December 9, 2024 • 42 mins
Bloomberg's Caroline Hyde break downs China's probe into Nvidia as global tech tensions escalate and adding to broader geopolitical risks weighing on US tech stocks. Plus, Alphabet reveals new details on the progress it's made in quantum computing.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Bloomberg Audio Studios, Podcasts, radio news from Mahart where innovation,
money and power.
Speaker 2 (00:11):
Collie in Silicon Vallet Nbon.
Speaker 3 (00:14):
This is Bloomberg Technology with Caroline Hyde and Ed Ludlove.
Speaker 4 (00:32):
Live from New York. This is Blouebag Technology. Coming up.
Speaker 5 (00:34):
China opens a probe into Nvidia regarding anti monopoly laws.
Speaker 4 (00:38):
That's the global tech.
Speaker 5 (00:40):
War escalates this adding to broader geopolitical risk.
Speaker 4 (00:43):
Weighing on US tech stocks.
Speaker 5 (00:45):
We get the market impact and Alphabet reveals new details
on the progress it's made in quantum computing.
Speaker 4 (00:53):
But first we check in on Invidio. Top story for
the day.
Speaker 5 (00:56):
We're currently shaving off a cool one hundred billion dollars
in terms of market capitalization for what is the most
valuable company out there and listed in the United States,
down three percent.
Speaker 4 (01:05):
Why Moores about.
Speaker 5 (01:07):
An investigation, a probe being opened up by China about
a deal and acquisition.
Speaker 4 (01:11):
It made back in twenty twenty.
Speaker 5 (01:13):
But this really fits into a context of global tit
for tap between China and the US. Peter Elstrom joins us.
Now just go into the intricacies of this particular investigation
regarding a deal made years ago.
Speaker 6 (01:26):
Yeah, this was a bit of a surprise. We got
this news last night from state television in China, where
CCTV disclosed that the Agency SAMR has opened this probe
into Nvidia and what it calls potentially anti monopolistic behavior.
It's a little curious because, as you say, they're focused
in or They mentioned specifically a deal that they did
(01:47):
four years ago to buy a company called Melanox. It's
a company in Israel that makes networking equipment. China had
to approve that deal when Nvidia decided to do it,
and they did approve it, as you mentioned in twenty twenty,
but they did with seven different conditions about invidious behavior
going forward. That included giving the customers in China access
to product information from Melanox and product access for Invidia's
(02:11):
own chips. Now it's not exactly clear what they're going
to dig into in this case, but as you noted,
in Vidia has been this company at the heart of
the US China conflicts over technology. Washington has cut off
in Vidia's ability to sell its most advanced chips into China,
so they're really at the heart of this conflict between
the two biggest economies.
Speaker 7 (02:32):
In the world.
Speaker 5 (02:33):
Now it's worth noting that the US Justice Department itself
is looking for information to in video and potentially violating
anti trust laws. It's not the only country investigating in video,
but set the context of what other companies are being
focused in on on this tip for tap, because Micron,
for example, is having difficulty with its relationship with China.
This just does seem to be around the context of
(02:57):
China trying to be one up on what respond to
the US limitations of technology into its country.
Speaker 6 (03:04):
Yeah, as you say, the anti monopoly concerns around in
Vidia are quite broad. It's not just China, it's also
the United States. France here has taken a look at it.
That's because they have essentially a monopoly one hundred percent share,
almost one hundred percent share in the most advanced chips
to train AI models. But you're exactly right. Also, there's
this broader trade conflict that we've seen that creates all
(03:26):
sorts of problems for companies in between the US and China.
In Vidia still sells a lot of chips into China.
It had sold more in the past, but two years
ago we got this news that Washington was going to
cut off their ability to sell the most high end
AI chips used to train AI models. Those restrictions have
been steadily tightened since then, so in Vidia has suffered
because of that.
Speaker 4 (03:46):
But it's beyond that too.
Speaker 6 (03:48):
It's the chip equipment makers that have been restrained, ASML
here in Europe has been restrained from selling their most
advanced gear. So in response, Beijing has taken a number
of actions that we're seeing them retaliate in kinds of
ways where they cut off a couple of minerals that
are very important to the chip industry, in the defense industry,
in particular Gallium and Germanium. So they're beginning to show
(04:10):
that they also have some cards to play some lovers
the poll in this broader trade conflict.
Speaker 4 (04:15):
Peter Elstrom always bringing us the context.
Speaker 5 (04:17):
Thank you so much, and look, let's just stick with
China and the US. Because President elect Donald Trump said
he did have an exchange with Chinese leader Jijingping in
recent days, the first clear indication of direct contact between
the two men since Trump's reelection in November. Here he
is in an interview with NBC's Meet the Press with
Kristin Walker.
Speaker 4 (04:37):
Have a listen, I.
Speaker 3 (04:37):
Had any agreement with President she who I got along
with very well, we've had communication as recently as this week,
and I had communication with him where they were going
to give the death penalty to anybody sending drugs into
the United States.
Speaker 5 (04:52):
Let's get out to Mike Shepherd in Washington, and that
was notable. The fact that conversation has already started in
the environment where we're seeing actually China having to respond
to its own economic issues, particularly as they have to
look to what Trump might do to their economy come
twenty twenty five.
Speaker 8 (05:11):
Well highlights the uncertainty surrounding the US China relationship as
we enter this change over here in Washington. On the
one hand, Trump is coming in with very much a
hawkish agenda towards China when you look at his promises
during the campaign to impose tariffs of as much as
sixty percent on Chinese goods and his follow on promise
(05:34):
to put ten percent tariffs on Chinese goods over his
complaint is mentioned in that interview over the flow of
fentanel from China into this country. But at the same time,
Beijing is sort of looking to the change in administration
as possibly a way to engage with an incoming president
who has shown a tendency to be a little bit
(05:56):
more transactional. Maybe they can engage with him in a
way they have been able to with President Joe Biden
and his team. So they have signaled that. And in fact,
when you were talking just now with Peter about some
of the restrictions that China recently imposed on some of
those sensitive critical minerals like gallium and germanium, their statement
also included an overture to the US that we look
(06:19):
forward to engaging more positively at some point. So there
is sort of a we don't know which exactly which
way this is going to go, but the direction of travel,
based on the people in Trump's team, looks to be pretty.
Speaker 5 (06:32):
Tough, tough, and particularly around tariffs. Just take listen to
what he said in that particular interview on tariffs too.
Speaker 3 (06:39):
I can't guarantee anything. I can't guarantee tomorrow, but I
can say that if you look at my just pre COVID,
we had the greatest economy of the history of our country,
and I had a lot of tariffs on a lot
of different countries, but in particular China. We took in
hundreds of billions of dollars and we had no inflation.
Speaker 5 (07:00):
Finding that to consumer impact in terms of inflationary pressure.
Speaker 4 (07:04):
But businesses are eyeing this very closely.
Speaker 8 (07:06):
To Mike, they should be, because businesses really are consumers.
A lot of those goods that are coming in from
China as key inputs to their industries and key components,
and an increase in tariffs would not just hit consumers
at the end of the very end of the supply chain,
it would also hit businesses and they would have to
(07:26):
think about.
Speaker 6 (07:27):
Well, what do we do with our costs?
Speaker 8 (07:29):
Do we pass those along? And from that interview you
really get the sense that Trump is not inclined to
back away from any of those pledges. He was pressed
during the interview about the impact on consumers and he
insisted that no, there was no cost to consumers. He
was asked that he said so twice during the interview
(07:49):
that there is no cost, and he actually tried to
tout the benefits of it, and moreover, he actually pointed
to tariffs as a way to gain in other areas
when it comes to non economic policy as well. So
it'll be interesting to see how he approaches this not
only as an economic tool, but maybe as a way
to gain leverage in other policy areas, including immigration and
(08:12):
who knows what else.
Speaker 5 (08:13):
Who knows what else Mike sheppin, We appreciate it now,
all of this geopolitical risk, but we haven't even mentioned
Syria yet. There's also this week the likelihood of another
sticky inflation print here in the United States. All of
it is sending US tech stocks lower on the day, despite,
of course, what had been a bit of optimism around
stimulus potential from China.
Speaker 4 (08:32):
That's going to market take.
Speaker 5 (08:33):
Brian Cosman is with US portfolio manager at GQG.
Speaker 4 (08:37):
You'll read on the.
Speaker 5 (08:37):
Market trends, look any of these latest issues in video
China giving you pause, Brian.
Speaker 2 (08:44):
Yeah, So further starts in terms of technology. We did
have a very heavy waiting to a lot of technology
names earlier in the year about in terms of our
US secuity portfolio, more than fifty percent of the portfolio
was in it and also communications services types of names.
And we have seen a lot of hype kind of
build up in a lot of these technius. In recent months,
(09:05):
we've adjusted and gotten a little bit more of a
neutral posture at this point in time. So, you know,
coming back to in Vidia specifically, they do seem to
be still garnering the majority of economics within sort of
this capital cycle and its tech cycle that we're seeing
at this point in time. And I think you're seeing
that in terms of the power and their ability to
sort of you know, capture the market and sort of
(09:25):
you know, circle of wagons, if you will, in that sense.
With regards to China, they do have a slightly reduced
exposure to China then they've had in recent years and months,
and I think that's because you're just seeing the overall
strength in the other areas that they're selling through to
the hyperscalers and things like that. So China's now fifteen
sixteen percent of their revenue at this point in time.
So I think this is a little bit less of
an issue than maybe what we saw a couple of
(09:46):
years ago when you saw some of those US restrictions
sort of pulling back you know, in video's ability to
sell into China.
Speaker 5 (09:53):
So ultimately, the valuations when we're still seeing in video
and excess of three trillion dollars that's been dedicat by
the extend to the market opportunity despite some curtailment to
China for example.
Speaker 2 (10:05):
Yeah, so when I think you look at in video,
you look at sort of the hardware implication and sort
of the technology that they have, and they have a
better mouse trap in terms of the GPUs that they're
selling into folks, and I think you're seeing that in
terms of the strength of their business. But I think
goes underappreciation is the software element of the thesis as well.
That gives them a lot longer tail in terms of
the ability to continue to sell through that software angle.
