saying that he's cleared it with his counterpart
in Beijing, but we haven't really seen a concrete response
on how that will manifest itself in terms of orders
(02:28):
or shipments.
Speaker 4 (02:29):
We know that the articulated total addressable market is fifty
billion dollars for Jensen Wang, But what does the H
two hundred bring that the H twenty doesn't.
Speaker 6 (02:38):
How much better is it?
Speaker 5 (02:40):
Yeah, I mean it's multiple times better. It's still a
mainstream chip. It's still in widespread use in the world's
data centers. It's not, of course in video's latest. It's
nowhererea as good as the Blackwell generation, but it's still very,
very good and according to analysiss a lot better, which
is more important than anything else that the Chinese can
make themselves.
Speaker 3 (03:01):
So that there is the what happens next. China has
two domestic players maybe three, that have AI accelerators. They
are not as performant as H two hundred, we think,
But the main issue for China is also the supply
constraint and manufacturing.
Speaker 2 (03:17):
Explain that part of the story.
Speaker 5 (03:18):
Yeah, I mean, as you know, China has been cut
off from advanced semiconductor manufacturing, including TSMC. So the chips
that it can make itself are constrained by the actual
amount that it can make, but also the quality that
it can make based upon the manufacturing capabilities of plants
inside China.
Speaker 4 (03:38):
And quickly, just when I reference H twenty, the fifteen
percent never went to the government because that was never
signed off legally. Do we know whether the H two hundred, Yes,
we've got a truth social but will it be signed
off legally by the US at least?
Speaker 5 (03:51):
Yeah, I mean, as we put in our story yesterday,
we had some reporting on that. And then exclusive for US,
this is going to be a different arrangement because these
ships are made in Taiwan. What's going to happen apparently
is that they're going to be imported into the US,
which would make them subject to an import tariff. So
that's how we basically the US government gets paid and
then they become re exported to China.
Speaker 6 (04:14):
The nuance always clear. Thank you in King.
Speaker 4 (04:17):
Now let's discuss the broader impact now on the tech markets.
We bring in Ebeks Deshkaya's analyst at Swiss quote IBEC.
What do you make of this potential for Invidio at least,
and whether or not China will indeed import the market
reaction from your perspective.
Speaker 7 (04:35):
Well, this is a win for Nvidia, but it has
to be taken with a pinch of salt because we
already hear that there are attentions among US politicians regarding
what this means for the national security for US is
a very sensitive segment of business. China is ensure to
let its own companies buy these chips for the same
national security issues and also because they're backing the domestic
(04:58):
chip production. So for Nvidia, which assumes that is China
sales are now zero, every chip that is going to
be sold to China is a bonus. It is a tailwind.
But for long term forecast with stale remain very much
coaches regarding what this me is and how stable the
stale is going to be.
Speaker 2 (05:16):
Upek.
Speaker 3 (05:17):
The base case assumption from Nvidia is right now in
the fiscal year and maybe next fiscal year, zero revenues
from China. That clearly this news has led everyone to recalculate.
Would you assume some upside from Chinese AI data censorship
revenue next year?
Speaker 7 (05:37):
Well, again, the geopolitical backshop is so uncertain that the
best thing to do in terms of the safest a
way to go with these assumptions is to keep that
forecast at zero percent. But obviously investors are happy to
hear that there is an opportunity in China for Nvidia,
but again that opportunity is very much uncertain right now.
(06:00):
We don't know if China is going to allow these
chips to be sold. And again we don't even know
if the Chinese companies will be willing to buy these chips,
given that it is politically very sensitive, and if they
could use domestic alternatives, even though they are less efficient,
maybe it's going to be a better deal for them,
especially from a political and geopolitical perspective. So in my opinion,
(06:23):
investors will continue to assume that the Chinese sales is
in the Nvidia's Chinese sales are going to be zero
with a little bonus maybe if some of these chips
actually get sold.
Speaker 4 (06:35):
If I bring us your global perspective, as you sit
in Switzerland, this tussle between US and China to say
who's got more dominance, But what do you hear from
your perspective as to how far along China really is
with its domestic chip manufacturing and use well.
Speaker 7 (06:51):
Obviously, US and the US technology is ahead of China.
What China is good at doing is basically to find
the chips to do manufacture from manufacturer chips in order
to serve their own needs and serve the basic AI needs.
So China is generally better in transforming their I investments
(07:13):
into real revenue, whereas the US is prestige technology. Hope,
that's really aiming for the best. But even from the
European perspective, where we could be a little bit more neutral.
While US is allies first of all, and US is
well ahead from a technology perspective than China, then China
is clearly coming and the fact that these chip restrictions
(07:35):
have hit the Chinese technology sector means that they will
be accelerating their efforts and do anything in their power
to catch the US technology.
Speaker 4 (07:44):
It really has been a mounting concern from some of
these tech leaders that perhaps were underestimating how sophisticated China is,
is the market, underestimating how superior TPUs are, how other
chips are being manufactured by big tech players for their
own inference.
Speaker 6 (07:59):
How much we need to take that into the Nvidia story.
Speaker 7 (08:03):
Well, I think that what TPUs are interesting, but they
are actually you know, serving in one particular area, and
that's AI. Obviously, the fact that Google is now out
there and commercializing the TPUs is a is a negative risk,
is a negative factor for Nvidia. It is a risk
in the sense that for those companies who are looking
(08:26):
only for their AI applications, they would find it cheaper
and as efficient as buying Mvdia's GPUs. But at the
end of the day, mvd is not only selling chips,
They're also selling the ecosystem that comes with it. So
not every company is going to be comparatable going into
the TPUs just because as cheaper and as efficient. I
think that Nvidia still has the GPU market and the
(08:49):
ecosystem well anchored, and Google, for example, with this TPUs
doesn't really have it.
Speaker 2 (08:55):
You peck us good.
Speaker 3 (08:56):
This guy a Swiss great senior market analyst. Thank you
very much for coming up. We'll come back on the
talk of Hollywood, Warner Brothers, Discovery and the antitrust concerns.
The proposed bids are definitely raising.
Speaker 8 (09:08):
That's next.
Speaker 2 (09:09):
This is Bloomberg Tech.
Speaker 4 (09:24):
Let's get to what everyone has been talking about in
Entertainment the rival bids for Warner Brothers Discovery after Paramount
came out with its own hostile takeover offer yesterday. Netflix
co CEO Ted Serranos says he's not too worried. He
spoke at the UBS Global Media and Communications Conference in
New York yesterday.
Speaker 6 (09:41):
Just take a listen.
Speaker 9 (09:43):
Today's move was entirely expected. We have a deal done,
and we are incredibly happy with the deal. We think
it's great for our shareholders, think it's great for consumers.
We think it's a great way to create and protect
jobs in the entertainment industry. We're super confident we're going
to get it across the line of finished.
Speaker 8 (10:00):
So we're excited.
Speaker 4 (10:02):
Let's bringing bloom bags. Luca Sure, author of the screen
Time newsletter. Should he be as confident Lucas?
Speaker 10 (10:09):
First all, I love those old photos that you have
of Ted back from I guess that's the release of Lillehimmer.
Speaker 8 (10:15):
Of course he's going to project confidence.
Speaker 10 (10:18):
He agreed to a deal, and what Paramount is proposing
at the moment is not different from what it proposed
when it lost this auction last week. Right, they're sticking
with their thirty dollars a share and their insistence is
that the board just didn't take the time to really
consider it, and that if more shareholders hear their point
of view, that they will be swayed. I think if
Paramount stays on that current course, Ted has reason to
(10:40):
be pretty confident because he already won with that. Risk
for Ted and for Netflix would be if Paramount comes
back and offers even more money, because then the board
and the shareholders would likely think about opening back up.
Speaker 3 (10:54):
Lucas you've spoken at length and in some detail in
the last year and two years with both Ted Surround
and with David Ellison. The reporting overnight was really focused
on the coalition that David Ellison has pulled together for
the financing, and part of that includes people that have
proximity to the administration, Jared Kushner and his private equity
(11:16):
firm being the easiest example. Just reflect on what you
know of both CEOs in the competition between them, but
also David Ellison's ability here to manage that relationship with
the White House if there is one.
Speaker 10 (11:29):
Right well, look, I think one of the reasons that
Netflix prevailed in the initial auction was that you put
Ted Surrandos in a room with the Warner Brothers, Discovery
board and leadership, and he is charming and charismatic and
it is generally very good in those situations.
Speaker 11 (11:49):
Right.
Speaker 10 (11:50):
It's one of the reasons that he is Netflix's primary
person when it comes to wooing talent. David Ellison has
a little bit less experience with that, right, he's younger.
He is very amiable, but he doesn't have that He's
not kind of like oozing the charm in the same
way that Ted is. We've now entered a different phase
of this to your point, where a lot of it
is more about who you know and their proximity to
(12:11):
people in positions of power. That's an area where Ted
surround us is still very strong. But he comes from
the generally from the democratic side of the political aisle,
and so Netflix is having to kind of find the
people who can help them get into the Trump administration.
To your point, David Allison, in part because of his
father Larry, and then also the various financiers he has,
(12:32):
he has a lot of different ways to access the president. Now,
of course all of this is moot, or should be
moot if the board decides one way or the other.
Speaker 8 (12:41):
But you know, times are different.
Speaker 6 (12:43):
Now, I guess times are different.
Speaker 4 (12:46):
But let's just go back to cold hard cash here,
because this really comes down to how much the investors
are going to think the cable networks are worth.
Speaker 8 (12:54):
Right, That's a big part of it.
Speaker 10 (12:56):
Yeah, I mean the fundamental breakdown between what Paramount thinks
of the state of play and what Warner Brothers Discovery
thinks of the state of play is the value of
the cable networks and the importance of cash. Paramounts bid
is all cash. Netflix is bid is primarily cash, but
not all cash. Paramount is valuing those cable networks at
about a dollar a share. By that valuation, their bid
(13:18):
is superior to Netflix. Warner Brothers Discovery is valuing it
at three or four dollars a share, which and.