(10:28):
The thesis pulls up the margin structure of the video
over the course of time. People talk about kudos or
being that base software layer, and then these software stacks
that they're building on top of this. This all creates
a very system wide approach and it's hard to sort
of disrupt. That creates really good economics for them, and
I think there's a plenty of hetero for them to
continue to grow on an ongoing basis. I think it
gives a little more stickiness than maybe some of their peers,
(10:50):
even within the GPU space.
Speaker 5 (10:52):
Interesting, Brian take us there for to software more broadly,
we have seen the shift out of perhaps chip stocks
semi conductors into the software space. Thinking how generative AI
is going to make their valuation suddenly ever more inflated.
Speaker 4 (11:04):
Is that the right way to go?
Speaker 2 (11:05):
Yeah, So, in terms of the software names, we have
seen so multiples really inflate and sort of expectations go up.
In terms of the ability for these folks to sort
of capture the economics from GENAI and things like that,
I think we're still in early stages. I think we
need to see where some of the impacts are going
and where folks are able to sort of monetize the
economics of how generative AI work. I think enterprise is
(11:27):
a little bit slower to adopt and it's going to
take some time to get there. Where we see the
impact more immediately is on the advertising space. So we
are seeing that by using sort of GENAI, using this
unstructured data to predict what people want to see, what
people want to sort of buy in terms of the
advertising and things along those lines, matching up that media
content with those advertising components. That's an area where we're
(11:49):
seeing real returns at this point in time. The economics
are actually going in that direction. So that's more of where
our focus has been on a your to day basis
and where we're positioned in the portfolio at the moment.
Speaker 5 (11:59):
Interesting of course, we've seen Meta outperform on the back
of that alphabet as well.
Speaker 4 (12:03):
Brian, What was so interesting about what you said.
Speaker 5 (12:05):
Earlier was that the fact that you have gone more neutral,
and I'm interested as to therefore are you putting allocation
into other AI tangential industry groups if it's not just
all about semiconductors and software right now.
Speaker 2 (12:17):
Yeah, So when I describe sort of the neutral posturing,
I think that's the overall sort of exposure that you've
seen in the portfolio has come down. We still have
higher conviction calls in those select names like an Nvidia,
like a Meta, even like a Microsoft, a little bit
lesser extent in our portfolio. You see some fairly decent
sizing in those anes. Lean knows is a tech show
and a tech program that we're talking about, But we
(12:37):
are seeing some more interesting areas and some of the
more boring areas of the economy as well. So consumer
staples to some extent utilities, not necessarily because they're defensive
by nature, but quite frankly because they're a little bit
or less trend. The multiples look more attractive and you're
getting a little bit more in a bank for your buck,
so to speak. In terms of the returns, you get
to a high single digit, low double digit type of
(12:58):
return with a divid and eel on top of it
for a reasonable multiple. You don't have to worry about
the higher So it's almost like a barbell approach, if
you will, having some of these punchier names, some of
these higher conviction names within the tech ecosystem where we
are seeing the economics coming through, and then you're seeing
some of these other things where we can get some
really good quality compounding at a reasonable price.
Speaker 4 (13:17):
Diversification key no matter who the audience is.
Speaker 5 (13:19):
Brian Kersman, we thank you, GQG portfolio manager there, appreciate it. Meanwhile,
let's just check in on super micro shares for a moment.
Jumping earlier in the training session, we're now up just
two percent.
Speaker 4 (13:31):
But all of this after the battle.
Speaker 5 (13:32):
Server makers said that the Nasdaq has granted the firm
more time to become compliant with listing rules. Ultimately, this
is all regarding whether or not they've got time to
serve their teen filings.
Speaker 4 (13:44):
Take whose filings.
Speaker 5 (13:45):
We're seeing that the company has now been given till
February the twenty fifth the file. It's financial report for
the fiscal year ended June thirtieth alphabet, and it's revealing
new details on the progress it's made in quantum computing.
The exponential advancements are being done with a new Willow
(14:08):
quantum chip that is helping overcome some of the biggest
challenges in the field. Now the next step is of
course coming up with real world use for all the
theoretical power. Here to talk about all of it, hartman
Naven here's the vice president of Engineering at Google as
well as a founder and manager of the Quantum Artificial
Intelligence Lab, And the biggest challenge thus far has been.
Speaker 4 (14:26):
Ultimately error correction. Can you just talk us through what
you have just achieved.
Speaker 9 (14:31):
Yeah, The biggest news for today is that we developed
a new very high quality quantum ship and this enabled
to break through computations. And one is an error correction
where we showed for the first time that as we
use more cubits, the error rate comes down, and it
(14:52):
comes down exponentially. The second important achievement breakthrough achievement, is
that we did benchmark computation took just five minutes on
our new chip that would take the top supercomputer ten
septillion years to perform, so ten septillion audience may not
(15:13):
know that's one with twenty five zero. So it's a
mind bogglingly long time and it.
Speaker 4 (15:19):
Took five minutes.
Speaker 5 (15:20):
What I want to understand is what this benchmark test
is benchmark random circuit sampling?
Speaker 4 (15:25):
What does this prove?
Speaker 9 (15:27):
Yes, So this benchmark problem we often say, it's not
a problem that people on mainstream have yet, but it's
a suitable benchmark to compare different quantum processors to each other,
or to compare quantum processors quantum chips against classical computers.
Speaker 5 (15:48):
What's so mind boggling about quantum more generally is that
we have been talking about it for decades, but it
always feels another five, another ten years off.
Speaker 4 (15:57):
Is it different this time? Hartlett?
Speaker 9 (16:00):
Yeah, Queen computing is definitely not that field. There's a
steady progress, and our new chips that we just released
a Willow will form the basis to start doing useful
computations that will enable or solve problems that humankind has.
Speaker 4 (16:21):
Let's talk about those problems.
Speaker 5 (16:23):
Then everyone immediately goes to healthcare drug discovery. Is that
the most obvious place with which this sort of computing
is going to be helpful.
Speaker 9 (16:32):
So the well understood killer application of Queen computers is
to model systems where quantum effects are important, and that
is more often the case than you may think. For example,
if you want to understand how drugs pharmaceuticals bind to cells,
or if you want to design a nuclear fusion reactor
(16:55):
or something as mundane as improving the batteries for an
electric are all these engineering challenges involve quantum problems, and
their quantum processes are just the tool of choice to
solve those problems.
Speaker 5 (17:10):
Put it in the perspective of generative AI, the hype
that our audience that we have had over the past
couple of years.
Speaker 4 (17:18):
What does quantum mean for that? Is it?
Speaker 5 (17:21):
Is it superior even in its effects versus generative AI,
and does itself the same issues in terms of energy
and infrastructure.
Speaker 9 (17:28):
I wouldn't compare it in those terms. AI as well
as quantum computing will be the most transformative technologies of
our time. But there's a lot of cross fertilization between
AI and quan computing. So AI can help work for
(17:49):
quantum ships to work better and vice versas. There are
many key computational tasks in AI where a quantum ship
will just be the tool of choice to handle laws.
Speaker 4 (18:01):
It's a phenomenal step.
Speaker 5 (18:02):
Please come back as you start to apply the significance
of this computing. Heartment name and he's the VP of
Engineering at Google and founder of the Quantum Malt Official
Intelligence Lab.
Speaker 4 (18:12):
Meanwhile, coming up, we're going to.
Speaker 5 (18:13):
Be discussing the limits of power grids and how they're
running up against the ambitions of both Big Tech and
Wall Street when it comes to generator of AI. What
we're just finishing on conversation there on. Meanwhile, just check
in on some key stories today. Comcast one of those,
off by seven percent, key drops on the SMP at
the moment and the Nasdaq shares sliding, and this after
(18:34):
the Comcast Cable president's CEO David Watson is saying is
projects broadband subscribing losses and more than one hundred thousand
in the fourth quarter, hurricans Helene and Milton causing around
ten thousand broad band losses. We're also looking at Vivendi
Big News out of France, as this company does look
to break itself apart, saying it's very likely though it
would drop out of the CAC forty after it did
(18:56):
indeed approve that three spin off listing December the sixth,
Steek investors approving that and plans spin off of three
multi billion euro units.
Speaker 4 (19:05):
This is bluembg technology time now for talking tech.
Speaker 5 (19:20):
First up, Hello Fresh, But it's facing allegations that one
of its facilities in Illinois used migrant teen labor chars
fell after ABC News reported it is under investigation by
the US Labor Department along with a staffing agency that
hired employees for the facility, plus Finance Holding CEO Richard
Teng told Bloomberg that quote, the future is bright for
crypto and encouraging other countries to follow the lead in
(19:41):
the US just.
Speaker 4 (19:41):
Take a listen.
Speaker 10 (19:43):
Yeah, Now, very crypto friendly president in the States and
extremely smart strategy appointing AI in crypto za on that
front because this tool represents the most innovative technology for
the future.
Speaker 5 (19:57):
And Jack Maher made a rare public appearance Sunday to
Mark and Group's twentieth anniversary, spoke optimistically about the future
of the company while acknowledging its challenges and also the
hype around the company's AI powered products. Let's just talk
about AI a little bit more because Blackstone has been
big data center investor and its ambitions are huge, but
not everyone is into it. After it bought data center
(20:21):
developer QTS in twenty twenty one, it's turned it into
a major profit driver for the firm.
Speaker 4 (20:26):
But around Fairttville, Georgia.
Speaker 5 (20:29):
The data center has stirred discord in the community over
its unprecedented power needs. No most Josh Saul is here
with more to deliver. This sort of infrastructure impacts real people,
real citizens', real ground.
Speaker 11 (20:39):
Yes, people in Fairville were unhappy when the local power
company came around and started knocking on the door saying
they wanted to buy easements. Basically wanted to run big
transmission lines through their yards. And that's because these data
centers take so much power. The one that QTS is
building in Georgia will take about as much power as
a million horns.
Speaker 5 (20:59):
Wow, what does this mean in terms of scaling that
across different states, different towns, Because there's not only just
one data center in.
Speaker 4 (21:05):
Georgia right now.
Speaker 11 (21:06):
Exactly, Georgia isn't even one of the hotspots. It's not.
It's not data center Alley, it's not it's not in Arizona.
So it means that there'll be more and more of
this more power infrastructure being built all over the US
and people unhappy when power lines runs run across. But
at the same time, data centers in AI we all
use all the time, so there's certainly no indication that
(21:28):
it's going to slow down.
Speaker 5 (21:29):
So is there a price point or is there just
money recompense what happens?
Speaker 11 (21:33):
It seems like pretty much all the systems are set
up for more data centers to get built faster and faster.
Power power companies have a lot of power sorry to
build transmission lines across in order to provide electricity, So
it seems like most of our systems and most of
the profit and in centers are set up to have more
data centers and more transmission lines at a pretty rapid
(21:55):
pace and.
Speaker 5 (21:56):
More discord perhaps among communities. It's great reporting. I urge
you to go and read the story in full.
Speaker 4 (22:10):
Welcome back to Blue Megtechnology and Karen Hide in New York.
Speaker 5 (22:13):
As the course of the show continues, we've actually seen
a build up in risk a version surrounding tech. At
the moment, we're off by seven tenths of a percent
all fixation on the CPI print that inflationary print here
on Wednesday in the United States. What does that mean
for FED policy going forward? What about the geopolitical risks
that are currently affecting in video for example, once again
a new probe being announced by China that tip for
tach continues to build anxiety.
Speaker 4 (22:34):
We're seeing though, Crypto off by about two percent.
Speaker 5 (22:36):
No, we're still in a ninety eight thousand level close
to that one hundred, but then too seeing a little
bit of a sell off as we get out of
those riskier assets. Move on to some of the individual
movers that I want to shine a light on. I mean,
PDD don't remember there is some optimism, particularly around in
China and its own monetary policy, fiscal policy stimulus, some
of the most positive that we've seen in more than
(22:57):
a decade. According to Morgan Stanley, we're up some eleven
percent as Chinese names do well off the.
Speaker 4 (23:01):
Hints coming from the politboro.
Speaker 5 (23:03):
We're looking at work Day up six percent this as
we get new announcements of who and who is in
and out of the next S.
Speaker 4 (23:09):
And P five hundred.
Speaker 5 (23:10):
It's a real and we're seeing Workday software company up
six percent because it will make the S and p
five hundred everyone, or rather I hope app love and
it's Investibatis thought it might, but it's down fourteen percent
as it doesn't make the cut. Meanwhile, let's just talk
about Reddit as well. Currently testing a new AI powered
chatbot called Reddit Answers to help users find information and
surface discussions from across the platform's forums. That's why I
(23:30):
am Bluemokes Asia Council. What's so interesting is I feel
as a user, you and I use Generator AI of Googles.
I often get Reddit citations and I'm being pushed to
Reddit more and more. But this is about keeping people
within the platform absolutely.
Speaker 12 (23:44):
Like you said, right, you search on Google and you
end up on Reddit. Sometimes I search Google specifically, so
if I'm trying to find la versus San Francisco, I
might search for that and then add Reddit at the end,
and that's how I find stuff on the site. Because
sometimes on Reddit site it can be a little bit
difficult to find things. They have like one hundred thousand communities,
billions of posts, and so with this new chatbot, you
(24:07):
can just search on the Reddit site and find all
the things that you're looking for. It or give you
a summary or give you links, and it just makes
the experience a lot easier.
Speaker 5 (24:14):
What's so interesting is Reddit, of course have been signing
deals with generative AI companies Google for example, and getting
money for its own website and data to be used
by these generative AI large language models.
Speaker 4 (24:27):
What then are they doing in terms of their own
lms to build its products? Are they agnostic? Are they
using one or a couple?
Speaker 12 (24:34):
They're using a few, so they have their own models.
They also use models from Open AI and Google. So
they have really interesting partnerships with these companies where they're
using their large language models. But then, like we mentioned,
if you go on Google and you find Reddit results,
that's because of a partnership that they have with Google
that allows Reddit posts to appear in Google searches. So
it's a mixture of all those different things. But they're
(24:55):
definitely building their own models.
Speaker 4 (24:57):
Internally, and then they're using some from some of these
other companies, becoming evan more integral to the business model
generator of aiish accounts. Great to have me, thank you.
Speaker 5 (25:05):
Now let's just stick with social media a little bit
more because TikTok's Chinese parent company, bite Dance now faces
of course a ban in the US next month as
a result of a federal appeals court decision Friday, But
the ruling has sparked opposition from civil liberties groups and users,
with only thirty two percent in fact of US adults
supporting the ban according to a recent Pew survey. Let's
bring in Annapam Chandra is professor of law at Georgetown University.
(25:29):
And what's so interesting is this is of course ban
or divest. You have written at length prior to this
talking about some of the legal issues, perhaps with the
government arguing, highlighting in particular, two key sentences in the
briefing to the DC circuit that you thought basically cast
doubt on the national security arguments.
Speaker 4 (25:47):
Didn't you just go through them for us?
Speaker 10 (25:49):
Sure?
Speaker 13 (25:50):
So, I think some of the key issues in the
case revolve around whether or not there is in fact
national security risk from the app, and the nationaleus risk,
as the government tells us, is that the Chinese government
might manipulate the app to push Chinese propaganda, or that
it might manipulate the app to surveil Americans. And in
(26:13):
order to avoid this, the United States government has negotiated
has been negotiated for the last handful of years with
TikTok to kind of locate the data entirely within Texas
essentially and also put all these controls over that data.
But the United States government has been dissatisfied with the
level of protection. But the US government says that it
(26:37):
has no information that there has in fact been any
Chinese government manipulation of the app. So it's all speculative,
and that's part of the difficulty for the government in
this case.
Speaker 5 (26:50):
But is potential and speculation enough ultimately to argue that
we're protecting citizens here exactly?
Speaker 13 (26:58):
The Circuit Court of Appeals believed it was that the
government had was uncomfortable, did not trust Bitedance, did not
trust TikTok, and therefore could move to ban it from
the country if it's so decided. And so I think
that is in fact the question now that we'll be
(27:20):
posed to the Supreme Court of the United States. Should
the Supreme Court take the case.
Speaker 5 (27:25):
Can you give us context ugus examples in the past
where we've really seen court side with a government in
this way around national security that even though it is
theoretical in nature, it's important enough to overcome any well
rights to free speech.
Speaker 13 (27:42):
So this is a very unusual case. There's nothing that
is on all fours with this case. The closest case
involving foreign propaganda in the United States is a case
from the nineteen sixties involving literally Chinese communist propaganda, and
the Supreme Court unanimously in that case said Americans have
a right to receive Chinese communist propaganda if they so want.
Speaker 8 (28:04):
So.
Speaker 13 (28:04):
There's not an ideal case. I think if cases around
national security, like the Pentagon Papers case, where the US
government said it's too dangerous for the New York Times
to reveal the secret history of the Vietnam War, and
the Supreme Court said, actually, no, we don't believe that
it's not a kind of immediate risk that you are
(28:25):
pointing to. And that's what the TikTok's argument is going
to be to the Supreme Court.
Speaker 5 (28:30):
Do you think that it's argument will carry weight if
indeed it does get to the level of the Supreme Court.
Speaker 13 (28:36):
Well, if it goes to the Supreme Court, I think
the same question is going to be before those judges
as was before the DC Circuit. Do you defer to
the judgment of the political branches of Congress and the
President when they insist that there is a national security
risk and there's no other way to protect Americans than
(28:59):
to ban the app.
Speaker 5 (29:01):
What's so interesting is it feels as though perhaps the
worries coming from the user base have diminished significantly, and
as we're just saying just a third of US users
or in the US population supports such a ban, So
is there context as to how that might need to change.
Speaker 4 (29:18):
I mean, you're no.
Speaker 5 (29:19):
Marketing expert, but it almost feels as though the government
needs to go out and make its case.
Speaker 4 (29:22):
A little louder.
Speaker 13 (29:24):
The government tried to make its case for the last
six months. So the ban has been in place, has
been you know, the law has been in place threatening
this ban since April, and so since that time, we've
seen the government say that China controls this app and
therefore it's unsafe. But Americans have continued using the app,
(29:46):
so they you know, Americans have decided that they don't
feel unsafe despite the government's use. In fact, Donald Trump
started using the app. In fact, Joe Biden started using
the app, Kamala Harris started using the app. Every politician
is now using the app, despite it's the fact that
it's apparently possibly prone to foreign manipulation, though the government
(30:10):
again has said we have no evidence of that thus far.
Speaker 4 (30:15):
January in the nineteenth.
Speaker 5 (30:16):
Therefore it gets banned, it has to be sold during
foresee that being the outcome.
Speaker 13 (30:22):
Here, I think that's becoming closer and closer to the
likely outcome that TikTok goes dark on January nineteenth.
Speaker 4 (30:32):
And it's great to have some time with you. Thank you.
Anamam Chandra is of Georgetown University Law Center.
Speaker 5 (30:38):
Meanwhile, let's talk about a network of Facebook accounts posting
more than forty one hundred political ads against Romania's pro
EU presidential candidate while also promoting far right figures. Digital
threat research groups reset tech and check First say the
posts were.
Speaker 4 (30:54):
Viewed by nearly two hundred million times.
Speaker 5 (30:56):
Facebook's parent company Meta did not comment on the report,
but says it has not seen any evidence of major
incidents on its platforms in Romania. Now coming up, Lily
Lyman underscore VC Managing Partner.
Speaker 4 (31:07):
Is here with a look VC funding where it's heading next.
In twenty twenty five, this Bloomberg technology.
Speaker 5 (31:22):
Investment strategies of major multi billion dollar firms and then
small specialized VC funds, Well, they're showing that venture the
asset class is kind of bierfurcating.
Speaker 4 (31:32):
Here heading into the new year.
Speaker 5 (31:34):
According to Linny Lyman, who is managing partner at Underscore VC,
joining us now from Boston and nearly the theory here
is that you're getting two very different strategies, not both
of particularly complementary.
Speaker 1 (31:45):
That's right, Carolin, It's great to be here. Thanks for
having me again. I'm Lily Lyman. I'm at Underscore BC.
We are a community driven early stage venor from BC
here in Boston. We invested the preceed and seed stages
in the world of B to B software. In your questions,
the right one. We're seeing the emergence of two very
different strategies in venture, and it's bifurcating the market in
(32:07):
a way that almost looks like a barbell. On one hand,
you have these big, multi billion dollar multi stage venture
firms that our megafirms. They're really sort of financial institutions
at this point, and I hear LPs refer to these
as venture beta. On the other side of the spectrum,
you have smaller funds, more specialized targeted strategies. That's where
we at Underscore play and it's because our that's our
(32:28):
belief of where you can still get venture alpha that
three to five even ten x return. What's happened over
the last eighteen months or so, though, is we're seeing
a heavy concentration of LP dollars going into the megafunds.