Speaker 8 (13:24):
By that metric, Netflix is superior. That's why I.
Speaker 10 (13:26):
Think it feels almost inevitable that Paramount if it really
wants it will have to come over the top and
just add let's say, three dollars to its bid.
Speaker 8 (13:35):
But we're not there yet.
Speaker 2 (13:37):
Bloom most Lucas, sure, thank you very much.
Speaker 3 (13:39):
President Trump first raised anti trust concerns over Netflix's market
share if the deal goes through. Now, Senator Elizabeth Warren
also weighing in on paramount spid, saying that merger would
be quote a five alarm antitrust fire. Let's break it
all down with Jennifer Huddleston, Senior tech Fellow at the
Cato Institute whose work covers antitrust. A deal, a proposed
(14:03):
deal of this scale and size and value is inevitably
going to be looked at by regulators and reviewed, in
part because we have a set of codified guidelines and
rules from twenty twenty three that allow for that to
be the case when regulators do look at it. Jennifer,
what is the top consideration about this market that they'll
(14:26):
be making.
Speaker 12 (14:27):
Thank you so much for having me, and you're right.
One of the things we're going to see in this
conversation is how do these twenty twenty three guidelines potentially
play out in this case and what does that mean
if such ends up in court, as well as what
does that mean for any challenges. One of the interesting
things is going to really be around this question of
market definition. What is the market that these two platforms
(14:49):
are potentially competing in and what does that mean for
the actual average consumer? Is this online streaming subscription services.
Is this some kind of bigger attention economy issue or
is this something you know specific about the amount of
time individuals spend on any type of entertainment.
Speaker 3 (15:09):
As it goes about Jennifer, will any of the regulators
involves look at what the benefit is to the consumer
in either merger or deal going through.
Speaker 12 (15:23):
I mean, we can certainly hope so. US antitrust law
is supposed to be based in the consumer Welfare Standard,
which means what regulators should really be looking at is
will this mergerer acquisition harm consumers in some way? Will
it negatively impact prices? Will it negatively impact the ability
to have to have a certain quality or certain other
(15:44):
elements that really should be based on sound economic factors.
This shouldn't be about a subjective idea of how many
players should be in the streaming market. It should really
be focused on this question of our consumers going to
be harmed. Our consumer is going to be worse off
if they have this combined company as opposed to two
(16:04):
separate companies.
Speaker 4 (16:06):
The Cato Institute is so interesting because it's a libertarian
think tank and it's really thinking about individual liberty, about
limited government, free markets. You come in it from that angle,
which Jennifer, give us your global perspective here a bit,
because this doesn't just get sign off from the United States.
Speaker 6 (16:19):
We go worldwide.
Speaker 4 (16:20):
We're just seeing what the EU is doing again today,
for example, with Alphabet and competition within AI. Are they
going to hit roadblocks there as well?
Speaker 12 (16:29):
I mean, that's certainly going to be a question depending
on the nature of any particular transaction. And we've seen
this happen in other cases when it comes to questions
around large US tech companies and their potential acquisitions. For example,
we certainly saw a large debate around the Microsoft Activision
acquisition in Europe as well, and ultimately that was able
(16:50):
to go through. We've also seen European regulators trying to
put up various structural roadblocks that would allow some of
America's leading tech companies to continue to innovate, continue to
go into new markets. So I think it's certainly a
worthy question of not only how might such transactions play
out when it comes to US regulators, but if this
is something that's a global international debate, are there other
(17:13):
regulators where there might be other conversations that get had
along the way.
Speaker 4 (17:17):
It's interesting Jennifer that Netflix has tried to front run
a lot of these concerns by putting out the statement
that consumers will win from their perspective because already they'll
signed up to Netflix and HBO and we'll broadly they'll
probably get more bang for the bug. But they're also
having to talk about how they're going to benefit content
creation or indeed the industry of Hollywood writ large.
Speaker 6 (17:36):
Should they have to do that?
Speaker 4 (17:38):
Will the narrative be based on that in any way
as to how this might help jobs and creativity.
Speaker 12 (17:45):
I think one of the big questions is going to
be around that market definition. Is the market the definition
definition include markets for content creation? And if so, what
does that mean in this today and age where we
have such significant amounts of user generator content. Is it
only about studio created content or is it also about
looking at content more holistically the way short form video
(18:08):
or even long form video on platforms like YouTube have
arisen to allow creative's new outlets. So we'll certainly see
that be part of the conversation as well, possibly if
that gets included in the market definition.
Speaker 4 (18:21):
Jennifer hudleson great to catch up with you, Senior Tech
fellow over at the Cato Institute. Now we're going to
think about entertainment a little bit more because Walt Disney
is bringing Jimmy Kimmel back for at least one more
year now, the late night star who was suspended in
September over remarks about slain GOP activist Charlie Kirk. Where
he's going to continue to host Jimmy Kimmel Live until
(18:42):
May twenty twenty seven under a new one year deal
with Disney's ABC network.