So eighty percent of capital rais last year went to
established managers of firm sizes over five hundred million, and
the representation of these firms is increased, so it's gone
(32:51):
from about two percent to twenty five percent in the
venture market. And that overrepresentations has a lot of implications
for LPs, for managers and for founders.
Speaker 5 (33:00):
Okay, let's talk about what means for onto school VC. Ultimately,
you're going to see more consolidation aman among the seed
the pre seed stage or actually do need the breadth
of more focused funds, and it's actually in consolidation that
you have in the middling.
Speaker 1 (33:16):
I think what it means is for I think we're
going to see, you know, more specialization at the earliest stages,
so firms like ours that you need to really lean
into differentiation and get to founders early and also back
perhaps non consensus, non not obvious both founders and ideas
because that's how you can capture true alpha. What's happening
with the multi stage strategy at the at the seed
(33:36):
stage is that they play more of an index game
and putting out a lot of a lot of checks
and then backing the ones that are that are winners.
Speaker 4 (33:44):
But the implications of that is that.
Speaker 1 (33:45):
They're not always very pre sensitive in that and it
can dilute the entire early stage class if you can't
know counter that and and really dig in and go
earliest and non consensus. So that's how we've been thinking
about it in order to counteract some of the implications
of the multi stage indexing model at seed stage.
Speaker 5 (34:02):
It kind of takes me back to when Softbad first unveiled.
There are one hundred billion dollar fund and everyone feeling
that the vision FuMB was just going to give enormous
amounts of money to individual founders or companies that sort
of drown out the rest of the competition. Is that
what happens or is it just more that it inflicts valuations.
What is a real impact on those very early stage bets.
Speaker 1 (34:26):
I think it can have multiple impacts. I think, on
one hand, the indexing can impact valuations for everyone, and
so it can actually dilute the returns across the board.
And but on the other hand, if you can get
in early, or if you can do you know, invest
in sort of non consensus, I think you still can
generate that.
Speaker 4 (34:44):
That real alpha.
Speaker 1 (34:45):
I think part of the implications are questions for founders
as well. I think we're seeing founders think a lot
about who do they bring around their table at which
stages of company building. You know, in the in the
market peak during the zerp era, there's a lot speed dating,
there's a lot of capital that went out of the
door very quickly, and then two years later we had
a much tougher environment and it was hard for founders.
(35:07):
And so I think we're seeing founders, both new founders
and sophisticated founders, think about, you know, what kind of
investments and investors do they want around their table for
each stage of company building, and thinking more about this
question of alignment. You know, what does that check mean
to that firm and how does that infer how they'll
show up during the company company building process.
Speaker 5 (35:27):
I'm assuming that most founders you're seeing are having to
lean into generative AI, even if they are not an
AI specific company.
Speaker 4 (35:34):
But what does that mean, in terms of trying to.
Speaker 5 (35:37):
Uncover the least covered areas or the particularly non consensus
builders and founders you want to find.
Speaker 1 (35:44):
We're certainly seeing a ton of opportunity in AI and
and you know, over thirty percent of venture capital this
past year went into AI companies, and we expect that
number to increase next year. But what we're seeing is
is companies that are AI native, both in their products
that they're offering, but so in terms of how they're
being built. I think one of the things that makes
this time particularly exciting to be an investor at the
(36:05):
earliest stages is companies are being built with a new model.
I think with a new capital structure you can do
a lot more things more efficiently. Engineering can be much
more efficient, sales and marketing can be much more efficient,
and so I think there's a mentality of you do
more with less, achieve more with less, and so in
many ways, you know, outside of a certain bucket of
types of investments like foundational models and infrastructure, I think
(36:28):
where we invest a lot of the applied AI vertical
AI applications, I actually think you can build pretty big
companies with a totally different capital structure given the increased
efficiencies of building with AI itself.
Speaker 4 (36:39):
How many of them are in Boston? Are you looking
across the US?
Speaker 1 (36:44):
We invest across the US, but we're very bullish on Boston.
So over half of our capital is deployed here here
in Boston. And it's because this is an ecosystem that
continuously produces the top talent in the world. It continuously
produces some of the top research and development the world,
and increasing you know, it consistently solves really hard problems,
so things like healthcare and biotech and life sciences and
(37:08):
financial services, and so cybersecurity and energy and climate. So
you know, these three factors are you know, are part
of what gets us very excited to invest in the
bus and ecosystem, particularly in this age of applied.
Speaker 5 (37:20):
Ailey Lyman, great to have you, managing partner of Underscore VC,
Wishing well for the rest of the year. Meanwhile, Elsewhere
and VC, a startup founded by the former lead researcher
of Open AI, is raised forty million dollars in funding.
Speaker 4 (37:33):
Waveforms AI.
Speaker 5 (37:35):
It makes AI audio software that improves verbal interactions with machines,
and is one of several AI companies developing.
Speaker 4 (37:41):
More human length voice features. Bloomberg's Rachel Mets is here
with more. It's come out of stealth.
Speaker 5 (37:46):
There's only five people there, but it's got quite a
head evaluation already.
Speaker 14 (37:50):
Yeah, this is It's still a pretty small company. They
don't have a demo to share publicly yet, but they
are working on software to improve the kind of interactions
that people have with computers when we're speaking to computers.
Speaker 5 (38:04):
And this is all around Alexis Kono, who was very
senior at open Ai. So I assume that even though
small and without a demo, people are betting on him
his expertise in his background.
Speaker 14 (38:16):
Yeah, he has quite a long tale of experience working
on voice computer human interactions and he was pretty instrumental
in helping Opening Eye build its advanced voice mode. This
is the more recent version of the voice mode that
works with chat GPT that is meant to feel more
lifelike than it had in the past, more feeling like
(38:38):
you're having a conversation with an actual person than with
a computer business model.
Speaker 5 (38:43):
I mean the fact that we don't have a demo,
but ultimately, how is this company going to be returning
value to its vcs?
Speaker 14 (38:49):
That is a great question. What they told me is
they are going to be releasing at some point products
that are both B to B and also consumer oriented.
So you might imagine something like an app perhaps that
would be for consumers from the company to kind of
give people a sense of what their technology is and
how it works. But then also perhaps they might work
(39:12):
with other companies to let other companies use their software
as well.
Speaker 4 (39:15):
And the San Francisco based yes, that is true. They
are based in San Francisco.
Speaker 5 (39:20):
Rachel Matz all things open ai and ex open ai.
Speaker 4 (39:23):
We appreciate it, thank you very much.
Speaker 5 (39:25):
Indeed, Oracle stock having its best year since nineteen ninety nine.
It's all thanks to the momentum of its cloud business.
We'll find out if that momentum will continue when the
company releases its earnings today after the closing bell.
Speaker 4 (39:45):
Luberg's Brodie Ford with a preview. I mean, what an
extraordinary chart.
Speaker 5 (39:48):
It's just done so well because people are now like, oh,
the cloud thing, it's actually relatively doing well compared to competitors.
Speaker 7 (39:54):
There were many years that people would make jokes to
me about Oracle's cloud about how we know, will this
ever actually be a serious player, And within the last
two years, the answer has become yes, right, And a
large part of that is because the AI workloads are
so demanding that the traditional giants like AWS and Azure
just don't have the capacity, right, And so it's become
(40:15):
that Oracle's cloud is seriously taken as the kind of
fourth hyperscaler, and it's been able to grow in such
an accelerated way that very few other software companies can
you know, compete with.
Speaker 5 (40:27):
It's got some I wouldn't call it a key Man risk,
but certainly I'm talking about Uber as a client, but
also TikTok, and we've been talking a lot about some
concerns around the deal around TikTok.
Speaker 7 (40:37):
Absolutely right. I mean, yeah, we saw the court ruling
that the there's increased odds that TikTok could go ahead
and go poof, But there is you know, at the
end of the day, if anybody buys TikTok, the odds
that you say, hey, let's go ahead and switch out
our cloud infrastructure. Nobody wants to do that, right, I mean,
so it is a risk for Oracle, But you know,
(40:58):
when TikTok was one of their only marque customers two
years ago for OCI, it was a bigger deal today,
they have a lot of customers, and I think the
risk isn't quite as acute as it once was.
Speaker 5 (41:08):
What about other things in infrastructure and cloud provision, i
mean software, other services.
Speaker 4 (41:13):
What else are they managing to do well on?
Speaker 12 (41:15):
Yeah.
Speaker 7 (41:15):
Well, one that's really interesting is that for a long
time they really tried to get you to run their
key iconic database software on prem and kind of control that.
And they finally said, you know what, you could run
it on Azure, you could run it on AWS. And
that's kind of been a pretty big exciting thing for
investors that you know, a lot of the world is
on AWS or Azure and they want to use the
(41:37):
Oracle database and so being able to kind of update
that customer segment, you know, breathe some growth into the
database segment, which you know dates back to the late seventies.
That is another thing that's been exciting folks about a Oracle.
Speaker 5 (41:50):
But a lot of it is probably priced in when
you're looking at an eighty percent run up and the shares.
So what do you tend to hear in these calls
post earnings.
Speaker 7 (41:58):
Oh god, we're building data centers across the world, you know,
no one's ever seen a data center build out like this, right,
I mean that's.
Speaker 4 (42:05):
The general energy informative. Yes, absolutely, I.
Speaker 7 (42:07):
Mean the amount of hype right now about the AI
fueled data center buildout is massive, right. You see you
sat pricing the shares. You know they are near an
all time high. And so the question is so much
that how quickly are they going to be able to
convert demand into actually offering these services, because you know
they just like Microsoft and Amazon, they can't get these
(42:29):
centers up quick enough. They can't get enough chips to
deal with the demand.
Speaker 5 (42:32):
Well, despite the run up, they've got twenty five buys,
fifteen holds and not a single cell rating.
Speaker 4 (42:37):
Bodie Ford, all things are coole.
Speaker 5 (42:39):
We'll be checking in on him tomorrow, I'm sure, after
those earning drop. But meanwhile, that does it for this
addition of Blomberg technology. You don't want to forget a
podcast find out on the terminal as well as online
on Apple, Spotify, and iHeart this is Bloomberg technology.
Bloomberg Audio Studios, Podcasts, radio news from Mahart where innovation,
money and power.
Speaker 2 (00:11):
Collie in Silicon Vallet Nbon.
Speaker 3 (00:14):
This is Bloomberg Technology with Caroline Hyde and Ed Ludlove.
Speaker 4 (00:32):
Live from New York. This is Blouebag Technology. Coming up.
Speaker 5 (00:34):
China opens a probe into Nvidia regarding anti monopoly laws.
Speaker 4 (00:38):
That's the global tech.
Speaker 5 (00:40):
War escalates this adding to broader geopolitical risk.
Speaker 4 (00:43):
Weighing on US tech stocks.
Speaker 5 (00:45):
We get the market impact and Alphabet reveals new details
on the progress it's made in quantum computing.
Speaker 4 (00:53):
But first we check in on Invidio. Top story for
the day.
Speaker 5 (00:56):
We're currently shaving off a cool one hundred billion dollars
in terms of market capitalization for what is the most
valuable company out there and listed in the United States,
down three percent.
Speaker 4 (01:05):
Why Moores about.
Speaker 5 (01:07):
An investigation, a probe being opened up by China about
a deal and acquisition.
Speaker 4 (01:11):
It made back in twenty twenty.
Speaker 5 (01:13):
But this really fits into a context of global tit
for tap between China and the US. Peter Elstrom joins us.
Now just go into the intricacies of this particular investigation
regarding a deal made years ago.
Speaker 6 (01:26):
Yeah, this was a bit of a surprise. We got
this news last night from state television in China, where
CCTV disclosed that the Agency SAMR has opened this probe
into Nvidia and what it calls potentially anti monopolistic behavior.
It's a little curious because, as you say, they're focused
in or They mentioned specifically a deal that they did
(01:47):
four years ago to buy a company called Melanox. It's
a company in Israel that makes networking equipment. China had
to approve that deal when Nvidia decided to do it,
and they did approve it, as you mentioned in twenty twenty,
but they did with seven different conditions about invidious behavior
going forward. That included giving the customers in China access
to product information from Melanox and product access for Invidia's
(02:11):
own chips. Now it's not exactly clear what they're going
to dig into in this case, but as you noted,
in Vidia has been this company at the heart of
the US China conflicts over technology. Washington has cut off
in Vidia's ability to sell its most advanced chips into China,
so they're really at the heart of this conflict between
the two biggest economies.
Speaker 7 (02:32):
In the world.
Speaker 5 (02:33):
Now it's worth noting that the US Justice Department itself
is looking for information to in video and potentially violating
anti trust laws. It's not the only country investigating in video,
but set the context of what other companies are being
focused in on on this tip for tap, because Micron,
for example, is having difficulty with its relationship with China.
This just does seem to be around the context of
(02:57):
China trying to be one up on what respond to
the US limitations of technology into its country.
Speaker 6 (03:04):
Yeah, as you say, the anti monopoly concerns around in
Vidia are quite broad. It's not just China, it's also
the United States. France here has taken a look at it.
That's because they have essentially a monopoly one hundred percent share,
almost one hundred percent share in the most advanced chips
to train AI models. But you're exactly right. Also, there's
this broader trade conflict that we've seen that creates all
(03:26):
sorts of problems for companies in between the US and China.
In Vidia still sells a lot of chips into China.
It had sold more in the past, but two years
ago we got this news that Washington was going to
cut off their ability to sell the most high end
AI chips used to train AI models. Those restrictions have
been steadily tightened since then, so in Vidia has suffered
because of that.
Speaker 4 (03:46):
But it's beyond that too.
Speaker 6 (03:48):
It's the chip equipment makers that have been restrained, ASML
here in Europe has been restrained from selling their most
advanced gear. So in response, Beijing has taken a number
of actions that we're seeing them retaliate in kinds of
ways where they cut off a couple of minerals that
are very important to the chip industry, in the defense industry,
in particular Gallium and Germanium. So they're beginning to show
(04:10):
that they also have some cards to play some lovers
the poll in this broader trade conflict.
Speaker 4 (04:15):
Peter Elstrom always bringing us the context.
Speaker 5 (04:17):
Thank you so much, and look, let's just stick with
China and the US. Because President elect Donald Trump said
he did have an exchange with Chinese leader Jijingping in
recent days, the first clear indication of direct contact between
the two men since Trump's reelection in November. Here he
is in an interview with NBC's Meet the Press with
Kristin Walker.
Speaker 4 (04:37):
Have a listen, I.
Speaker 3 (04:37):
Had any agreement with President she who I got along
with very well, we've had communication as recently as this week,
and I had communication with him where they were going
to give the death penalty to anybody sending drugs into
the United States.
Speaker 5 (04:52):
Let's get out to Mike Shepherd in Washington, and that
was notable. The fact that conversation has already started in
the environment where we're seeing actually China having to respond
to its own economic issues, particularly as they have to
look to what Trump might do to their economy come
twenty twenty five.
Speaker 8 (05:11):
Well highlights the uncertainty surrounding the US China relationship as
we enter this change over here in Washington. On the
one hand, Trump is coming in with very much a
hawkish agenda towards China when you look at his promises
during the campaign to impose tariffs of as much as
sixty percent on Chinese goods and his follow on promise
(05:34):
to put ten percent tariffs on Chinese goods over his
complaint is mentioned in that interview over the flow of
fentanel from China into this country. But at the same time,
Beijing is sort of looking to the change in administration
as possibly a way to engage with an incoming president
who has shown a tendency to be a little bit
(05:56):
more transactional. Maybe they can engage with him in a
way they have been able to with President Joe Biden
and his team. So they have signaled that. And in fact,
when you were talking just now with Peter about some
of the restrictions that China recently imposed on some of
those sensitive critical minerals like gallium and germanium, their statement
also included an overture to the US that we look
(06:19):
forward to engaging more positively at some point. So there
is sort of a we don't know which exactly which
way this is going to go, but the direction of travel,
based on the people in Trump's team, looks to be pretty.
Speaker 5 (06:32):
Tough, tough, and particularly around tariffs. Just take listen to
what he said in that particular interview on tariffs too.
Speaker 3 (06:39):
I can't guarantee anything. I can't guarantee tomorrow, but I
can say that if you look at my just pre COVID,
we had the greatest economy of the history of our country,
and I had a lot of tariffs on a lot
of different countries, but in particular China. We took in
hundreds of billions of dollars and we had no inflation.
Speaker 5 (07:00):
Finding that to consumer impact in terms of inflationary pressure.
Speaker 4 (07:04):
But businesses are eyeing this very closely.
Speaker 8 (07:06):
To Mike, they should be, because businesses really are consumers.
A lot of those goods that are coming in from
China as key inputs to their industries and key components,
and an increase in tariffs would not just hit consumers
at the end of the very end of the supply chain,
it would also hit businesses and they would have to
(07:26):
think about.
Speaker 6 (07:27):
Well, what do we do with our costs?
Speaker 8 (07:29):
Do we pass those along? And from that interview you
really get the sense that Trump is not inclined to
back away from any of those pledges. He was pressed
during the interview about the impact on consumers and he
insisted that no, there was no cost to consumers. He
was asked that he said so twice during the interview
(07:49):
that there is no cost, and he actually tried to
tout the benefits of it, and moreover, he actually pointed
to tariffs as a way to gain in other areas
when it comes to non economic policy as well. So
it'll be interesting to see how he approaches this not
only as an economic tool, but maybe as a way
to gain leverage in other policy areas, including immigration and
(08:12):
who knows what else.
Speaker 5 (08:13):
Who knows what else Mike sheppin, We appreciate it now,
all of this geopolitical risk, but we haven't even mentioned
Syria yet. There's also this week the likelihood of another
sticky inflation print here in the United States. All of
it is sending US tech stocks lower on the day, despite,
of course, what had been a bit of optimism around
stimulus potential from China.
Speaker 4 (08:32):
That's going to market take.
Speaker 5 (08:33):
Brian Cosman is with US portfolio manager at GQG.
Speaker 4 (08:37):
You'll read on the.
Speaker 5 (08:37):
Market trends, look any of these latest issues in video
China giving you pause, Brian.
Speaker 2 (08:44):
Yeah, So further starts in terms of technology. We did
have a very heavy waiting to a lot of technology
names earlier in the year about in terms of our
US secuity portfolio, more than fifty percent of the portfolio
was in it and also communications services types of names.
And we have seen a lot of hype kind of
build up in a lot of these technius. In recent months,
(09:05):
we've adjusted and gotten a little bit more of a
neutral posture at this point in time. So, you know,
coming back to in Vidia specifically, they do seem to
be still garnering the majority of economics within sort of
this capital cycle and its tech cycle that we're seeing
at this point in time. And I think you're seeing
that in terms of the power and their ability to
sort of you know, capture the market and sort of
(09:25):
you know, circle of wagons, if you will, in that sense.
With regards to China, they do have a slightly reduced
exposure to China then they've had in recent years and months,
and I think that's because you're just seeing the overall
strength in the other areas that they're selling through to
the hyperscalers and things like that. So China's now fifteen
sixteen percent of their revenue at this point in time.
So I think this is a little bit less of
an issue than maybe what we saw a couple of
(09:46):
years ago when you saw some of those US restrictions
sort of pulling back you know, in video's ability to
sell into China.
Speaker 5 (09:53):
So ultimately, the valuations when we're still seeing in video
and excess of three trillion dollars that's been dedicat by
the extend to the market opportunity despite some curtailment to
China for example.
Speaker 2 (10:05):
Yeah, so when I think you look at in video,
you look at sort of the hardware implication and sort
of the technology that they have, and they have a
better mouse trap in terms of the GPUs that they're
selling into folks, and I think you're seeing that in
terms of the strength of their business. But I think
goes underappreciation is the software element of the thesis as well.
That gives them a lot longer tail in terms of
the ability to continue to sell through that software angle.
(10:28):
The thesis pulls up the margin structure of the video
over the course of time. People talk about kudos or
being that base software layer, and then these software stacks
that they're building on top of this. This all creates
a very system wide approach and it's hard to sort
of disrupt. That creates really good economics for them, and
I think there's a plenty of hetero for them to
continue to grow on an ongoing basis. I think it
gives a little more stickiness than maybe some of their peers,
(10:50):
even within the GPU space.