Speaker 6 (18:47):
It's all according to sources.
Speaker 2 (18:49):
Ed okay, coming up.
Speaker 3 (18:50):
Investors have been disappointed in Apple's lack of AI strategy.
Now that weakness maybe a strength. We're going to talk
about why next. This has been big ten.
Speaker 6 (19:13):
Take a look at Apple shares since June. Have you
noticed this?
Speaker 4 (19:17):
Back then, analysts were talking about the iPhone maker falling
behind big rivals in the AI race, But now the
company is the beneficiary of rising doubts about the AI trade.
Speaker 6 (19:26):
Here to talk us through is Brumberg Gequiti's reporter.
Speaker 4 (19:28):
Carmen Ranicky, who is gonna help us really illustrate a
story that everyone is reading today. If almost we'd forgot,
we'd all thought Apple had been beaten up, and actually
it is worth four point one trillion dollars and the
second most valuable company on the SP five hundred.
Speaker 6 (19:42):
The narrative shifted, Yeah, it totally shifted.
Speaker 13 (19:45):
It's really like a complete turnaround from the beginning of
the year. Now it's you know, in spitting distance of
taking over there's the biggest company from Nvidia, right that
gap is ever narrowing in terms of market caps. So really,
what we've seen is that people are getting more worried
about the AI trend. They're looking at these companies more critically,
and Apple sort of at the same time has benefited
(20:06):
because it's not seen as this huge AI play I think.
At the same time as well, what we've seen is
that there's been a little bit more shine around the
new iPhone iPhone seventeen. The big story for Apple, at
least I've been hearing for the last few years, is
the upgrade cycle, right, people getting the new phones that
they've had for four or five years, and analysts and
investors are starting to see that happen and they're excited
(20:28):
about it more and more going forward. But you know,
at the same time, Apple's valuation is near record high
as it trades at thirty two times forward earnings or
nearly thirty three times forward earnings. That's more than any
other stock in the mag seven aside from Tesla.
Speaker 4 (20:44):
So it's expensive. And what's extraordinary is that thirty three times.
I mean, we hardly have.
Speaker 13 (20:48):
A C trade at those sources of levels, right, Yeah,
And if you think about it, it's pretty incredible because
Apple's revenue growth is very small compared to Nvidia, for example.
I mean, they've been doubling revenue and having these you know,
huge numbers. Apple's not been doing that.
Speaker 6 (21:03):
But its revenue is so massive.
Speaker 13 (21:05):
It is, it's so mad. It's not that it's not
putting up big numbers, but the growth, the percentage gain,
you know, year over year, is not really comparable to
in videos. So it's interesting to see it become more expensive.
Speaker 2 (21:18):
Have you studied the technical charts?
Speaker 3 (21:19):
We think Apple in the short term might fade a little,
but long term, the ball thesis is really simple. Right
when AI goes mainstream common, where are people going to
be using that technology?
Speaker 2 (21:30):
Yeah?
Speaker 13 (21:30):
So one thing that I keep hearing, and I just
spoke to an investor about this, is they really want
in a gentic series. They want that you know, conversational
experience with the phone. And so this is something that
actually I think we see from Apple we've seen in
the past, is that they wait on the new technology
until they can really get it right. Obviously they've introduced
some you know, Apple AI, but I think we're going
(21:51):
to see just leaps and bounds going forward. And that's
when investors are looking at that true you know conversational.
Speaker 3 (21:56):
Assistant, looking at A and R, the consensus ratings on
the stock. Where are we right now in the cell
side's position on Apple.
Speaker 13 (22:06):
So I think what we've been seeing, or what I've
heard at least from technical analysts, is that there could
be a little bit of downside sort of baked in here.
I mean, Apple is near record highs. It's you know,
price to earnings ratio is near fifteen year highs I
think for the stock, or at least the highest since twenty twenty,
and they think that there could be a little bit
of a shakeout, right, or that some healthy selling might
(22:29):
give a new place for investors to jump back in
and actually really start buying with momentum.
Speaker 3 (22:34):
Bloombos come and Ryan Okey with the Apple stock story.
Thank you very much like coming out on the show.
We're going to get back to Warner Brothers saga and
discuss with Ross Gerba from Gerba Kawasaki, who's got skin
on the game with Netflix. What's his take We'll find
out next. This is Bloomberg Tech. Welcome back to Bloomberg Tech.
(23:05):
Our top story is Nvidia getting approval from the government
to export H two hundred chips to China, but with
the condition of a twenty five percent surcharge. And yes
they session the stock spike tire we're now softer two
tens to one percent. That same rule applies to both
AMD and Intel, but in their cases, of course, much
(23:26):
smaller exposure in the market, although it'll be interesting to
see how the dynamic goes going forward. The reports overnight
very much focused that China will put in place some
rules that will move to restrict the access of Chinese
technology companies to those chips, even if America is allowing it.
The other top story the saga behind Warner Brothers Discovery.