Speaker 5 (10:52):
Interesting, Brian take us there for to software more broadly,
we have seen the shift out of perhaps chip stocks
semi conductors into the software space. Thinking how generative AI
is going to make their valuation suddenly ever more inflated.
Speaker 4 (11:04):
Is that the right way to go?
Speaker 2 (11:05):
Yeah, So, in terms of the software names, we have
seen so multiples really inflate and sort of expectations go up.
In terms of the ability for these folks to sort
of capture the economics from GENAI and things like that,
I think we're still in early stages. I think we
need to see where some of the impacts are going
and where folks are able to sort of monetize the
economics of how generative AI work. I think enterprise is
(11:27):
a little bit slower to adopt and it's going to
take some time to get there. Where we see the
impact more immediately is on the advertising space. So we
are seeing that by using sort of GENAI, using this
unstructured data to predict what people want to see, what
people want to sort of buy in terms of the
advertising and things along those lines, matching up that media
content with those advertising components. That's an area where we're
(11:49):
seeing real returns at this point in time. The economics
are actually going in that direction. So that's more of where
our focus has been on a your to day basis
and where we're positioned in the portfolio at the moment.
Speaker 5 (11:59):
Interesting of course, we've seen Meta outperform on the back
of that alphabet as well.
Speaker 4 (12:03):
Brian, What was so interesting about what you said.
Speaker 5 (12:05):
Earlier was that the fact that you have gone more neutral,
and I'm interested as to therefore are you putting allocation
into other AI tangential industry groups if it's not just
all about semiconductors and software right now.
Speaker 2 (12:17):
Yeah, So when I describe sort of the neutral posturing,
I think that's the overall sort of exposure that you've
seen in the portfolio has come down. We still have
higher conviction calls in those select names like an Nvidia,
like a Meta, even like a Microsoft, a little bit
lesser extent in our portfolio. You see some fairly decent
sizing in those anes. Lean knows is a tech show
and a tech program that we're talking about, But we
(12:37):
are seeing some more interesting areas and some of the
more boring areas of the economy as well. So consumer
staples to some extent utilities, not necessarily because they're defensive
by nature, but quite frankly because they're a little bit
or less trend. The multiples look more attractive and you're
getting a little bit more in a bank for your buck,
so to speak. In terms of the returns, you get
to a high single digit, low double digit type of
(12:58):
return with a divid and eel on top of it
for a reasonable multiple. You don't have to worry about
the higher So it's almost like a barbell approach, if
you will, having some of these punchier names, some of
these higher conviction names within the tech ecosystem where we
are seeing the economics coming through, and then you're seeing
some of these other things where we can get some
really good quality compounding at a reasonable price.
Speaker 4 (13:17):
Diversification key no matter who the audience is.
Speaker 5 (13:19):
Brian Kersman, we thank you, GQG portfolio manager there, appreciate it. Meanwhile,
let's just check in on super micro shares for a moment.
Jumping earlier in the training session, we're now up just
two percent.
Speaker 4 (13:31):
But all of this after the battle.
Speaker 5 (13:32):
Server makers said that the Nasdaq has granted the firm
more time to become compliant with listing rules. Ultimately, this
is all regarding whether or not they've got time to
serve their teen filings.
Speaker 4 (13:44):
Take whose filings.
Speaker 5 (13:45):
We're seeing that the company has now been given till
February the twenty fifth the file. It's financial report for
the fiscal year ended June thirtieth alphabet, and it's revealing
new details on the progress it's made in quantum computing.
The exponential advancements are being done with a new Willow
(14:08):
quantum chip that is helping overcome some of the biggest
challenges in the field. Now the next step is of
course coming up with real world use for all the
theoretical power. Here to talk about all of it, hartman
Naven here's the vice president of Engineering at Google as
well as a founder and manager of the Quantum Artificial
Intelligence Lab, And the biggest challenge thus far has been.
Speaker 4 (14:26):
Ultimately error correction. Can you just talk us through what
you have just achieved.
Speaker 9 (14:31):
Yeah, The biggest news for today is that we developed
a new very high quality quantum ship and this enabled
to break through computations. And one is an error correction
where we showed for the first time that as we
use more cubits, the error rate comes down, and it
(14:52):
comes down exponentially. The second important achievement breakthrough achievement, is
that we did benchmark computation took just five minutes on
our new chip that would take the top supercomputer ten
septillion years to perform, so ten septillion audience may not
(15:13):
know that's one with twenty five zero. So it's a
mind bogglingly long time and it.
Speaker 4 (15:19):
Took five minutes.
Speaker 5 (15:20):
What I want to understand is what this benchmark test
is benchmark random circuit sampling?
Speaker 4 (15:25):
What does this prove?
Speaker 9 (15:27):
Yes, So this benchmark problem we often say, it's not
a problem that people on mainstream have yet, but it's
a suitable benchmark to compare different quantum processors to each other,
or to compare quantum processors quantum chips against classical computers.
Speaker 5 (15:48):
What's so mind boggling about quantum more generally is that
we have been talking about it for decades, but it
always feels another five, another ten years off.
Speaker 4 (15:57):
Is it different this time? Hartlett?
Speaker 9 (16:00):
Yeah, Queen computing is definitely not that field. There's a
steady progress, and our new chips that we just released
a Willow will form the basis to start doing useful
computations that will enable or solve problems that humankind has.
Speaker 4 (16:21):
Let's talk about those problems.
Speaker 5 (16:23):
Then everyone immediately goes to healthcare drug discovery. Is that
the most obvious place with which this sort of computing
is going to be helpful.
Speaker 9 (16:32):
So the well understood killer application of Queen computers is
to model systems where quantum effects are important, and that
is more often the case than you may think. For example,
if you want to understand how drugs pharmaceuticals bind to cells,
or if you want to design a nuclear fusion reactor
(16:55):
or something as mundane as improving the batteries for an
electric are all these engineering challenges involve quantum problems, and
their quantum processes are just the tool of choice to
solve those problems.
Speaker 5 (17:10):
Put it in the perspective of generative AI, the hype
that our audience that we have had over the past
couple of years.
Speaker 4 (17:18):
What does quantum mean for that? Is it?
Speaker 5 (17:21):
Is it superior even in its effects versus generative AI,
and does itself the same issues in terms of energy
and infrastructure.
Speaker 9 (17:28):
I wouldn't compare it in those terms. AI as well
as quantum computing will be the most transformative technologies of
our time. But there's a lot of cross fertilization between
AI and quan computing. So AI can help work for
(17:49):
quantum ships to work better and vice versas. There are
many key computational tasks in AI where a quantum ship
will just be the tool of choice to handle laws.
Speaker 4 (18:01):
It's a phenomenal step.
Speaker 5 (18:02):
Please come back as you start to apply the significance
of this computing. Heartment name and he's the VP of
Engineering at Google and founder of the Quantum Malt Official
Intelligence Lab.
Speaker 4 (18:12):
Meanwhile, coming up, we're going to.
Speaker 5 (18:13):
Be discussing the limits of power grids and how they're
running up against the ambitions of both Big Tech and
Wall Street when it comes to generator of AI. What
we're just finishing on conversation there on. Meanwhile, just check
in on some key stories today. Comcast one of those,
off by seven percent, key drops on the SMP at
the moment and the Nasdaq shares sliding, and this after
(18:34):
the Comcast Cable president's CEO David Watson is saying is
projects broadband subscribing losses and more than one hundred thousand
in the fourth quarter, hurricans Helene and Milton causing around
ten thousand broad band losses. We're also looking at Vivendi
Big News out of France, as this company does look
to break itself apart, saying it's very likely though it
would drop out of the CAC forty after it did
(18:56):
indeed approve that three spin off listing December the sixth,
Steek investors approving that and plans spin off of three
multi billion euro units.
Speaker 4 (19:05):
This is bluembg technology time now for talking tech.
Speaker 5 (19:20):
First up, Hello Fresh, But it's facing allegations that one
of its facilities in Illinois used migrant teen labor chars
fell after ABC News reported it is under investigation by
the US Labor Department along with a staffing agency that
hired employees for the facility, plus Finance Holding CEO Richard
Teng told Bloomberg that quote, the future is bright for
crypto and encouraging other countries to follow the lead in
(19:41):
the US just.
Speaker 4 (19:41):
Take a listen.
Speaker 10 (19:43):
Yeah, Now, very crypto friendly president in the States and
extremely smart strategy appointing AI in crypto za on that
front because this tool represents the most innovative technology for
the future.
Speaker 5 (19:57):
And Jack Maher made a rare public appearance Sunday to
Mark and Group's twentieth anniversary, spoke optimistically about the future
of the company while acknowledging its challenges and also the
hype around the company's AI powered products. Let's just talk
about AI a little bit more because Blackstone has been
big data center investor and its ambitions are huge, but
not everyone is into it. After it bought data center
(20:21):
developer QTS in twenty twenty one, it's turned it into
a major profit driver for the firm.
Speaker 4 (20:26):
But around Fairttville, Georgia.
Speaker 5 (20:29):
The data center has stirred discord in the community over
its unprecedented power needs. No most Josh Saul is here
with more to deliver. This sort of infrastructure impacts real people,
real citizens', real ground.
Speaker 11 (20:39):
Yes, people in Fairville were unhappy when the local power
company came around and started knocking on the door saying
they wanted to buy easements. Basically wanted to run big
transmission lines through their yards. And that's because these data
centers take so much power. The one that QTS is
building in Georgia will take about as much power as
a million horns.
Speaker 5 (20:59):
Wow, what does this mean in terms of scaling that
across different states, different towns, Because there's not only just
one data center in.
Speaker 4 (21:05):
Georgia right now.
Speaker 11 (21:06):
Exactly, Georgia isn't even one of the hotspots. It's not.
It's not data center Alley, it's not it's not in Arizona.
So it means that there'll be more and more of
this more power infrastructure being built all over the US
and people unhappy when power lines runs run across. But
at the same time, data centers in AI we all
use all the time, so there's certainly no indication that
(21:28):
it's going to slow down.
Speaker 5 (21:29):
So is there a price point or is there just
money recompense what happens?
Speaker 11 (21:33):
It seems like pretty much all the systems are set
up for more data centers to get built faster and faster.
Power power companies have a lot of power sorry to
build transmission lines across in order to provide electricity, So
it seems like most of our systems and most of
the profit and in centers are set up to have more
data centers and more transmission lines at a pretty rapid
(21:55):
pace and.
Speaker 5 (21:56):
More discord perhaps among communities. It's great reporting. I urge
you to go and read the story in full.
Speaker 4 (22:10):
Welcome back to Blue Megtechnology and Karen Hide in New York.
Speaker 5 (22:13):
As the course of the show continues, we've actually seen
a build up in risk a version surrounding tech. At
the moment, we're off by seven tenths of a percent
all fixation on the CPI print that inflationary print here
on Wednesday in the United States. What does that mean
for FED policy going forward? What about the geopolitical risks
that are currently affecting in video for example, once again
a new probe being announced by China that tip for
tach continues to build anxiety.
Speaker 4 (22:34):
We're seeing though, Crypto off by about two percent.
Speaker 5 (22:36):
No, we're still in a ninety eight thousand level close
to that one hundred, but then too seeing a little
bit of a sell off as we get out of
those riskier assets. Move on to some of the individual
movers that I want to shine a light on. I mean,
PDD don't remember there is some optimism, particularly around in
China and its own monetary policy, fiscal policy stimulus, some
of the most positive that we've seen in more than
(22:57):
a decade. According to Morgan Stanley, we're up some eleven
percent as Chinese names do well off the.
Speaker 4 (23:01):
Hints coming from the politboro.
Speaker 5 (23:03):
We're looking at work Day up six percent this as
we get new announcements of who and who is in
and out of the next S.
Speaker 4 (23:09):
And P five hundred.
Speaker 5 (23:10):
It's a real and we're seeing Workday software company up
six percent because it will make the S and p
five hundred everyone, or rather I hope app love and
it's Investibatis thought it might, but it's down fourteen percent
as it doesn't make the cut. Meanwhile, let's just talk
about Reddit as well. Currently testing a new AI powered
chatbot called Reddit Answers to help users find information and
surface discussions from across the platform's forums. That's why I
(23:30):
am Bluemokes Asia Council. What's so interesting is I feel
as a user, you and I use Generator AI of Googles.
I often get Reddit citations and I'm being pushed to
Reddit more and more. But this is about keeping people
within the platform absolutely.
Speaker 12 (23:44):
Like you said, right, you search on Google and you
end up on Reddit. Sometimes I search Google specifically, so
if I'm trying to find la versus San Francisco, I
might search for that and then add Reddit at the end,
and that's how I find stuff on the site. Because
sometimes on Reddit site it can be a little bit
difficult to find things. They have like one hundred thousand communities,
billions of posts, and so with this new chatbot, you
(24:07):
can just search on the Reddit site and find all
the things that you're looking for. It or give you
a summary or give you links, and it just makes
the experience a lot easier.
Speaker 5 (24:14):
What's so interesting is Reddit, of course have been signing
deals with generative AI companies Google for example, and getting
money for its own website and data to be used
by these generative AI large language models.
Speaker 4 (24:27):
What then are they doing in terms of their own
lms to build its products? Are they agnostic? Are they
using one or a couple?
Speaker 12 (24:34):
They're using a few, so they have their own models.
They also use models from Open AI and Google. So
they have really interesting partnerships with these companies where they're
using their large language models. But then, like we mentioned,
if you go on Google and you find Reddit results,
that's because of a partnership that they have with Google
that allows Reddit posts to appear in Google searches. So
it's a mixture of all those different things. But they're
(24:55):
definitely building their own models.
Speaker 4 (24:57):
Internally, and then they're using some from some of these
other companies, becoming evan more integral to the business model
generator of aiish accounts. Great to have me, thank you.
Speaker 5 (25:05):
Now let's just stick with social media a little bit
more because TikTok's Chinese parent company, bite Dance now faces
of course a ban in the US next month as
a result of a federal appeals court decision Friday, But
the ruling has sparked opposition from civil liberties groups and users,
with only thirty two percent in fact of US adults
supporting the ban according to a recent Pew survey. Let's
bring in Annapam Chandra is professor of law at Georgetown University.
(25:29):
And what's so interesting is this is of course ban
or divest. You have written at length prior to this
talking about some of the legal issues, perhaps with the
government arguing, highlighting in particular, two key sentences in the
briefing to the DC circuit that you thought basically cast
doubt on the national security arguments.
Speaker 4 (25:47):
Didn't you just go through them for us?
Speaker 10 (25:49):
Sure?
Speaker 13 (25:50):
So, I think some of the key issues in the
case revolve around whether or not there is in fact
national security risk from the app, and the nationaleus risk,
as the government tells us, is that the Chinese government
might manipulate the app to push Chinese propaganda, or that
it might manipulate the app to surveil Americans. And in
(26:13):
order to avoid this, the United States government has negotiated
has been negotiated for the last handful of years with
TikTok to kind of locate the data entirely within Texas
essentially and also put all these controls over that data.
But the United States government has been dissatisfied with the
level of protection. But the US government says that it
(26:37):
has no information that there has in fact been any
Chinese government manipulation of the app. So it's all speculative,
and that's part of the difficulty for the government in
this case.
Speaker 5 (26:50):
But is potential and speculation enough ultimately to argue that
we're protecting citizens here exactly?
Speaker 13 (26:58):
The Circuit Court of Appeals believed it was that the
government had was uncomfortable, did not trust Bitedance, did not
trust TikTok, and therefore could move to ban it from
the country if it's so decided. And so I think
that is in fact the question now that we'll be
(27:20):
posed to the Supreme Court of the United States. Should
the Supreme Court take the case.
Speaker 5 (27:25):
Can you give us context ugus examples in the past
where we've really seen court side with a government in
this way around national security that even though it is
theoretical in nature, it's important enough to overcome any well
rights to free speech.
Speaker 13 (27:42):
So this is a very unusual case. There's nothing that
is on all fours with this case. The closest case
involving foreign propaganda in the United States is a case
from the nineteen sixties involving literally Chinese communist propaganda, and
the Supreme Court unanimously in that case said Americans have
a right to receive Chinese communist propaganda if they so want.
Speaker 8 (28:04):
So.
Speaker 13 (28:04):
There's not an ideal case. I think if cases around
national security, like the Pentagon Papers case, where the US
government said it's too dangerous for the New York Times
to reveal the secret history of the Vietnam War, and
the Supreme Court said, actually, no, we don't believe that
it's not a kind of immediate risk that you are
(28:25):
pointing to. And that's what the TikTok's argument is going
to be to the Supreme Court.
Speaker 5 (28:30):
Do you think that it's argument will carry weight if
indeed it does get to the level of the Supreme Court.
Speaker 13 (28:36):
Well, if it goes to the Supreme Court, I think
the same question is going to be before those judges
as was before the DC Circuit. Do you defer to
the judgment of the political branches of Congress and the
President when they insist that there is a national security
risk and there's no other way to protect Americans than
(28:59):
to ban the app.
Speaker 5 (29:01):
What's so interesting is it feels as though perhaps the
worries coming from the user base have diminished significantly, and
as we're just saying just a third of US users
or in the US population supports such a ban, So
is there context as to how that might need to change.
Speaker 4 (29:18):
I mean, you're no.
Speaker 5 (29:19):
Marketing expert, but it almost feels as though the government
needs to go out and make its case.
Speaker 4 (29:22):
A little louder.
Speaker 13 (29:24):
The government tried to make its case for the last
six months. So the ban has been in place, has
been you know, the law has been in place threatening
this ban since April, and so since that time, we've
seen the government say that China controls this app and
therefore it's unsafe. But Americans have continued using the app,
(29:46):
so they you know, Americans have decided that they don't
feel unsafe despite the government's use. In fact, Donald Trump
started using the app. In fact, Joe Biden started using
the app, Kamala Harris started using the app. Every politician
is now using the app, despite it's the fact that
it's apparently possibly prone to foreign manipulation, though the government
(30:10):
again has said we have no evidence of that thus far.
Speaker 4 (30:15):
January in the nineteenth.
Speaker 5 (30:16):
Therefore it gets banned, it has to be sold during
foresee that being the outcome.
Speaker 13 (30:22):
Here, I think that's becoming closer and closer to the
likely outcome that TikTok goes dark on January nineteenth.
Speaker 4 (30:32):
And it's great to have some time with you. Thank you.
Anamam Chandra is of Georgetown University Law Center.
Speaker 5 (30:38):
Meanwhile, let's talk about a network of Facebook accounts posting
more than forty one hundred political ads against Romania's pro
EU presidential candidate while also promoting far right figures. Digital
threat research groups reset tech and check First say the
posts were.
Speaker 4 (30:54):
Viewed by nearly two hundred million times.
Speaker 5 (30:56):
Facebook's parent company Meta did not comment on the report,
but says it has not seen any evidence of major
incidents on its platforms in Romania. Now coming up, Lily
Lyman underscore VC Managing Partner.
Speaker 4 (31:07):
Is here with a look VC funding where it's heading next.
In twenty twenty five, this Bloomberg technology.
Speaker 5 (31:22):
Investment strategies of major multi billion dollar firms and then
small specialized VC funds, Well, they're showing that venture the
asset class is kind of bierfurcating.
Speaker 4 (31:32):
Here heading into the new year.
Speaker 5 (31:34):
According to Linny Lyman, who is managing partner at Underscore VC,
joining us now from Boston and nearly the theory here
is that you're getting two very different strategies, not both
of particularly complementary.
Speaker 1 (31:45):
That's right, Carolin, It's great to be here. Thanks for
having me again. I'm Lily Lyman. I'm at Underscore BC.
We are a community driven early stage venor from BC
here in Boston. We invested the preceed and seed stages
in the world of B to B software. In your questions,
the right one. We're seeing the emergence of two very
different strategies in venture, and it's bifurcating the market in
(32:07):
a way that almost looks like a barbell. On one hand,
you have these big, multi billion dollar multi stage venture
firms that our megafirms. They're really sort of financial institutions
at this point, and I hear LPs refer to these
as venture beta. On the other side of the spectrum,
you have smaller funds, more specialized targeted strategies. That's where
we at Underscore play and it's because our that's our
(32:28):
belief of where you can still get venture alpha that
three to five even ten x return. What's happened over
the last eighteen months or so, though, is we're seeing
a heavy concentration of LP dollars going into the megafunds.