(23:47):
We have the bid from Netflix cash and stock twenty
seven dollars seventy five cents a share. We have the
bid now from paramount S Guidance thirty dollars per share,
and we are looking deeply into the financing of both.
There is the Wall Street take, which we'll get into,
but there's also the Hollywood takes. What is this moo
for the industry going forward Later in the program, Carroc,
we'll get to that.
Speaker 2 (24:07):
What are the news?
Speaker 4 (24:08):
Yeah, it's time now for talking tech head and first start,
we're talking Microsoft committing seventeen point five billion dollars to
help build India's cloud and AI infrastructure.
Speaker 6 (24:16):
The spend coming over four years.
Speaker 4 (24:18):
We'll focus on scale, skills and sovereignty, according to Microsoft,
so aligning with Prime Minister Nerenda.
Speaker 6 (24:23):
Modi's AI goals for his country.
Speaker 4 (24:26):
Then there's the EU who's investigating whether Google is abusing
its market power in its AI rollout. Regulators are probing
whether Google is favoring its own model and how AI
overviews and AI mode use and compensate publishers for content.
Google argues the probe quote risks stifling innovation and then
sticking with Google remember these Google Glass been ahead of
(24:49):
its time, now developing two new smart glass designs as
it battles Meta in the AI device race. We understand
one with displays, the other audio only first pair's expected
next year, and.
Speaker 2 (25:03):
Let's go back to that in video story.
Speaker 3 (25:05):
President Trump has announced a major policy shift on Invidia,
allowing the company to sell its H two hundred AI
chips to China in exchange for a twenty five percent surcharge.
And as Bloombozi and King explained, because the chips are
manufactured in Taiwan, they enter the United States, that's the
point of surcharge in tariff. But there's more about the
market opportunity Bloomberg Intelligence, as man Deep Singh joins, us,
(25:28):
I was really interested in Bi's reaction to this because
prior to yesterday, this was about opportunity lost in China,
a market that Jensen Wang has said is fifty billion
dollars of addressable market and market opportunity. You've done some
math on where you think at BI there is some
revenue to now be had in China.
Speaker 2 (25:47):
What is that number?
Speaker 14 (25:48):
Yeah?
Speaker 11 (25:49):
Look, I mean it's hard to pinpoint exactly what portion
of their fifty billion and Video can capture through H
two hundreds, But there is no doubt that you know
frontier ll M companies, you know from Deep Seek to
you know the Alibaba, Quinn and Kimmy all these models
(26:09):
have been trained and they have kept up in terms
of you know, functionality with the frontier models here, whether
it's Gemini or open Ai. And so from that perspective,
you have to ask yourself, how have these companies trained
their models and is it all based on their in
house or you know, Huawei chips and the answer it's
(26:30):
hard to discern, you know, sitting here, but to my mind,
they would welcome any opportunity to get a big Nvidia
cluster because at the end of the day, when it
comes to training, Nvidia is proven to be the one
chip company that is the most useful for building the
big training clusters. Yes, we have the TPU news and
(26:50):
all that, but everyone universally wants to train their models
on Nvidia. And so from that perspective, it's two hundred,
just on the training side, could be a pretty sizable
twenty five to thirty billion dollars opportunity next year.
Speaker 4 (27:04):
Twenty five to thirty billion when the total addressable market
is fifty billion. From Jensen's perspective, so the narrative goes
that maybe Ali Baba and the others are able to
access h two hundreds, will there than be this building
of momentum that they can convince the Chinese government that
they should be allowed them and indeed that they can
therefore jostle in for even more sophisticated Blackwell architecture even
(27:27):
the next.
Speaker 11 (27:27):
Yeah, and that's where you know, the continuity is the
main point, because what Nvidia gives you is that backward comparatibility.
Even if you move to you know, the newer version
of their architecture, which Nvidia will be releasing Rubin and so, yes,
Chinese market will be delayed from that standpoint, But what
you want to see is they being able to use
(27:48):
you know, whenever the black Hole version is available to
the Chinese market, they should be able to use that,
because then it becomes a cluster that they could use
for inferencing down the line. Right now they use you know,
the same chips for training, but over time they could
use it for inferencing. And what we have heard from
the neocloud providers here is the nvideo chips have a
(28:09):
very long useful life, and you know, everyone wants to
use them for as long as possible. And the Chinese
market has no shortage of power, unlike you know the
market here. So from that perspective, it does make sense,
you know, for them to use those clusters for as
long as they can.
Speaker 2 (28:28):
In considering policy.
Speaker 3 (28:30):
I'm sure the White House made an assessment of China's
energy capabilities, right, Has Bloomberg Intelligence done any of its
own study or aggregation of maybe third party data on
realistically the production volume capabilities of a Huawei, a Cambra
con because no matter the technology's performance visa VI any
(28:52):
generation of American technology, the limit to China seems to
be supply constraint.
Speaker 11 (28:58):
Yeah, and look, every executive has called out, you know,
tokens per what as a key metric that they are
focused on, and that's a measure of intelligence. Right, So
per unit of power, how many tokens can you generate?
So from that perspective, you know, the Chinese market is
now different. They want to maximize the tokens generated per
(29:19):
what for you know, their model companies, even though they
have more wats available. And look over here to your
prior question, I mean right now, the estimates are we
will probably be adding up to hundred gigawatts of capacity
over the next five years. And we've heard commitments from
Open Ai which everyone has started to question. Now in
(29:41):
terms of you know the twenty six gigawatts that they
have been talking about, but clearly the numbers are big
in terms of, you know, the gigawatts that are in
the pipeline, in terms of, you know, the capacity that
companies want to add. But you know, the efficiency part
when it comes to token per what will remain a
key metric because you won't be able to bring all
(30:01):
their power online, you know, in the next twelve months,
and it's going to be a lot more staggered than
people can imagine.
Speaker 4 (30:08):
You recently visited Asia, Yeah, do they think China is
as far behind the US as US thinks China's behind
the US.
Speaker 11 (30:16):
Well, their notion of you know, model layer as an
intelligence layer is to open source a lot of their models,
and that's where they have kind of taken a different approach.
The models are available for any company to use. In fact,
they want to build an ecosystem outside of China where
Europeans or other companies adopt their open source models. And
(30:38):
from that perspective, they're trying to distill a lot of
their models to smaller versions so that they can run
on the phones. I just feel like they have espoused
that open source concept and you see that with Bydance
and all these companies, and they're trying to get to
agentic use case a lot quicker in terms of models
(30:58):
being able to you know, book your travel or book
your restaurant. And I mean we've seen that how good
the v chat app is in terms of you know,
being that super apps. So that's what they're trying to
get to with the agentic use cases is deploy a
lot more of that using the LLLM layer and do
that at a very efficient price. Because at the end
of the day, you know, they have a spouse cost
(31:20):
and efficiency when it comes to running their infrastructure, so
they want that to be reflected at the agent layer
as well.
Speaker 4 (31:25):
And they have to be And yep, saying great analysis
from Bloemberg Intelligence.
Speaker 6 (31:30):
We appreciate him. Now coming up, we're going.
Speaker 4 (31:32):
To be speaking with the CEO of Generative Media platform
File about the startup's latest fundraise.
Speaker 6 (31:37):
What g youp use there using this is blue bag tech.
Speaker 3 (31:56):
AI starts up Fall has raised a one hundred and
forty million dollars year. According to a Bloomberg source, It
gives the company a four point.
Speaker 2 (32:04):
Five billion dollar valuation.
Speaker 3 (32:05):
Just months after closing a Series C round at a
one point.
Speaker 2 (32:09):
Five billion dollar valuation.
Speaker 3 (32:11):
File hosts AI models specifically for images, video and audio generation,
aiming to give developers and enterprises unified access to those tools.
Files CEO and co founder Muckiger joins us now here
in San Francisco.
Speaker 2 (32:26):
Congratulations on the round.
Speaker 8 (32:27):
Thank you, Ed.
Speaker 3 (32:29):
You have used m and A as a strategy. You
have very interesting infrastructure where you're running h one hundreds
principally upgrading to Blackwell generation. Yes, those are very specific
to the modalities of image and audio, that's right, right.
But outside of that infrastructure, you've gone out buying companies.
Are you going to use some of the money raised
(32:50):
for that purpose?
Speaker 15 (32:51):
Absolutely?
Speaker 2 (32:51):
Yes, Why what's the logic hit?
Speaker 8 (32:53):
Definitely?
Speaker 15 (32:54):
I mean it's definitely a very fast race right now
trying to become one of the largest companies right now
in the space. And you know, one of the ways
we get talent is buying buying companies. So MNA is
definitely a very good way of doing that.
Speaker 4 (33:13):
Well, I want to go back a step, because you
claim to have the fastest inference that for modern generative
models multimodus. How what are you doing to remove the
technical economic barriers that you say exist when deploying GPUs.
Speaker 14 (33:27):
Yeah.
Speaker 2 (33:28):
Absolutely.
Speaker 15 (33:28):
So what we have is we have this inference engine
that is a proprietary engine that can run image and
video models up to three to four times faster, and
we do that by optimizing en video chips to run
these workloads.
Speaker 8 (33:42):
So this is a proprietary technology that we built in house.
Speaker 2 (33:45):
It's software that you've built.
Speaker 8 (33:47):
It is software, that's right.
Speaker 2 (33:48):
You know.
Speaker 3 (33:49):
That's the interesting debate in the world we're in. If
you look at some of the ASEX and Customs Silicon solutions,
the concern is they don't have the library of software
to go with it. We're here to talk about FOLE
but Wohild, you just reflect on the reality of needing
to use in video for that reason, but also you
build your own software.
Speaker 15 (34:07):
Absolutely, yeah, So I'll give you a quick summary of
the dilemma there. So we could run our workloads on
a six, but the scale at which we operate is
a lot larger, and there's no real ACIK out there
that has the scale.
Speaker 14 (34:22):
That's one.