So eighty percent of capital rais last year went to
established managers of firm sizes over five hundred million, and
the representation of these firms is increased, so it's gone
(32:51):
from about two percent to twenty five percent in the
venture market. And that overrepresentations has a lot of implications
for LPs, for managers and for founders.
Speaker 5 (33:00):
Okay, let's talk about what means for onto school VC. Ultimately,
you're going to see more consolidation aman among the seed
the pre seed stage or actually do need the breadth
of more focused funds, and it's actually in consolidation that
you have in the middling.
Speaker 1 (33:16):
I think what it means is for I think we're
going to see, you know, more specialization at the earliest stages,
so firms like ours that you need to really lean
into differentiation and get to founders early and also back
perhaps non consensus, non not obvious both founders and ideas
because that's how you can capture true alpha. What's happening
with the multi stage strategy at the at the seed
(33:36):
stage is that they play more of an index game
and putting out a lot of a lot of checks
and then backing the ones that are that are winners.
Speaker 4 (33:44):
But the implications of that is that.
Speaker 1 (33:45):
They're not always very pre sensitive in that and it
can dilute the entire early stage class if you can't
know counter that and and really dig in and go
earliest and non consensus. So that's how we've been thinking
about it in order to counteract some of the implications
of the multi stage indexing model at seed stage.
Speaker 5 (34:02):
It kind of takes me back to when Softbad first unveiled.
There are one hundred billion dollar fund and everyone feeling
that the vision FuMB was just going to give enormous
amounts of money to individual founders or companies that sort
of drown out the rest of the competition. Is that
what happens or is it just more that it inflicts valuations.
What is a real impact on those very early stage bets.
Speaker 1 (34:26):
I think it can have multiple impacts. I think, on
one hand, the indexing can impact valuations for everyone, and
so it can actually dilute the returns across the board.
And but on the other hand, if you can get
in early, or if you can do you know, invest
in sort of non consensus, I think you still can
generate that.
Speaker 4 (34:44):
That real alpha.
Speaker 1 (34:45):
I think part of the implications are questions for founders
as well. I think we're seeing founders think a lot
about who do they bring around their table at which
stages of company building. You know, in the in the
market peak during the zerp era, there's a lot speed dating,
there's a lot of capital that went out of the
door very quickly, and then two years later we had
a much tougher environment and it was hard for founders.
(35:07):
And so I think we're seeing founders, both new founders
and sophisticated founders, think about, you know, what kind of
investments and investors do they want around their table for
each stage of company building, and thinking more about this
question of alignment. You know, what does that check mean
to that firm and how does that infer how they'll
show up during the company company building process.
Speaker 5 (35:27):
I'm assuming that most founders you're seeing are having to
lean into generative AI, even if they are not an
AI specific company.
Speaker 4 (35:34):
But what does that mean, in terms of trying to.
Speaker 5 (35:37):
Uncover the least covered areas or the particularly non consensus
builders and founders you want to find.
Speaker 1 (35:44):
We're certainly seeing a ton of opportunity in AI and
and you know, over thirty percent of venture capital this
past year went into AI companies, and we expect that
number to increase next year. But what we're seeing is
is companies that are AI native, both in their products
that they're offering, but so in terms of how they're
being built. I think one of the things that makes
this time particularly exciting to be an investor at the
(36:05):
earliest stages is companies are being built with a new model.
I think with a new capital structure you can do
a lot more things more efficiently. Engineering can be much
more efficient, sales and marketing can be much more efficient,
and so I think there's a mentality of you do
more with less, achieve more with less, and so in
many ways, you know, outside of a certain bucket of
types of investments like foundational models and infrastructure, I think
(36:28):
where we invest a lot of the applied AI vertical
AI applications, I actually think you can build pretty big
companies with a totally different capital structure given the increased
efficiencies of building with AI itself.
Speaker 4 (36:39):
How many of them are in Boston? Are you looking
across the US?
Speaker 1 (36:44):
We invest across the US, but we're very bullish on Boston.
So over half of our capital is deployed here here
in Boston. And it's because this is an ecosystem that
continuously produces the top talent in the world. It continuously
produces some of the top research and development the world,
and increasing you know, it consistently solves really hard problems,
so things like healthcare and biotech and life sciences and
(37:08):
financial services, and so cybersecurity and energy and climate. So
you know, these three factors are you know, are part
of what gets us very excited to invest in the
bus and ecosystem, particularly in this age of applied.
Speaker 5 (37:20):
Ailey Lyman, great to have you, managing partner of Underscore VC,
Wishing well for the rest of the year. Meanwhile, Elsewhere
and VC, a startup founded by the former lead researcher
of Open AI, is raised forty million dollars in funding.
Speaker 4 (37:33):
Waveforms AI.
Speaker 5 (37:35):
It makes AI audio software that improves verbal interactions with machines,
and is one of several AI companies developing.
Speaker 4 (37:41):
More human length voice features. Bloomberg's Rachel Mets is here
with more. It's come out of stealth.
Speaker 5 (37:46):
There's only five people there, but it's got quite a
head evaluation already.
Speaker 14 (37:50):
Yeah, this is It's still a pretty small company. They
don't have a demo to share publicly yet, but they
are working on software to improve the kind of interactions
that people have with computers when we're speaking to computers.
Speaker 5 (38:04):
And this is all around Alexis Kono, who was very
senior at open Ai. So I assume that even though
small and without a demo, people are betting on him
his expertise in his background.
Speaker 14 (38:16):
Yeah, he has quite a long tale of experience working
on voice computer human interactions and he was pretty instrumental
in helping Opening Eye build its advanced voice mode. This
is the more recent version of the voice mode that
works with chat GPT that is meant to feel more
lifelike than it had in the past, more feeling like
(38:38):
you're having a conversation with an actual person than with
a computer business model.
Speaker 5 (38:43):
I mean the fact that we don't have a demo,
but ultimately, how is this company going to be returning
value to its vcs?
Speaker 14 (38:49):
That is a great question. What they told me is
they are going to be releasing at some point products
that are both B to B and also consumer oriented.
So you might imagine something like an app perhaps that
would be for consumers from the company to kind of
give people a sense of what their technology is and
how it works. But then also perhaps they might work
(39:12):
with other companies to let other companies use their software
as well.
Speaker 4 (39:15):
And the San Francisco based yes, that is true. They
are based in San Francisco.
Speaker 5 (39:20):
Rachel Matz all things open ai and ex open ai.
Speaker 4 (39:23):
We appreciate it, thank you very much.
Speaker 5 (39:25):
Indeed, Oracle stock having its best year since nineteen ninety nine.
It's all thanks to the momentum of its cloud business.
We'll find out if that momentum will continue when the
company releases its earnings today after the closing bell.
Speaker 4 (39:45):
Luberg's Brodie Ford with a preview. I mean, what an
extraordinary chart.
Speaker 5 (39:48):
It's just done so well because people are now like, oh,
the cloud thing, it's actually relatively doing well compared to competitors.
Speaker 7 (39:54):
There were many years that people would make jokes to
me about Oracle's cloud about how we know, will this
ever actually be a serious player, And within the last
two years, the answer has become yes, right, And a
large part of that is because the AI workloads are
so demanding that the traditional giants like AWS and Azure
just don't have the capacity, right, And so it's become
(40:15):
that Oracle's cloud is seriously taken as the kind of
fourth hyperscaler, and it's been able to grow in such
an accelerated way that very few other software companies can
you know, compete with.
Speaker 5 (40:27):
It's got some I wouldn't call it a key Man risk,
but certainly I'm talking about Uber as a client, but
also TikTok, and we've been talking a lot about some
concerns around the deal around TikTok.
Speaker 7 (40:37):
Absolutely right. I mean, yeah, we saw the court ruling
that the there's increased odds that TikTok could go ahead
and go poof, But there is you know, at the
end of the day, if anybody buys TikTok, the odds
that you say, hey, let's go ahead and switch out
our cloud infrastructure. Nobody wants to do that, right, I mean,
so it is a risk for Oracle, But you know,
(40:58):
when TikTok was one of their only marque customers two
years ago for OCI, it was a bigger deal today,
they have a lot of customers, and I think the
risk isn't quite as acute as it once was.
Speaker 5 (41:08):
What about other things in infrastructure and cloud provision, i
mean software, other services.
Speaker 4 (41:13):
What else are they managing to do well on?
Speaker 12 (41:15):
Yeah.
Speaker 7 (41:15):
Well, one that's really interesting is that for a long
time they really tried to get you to run their
key iconic database software on prem and kind of control that.
And they finally said, you know what, you could run
it on Azure, you could run it on AWS. And
that's kind of been a pretty big exciting thing for
investors that you know, a lot of the world is
on AWS or Azure and they want to use the
(41:37):
Oracle database and so being able to kind of update
that customer segment, you know, breathe some growth into the
database segment, which you know dates back to the late seventies.
That is another thing that's been exciting folks about a Oracle.
Speaker 5 (41:50):
But a lot of it is probably priced in when
you're looking at an eighty percent run up and the shares.
So what do you tend to hear in these calls
post earnings.
Speaker 7 (41:58):
Oh god, we're building data centers across the world, you know,
no one's ever seen a data center build out like this, right,
I mean that's.
Speaker 4 (42:05):
The general energy informative. Yes, absolutely, I.
Speaker 7 (42:07):
Mean the amount of hype right now about the AI
fueled data center buildout is massive, right. You see you
sat pricing the shares. You know they are near an
all time high. And so the question is so much
that how quickly are they going to be able to
convert demand into actually offering these services, because you know
they just like Microsoft and Amazon, they can't get these
(42:29):
centers up quick enough. They can't get enough chips to
deal with the demand.
Speaker 5 (42:32):
Well, despite the run up, they've got twenty five buys,
fifteen holds and not a single cell rating.
Speaker 4 (42:37):
Bodie Ford, all things are coole.
Speaker 5 (42:39):
We'll be checking in on him tomorrow, I'm sure, after
those earning drop. But meanwhile, that does it for this
addition of Blomberg technology. You don't want to forget a
podcast find out on the terminal as well as online
on Apple, Spotify, and iHeart this is Bloomberg technology.
Popular Podcasts
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.
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
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.