Speaker 15 (34:23):
And then the second point is that one of our
key value props is being day zero and if you
have a model and you want to run that model
whenever it's available, the first date is available, you actually
have to have like really good software that's already ready
to deploy. With A six typically what you have to
do is actually customizing the software a lot before you
(34:46):
can actually launch it. So with Nvidia, the software stack
is so much more mature, so we can actually do
these day zero releases. So this is one of the
key reasons why we prefer Nvidia.
Speaker 4 (34:57):
Right now, Okay, we now have the news maybe be
able to ship is H two hundreds to China. How
easily accessible is the supply that you need right now?
Is that a concern for you that sort of power
going to China as well?
Speaker 14 (35:11):
Right?
Speaker 15 (35:12):
So, for our worklodes, specifically for image and video, H
two hundreds are a little bit of overkill for us,
at least for the Hopper generation, so we actually prefer
H one hundreds and then H two hundreds are more
preferred for like language models, which is not our domain.
But when we're actually moving away from the Hopper architecture
(35:34):
to black Mail architecture, we're actually just going directly from
H one hundreds to B two hundreds or B three
hundreds because that actually gives us more flops per watt
or per dollar, so that we're already rolling this out
as B two hundreds and B three hundreds are becoming available,
we're actually moving over. I think you know the economics
(35:57):
of this is that many of the players are actually
going to move to Blackwells because you just get better
price performance.
Speaker 6 (36:04):
Fascinating.
Speaker 4 (36:05):
You're providing developers with over six hundred image, video, audio,
and three D models.
Speaker 6 (36:09):
So the other person to talk about it all Okaiga.
Speaker 4 (36:11):
Thank you very much, CEO and co founder Afar on
the fundraising. Now coming up, we'll get back to the
Warner Brothers saga with Ros Gerbert Gerbert Karasaki.
Speaker 6 (36:19):
This is Bloomberg Tech.
Speaker 3 (36:29):
President Trump has announced a major policy shift on Nvideo,
allowing the company to sell its H two hundred AI
chips in China in exchange for a twenty five percent
surch once. The Financial Times reported regulators in Beijing are
considering ways to allow only limited access to the H
two hundred bluemogs. Mike Shepherd joins us from Washington. Mike,
what do we need to know about the China response
(36:50):
and the national security implications?
Speaker 16 (36:53):
Well, really, the national security implications are the focus here
in Washington. The President's decision really broken article of faith
that has been held here in the Nation's capital for
so long, and that is that the US must deny
China and other adversaries access to the latest American technology.
And yet with this the administration is arguing that, look,
(37:14):
if we want to compete with China globally and maintain
the US edge and artificial intelligence, we have to be
able to compete inside China. And that's an argument we
have word heard Jensen Wang make here in the Nation's Capital,
and he's also done so privately with President Donald Trump,
and the two men did discuss export controls during a
private meeting last week here in DC while he was
(37:37):
also visiting members of Congress. So in a way, this
is giving Nvidia a lot of what it wants, maybe
not everything. They won't be able to sell their latest
generation Blackwell chips to China, but the Age two hundred
is far more powerful.
Speaker 2 (37:49):
Now.
Speaker 16 (37:50):
The question remains what is the US getting in return
and how much further will the president go in terms
of looking at national security controls as a trade instrument.
Speaker 4 (38:00):
Mike Shepherd, as always keeping up today, we thank you. Look,
the government has been potentially weighing in elsewhere as well,
and it comes to the purchase of Warner Brothers, Discovery,
Netflix paramount fighting for the assets.
Speaker 6 (38:12):
Let's bring in.
Speaker 4 (38:13):
Ross Gerbert CEO and president of Gerbert Kawasaki for more.
What's interesting is we've heard from President Trump and there
might be an issue of Netflix becoming so significant in size.
Speaker 6 (38:22):
You own Netflix shares. What do you make of it?
Speaker 14 (38:25):
Well, I think there's some truth to that.
Speaker 17 (38:27):
Netflix is sort of the big monster out here in Hollywood,
and there's only so many places to sell your movie.
So if you take out Warner Brothers, you're basically going
to Ted even more often to try to sell projects.
Speaker 14 (38:38):
So there is some truth to that. I think when
you look at.
Speaker 17 (38:41):
The broader media landscape, I'm not worried on a monopolistic
sort of perspective, but it certainly creates less competition, not more.
Speaker 14 (38:51):
Russ.
Speaker 3 (38:51):
One of the distinctions between the two bids is that
Netflix is not just being just the studios and streaming
platform business, but it is a hash and stock deal
that might seem a little bit dry, but as a
Netflix shareholder, it assumes that there will be some increase
in value, right if the merger goes through. What is
your kind of perspective on that.
Speaker 17 (39:13):
Well, that was what I was thinking about last night
and what I tweeted this morning that when I really
sat down and thought about it, I still think the
Netflix bid is superior because of this issue, because when
you do an all cash deal, you're basically say all
current owners will not participate in any future profits of
the company. And if I was a Warner Brothers shareholder,
(39:35):
I would want some of the Netflix stock as part
of the deal, because why would you want to give
up all the upside of now being a part of
basically the most powerful company in the history of Hollywood.
Speaker 14 (39:45):
So, you know, that's what I tweeted this morning.
Speaker 17 (39:48):
I think Ellison's going to have to come up with
another two to three dollars one way or another in
this bid to really be superior to the Netflix price.
Speaker 4 (39:57):
From what I understand, you don't own Paramount's guidar used
to No, I don't, but I don't know.
Speaker 17 (40:02):
I bought it, Yeah, I bought it recently because you know,
with several of the people I know in Hollywood are
extremely bullish on Allison, and so I bought it recently.
But when the deal started and it turned into a
bidding war, Warner isn't worth even close to thirty dollars,
So everybody's overpaying for this asset right now because.
Speaker 14 (40:21):
It's so rare.
Speaker 17 (40:22):
But what I fear and why I immediately turned around
and sold the stock, is they're going to end up
overpaying for this asset. And Warner has been an albatross
on every company that's that's ever owned it. I've known
Warner since the days, from the early days. Nobody's ever
made money owning Warner.
Speaker 4 (40:39):
So why do you think Netflix is just good buying
the bits of it that it should do? You not
like their what about paramount Skuyouts wanting the whole kaboodle?
Speaker 6 (40:47):
Is two to three dollars the.
Speaker 4 (40:49):
Right amount they should be forced to pay to inherit
the cables as.
Speaker 14 (40:51):
Well well, the cable assets.
Speaker 17 (40:54):
By keeping it is really a power play because they
want to take control of CNN, So it's a way
for the right to take control of CNN. And it
really has nothing to do with any value because cable
assets are being spun off, for example with Comcast, and
I think Disney needs to do the same thing because cable's,
you know, a dying business and it's the numbers go
down so when you think about why would they really
(41:14):
want this, that's just politics and that's I think a
big issue in this deal that people in Hollywood who
are generally on the left eating Netflix, don't want the
right to get a hold of these assets. So there's
a lot more to this than just money, and that's
the cable assets.
Speaker 4 (41:34):
I just want to push back on what makes you
articulate you think this is about politics, because I think
more broadly, certainly the Edison family would come in and say,
this is about the future of content, about delivery. This
is not about in any way us having any bearing
on the editorial decisions that we're making at CNN in
the same way that they've articulated that about CBS.
Speaker 17 (41:54):
Well, I don't look at Paramount Sky Garance is really
innovating anything, you know, like they're actually more like a
traditional studio than anybody else. And that's kind of why
Hollywood wants them to succeed, is because it just they
look like a traditional studio where Netflix is all about
innovation and they're the big disruptor in the industry that
most people in Hollywood don't like because of the way
(42:15):
the contract and the terms are. So if you're in
with Netflix, you like them. But if you're not in
with Netflix, you hate them. And that's Hollywood. So so
you know, once again, I think Paramount Skydance needs Warner
Brothers to really be a big player, and that's why.
Speaker 14 (42:30):
They're willing to pay so much. And politics isn't the
main driver.
Speaker 17 (42:34):
I'm not trying to say that it's a secondary driver
of this, but it's part of it, and I think
people need to keep that into consideration. And I still
think in the end, Paramount's going to do whatever's necessary
to get this.
Speaker 2 (42:44):
Deal in the time we have left.
Speaker 3 (42:46):
There is the Wall Street view on this, which thank
you you've given us the investor and Netflix investors take
the politics will put to one side. Actually more interesting
is the Hollywood view. How does this change Hollywood? And
one of the things that comes up on this program
is Netflix powerful algorithm combined with HBO Max Library. What
is your view on that as a Netflix shareholder?
Speaker 17 (43:08):
See, I don't see any value add on taking HBO
from the sense of like I need another streamer.
Speaker 14 (43:14):
You know, it's like they're not going to merge them.
Speaker 17 (43:17):
That doesn't make a lot of sense, and nor do
people want to pay fifty dollars a month for merged streamer,
even though they're paying that for two Seblis streamers.
Speaker 14 (43:24):
So I think for Netflix that's.
Speaker 17 (43:27):
Really about where do they go in live experiences and
what they're doing in the malls is really interesting with
the Netflix experiences that they've opened, like one in Philadelphia,
and then maybe they just completely changed the movie going
experience to being something much better than what it is now.
Speaker 14 (43:43):
So that's the Netflix show.
Speaker 3 (43:45):
Us movie theaters is what we wanted to get to.
We will have you back because this process is going
to go on for a very long time. Ross Gerber
of Goerba Kawasaki, thank you very much. That does it
for this edition of Bloomberg Tech. Towards the end of
the year.
Speaker 2 (44:00):
I cannot believe this new cycle character.
Speaker 4 (44:02):
I can't and we don't have enough to time and
bandwidth to get it all in in one hour, but.
Speaker 6 (44:06):
Boy do we try.
Speaker 4 (44:07):
And you've got to go and check it all out
again on our podcast.
Speaker 6 (44:10):
This is Bloomberg Tech