From the heart of where innovation, money and power collide in Silicon Valley and beyond. This is Bloomberg Technology with Caroline Hyde and Ed Ludlow.
Live from San Francisco. This is Bloomberg Technology coming up. Big Tech volatile after Moody's downgrades, America's credit rating, and Nvidia allowing customers to use rival chip makers in an effort to expand its AI ecosystem, And why Apple has been spending billions of dollars on AI and still hasn't managed to crack the code. This is what financial markets
look like across major averages. We've paired some of the early declines in Monday's session, but the story broadly is tech is pulling us down and leading us lower, and that center is particularly on semiconductors.
And AI infrastructure names.
In terms of the single names that we're watching, you have stocks like Tesla under pressure, and Video has actually paid a lot.
Of it's declined from the open.
It was down two percent at one point in Video is at the heart of the news cycle and it's going to be at the heart of today's program as well.
With a lot of news out of Taiwan.
Let's get to the big picture on technology and financial markets. Here to break down the tech moves, Bloomberg's Ryan Vostelica, Ryan, what's going on?
Hey, good morning, Thanks for having me. So there is a lot of uncertainty, a lot of volatility in the market today. You mentioned the Moody's downgrade. I think that just speaks to the level of political uncertainty.
In particular.
Of this, of course comes in the back of all the back and forth on tariffs in the trade war. There's just so much uncertainty in terms of policy, and the implications of this, I think are for a lot of people to just continue sort of selling the US look for defensive areas, look for safety. In that case, that means a lot of big tech names, especially since tech has been pretty over the past couple of weeks.
So this could just be some profit taking, but I do think it underlines how uncertain people feel about the economic outlook.
The mag seven have a sort of disproportionate impact to the index level, but it's AI, infrastructure and semiconductors where I see most weakness.
Basically this Monday. Why is that?
Yeah, So, I think you know, this is an area that is very much tied to the trade situation given to sort of the nature of the products that we're talking about here, there's still a lot of uncertainty about you know, chip restrictions and all these other kinds of things. And this also comes on the back of, you know, people still continue to want to see more return on all this AI related spending that is going on. There's just a lot of concerns maybe that you know, has
there been sort of a bubble in AI infrastructure. So I think this is an area where because of these group has done so well because the outlook is so debated right now, it's just a very natural place for people to be taking profits, especially on any kind of sort of negative headline that you see cross the wire.
Bloomboas round Vastelica with the market moves this Monday, thank you, let's get more on markets. Hillary Frisch, senior Research analysts for Software and IT services at Clearbridge Investments.
Is with us.
Last week, the tailwind was in all of the headlines about the billions and billions of dollars of AI infrastructure deals the United States and the Gulf Nations, Nvidia, super Micro, Dell.
But that's the area of weakness.
This Monday morning, is it just a little bit of a reality check on how soon some of that investment will be felt.
It's a good question, ed.
It may be a reality check, it may just also be profit ticking. And we went from overbought to severely oversold to overbought again within just about the span of a month. So the tech sector in general has had a very strong run recently, and we can talk more about why that's the case, but I think investors are just looking seeing, okay, we can take some profits at this point.
Well, in part of why that's the case is earnings, right, I think we're still a little bit digesting what the net takeaways from the calendar first quarter earnings were. What were your main takeaways from that period?
Sure, well, first of all, it was I think what was notable was just how resilient Q one results were. And it wasn't just in Q one, it was in the subsequent weeks through to the quarterly reports, and that encompassed a period of Teriff related headlines and uncertainties. So that was refreshing and encouraging. And it wasn't just AI or a series of AI proof of concepts at the expense of all else, which is largely what it felt like last year. It was AI strength plus real core
business strength, and that was refreshing. Now we're coming up upon April quarter earnings reports, and I could imagine that some of those companies who would have had to close their quarters in the midst of tariff uncertainty might see some variability, might have seen some pausing and spend, but overall the underlying results are very encouraging.
What I found super interesting about that earnings period. In technologies, you have like primary data and secondary data. The primary data like the commentary and what's in the earning statement, but some of the secondary data we were tracking, for example, in the AI infrastructure space, was leases, data center leases, and that caused all kinds of noise going into the prints and we had to decipher what it was that was the truth, you know, based on executive comments, How did you navigate that?
Yes, set of data, tons of noise in the space. That was probably maximum noise that we've seen in a long time.
But it's fascinating.
Starting with the underlying demand trends, you saw hyperscale vendors reaffirming CAPEX plans. You saw some of them increase them. Microsoft in particular talked about rising near and long term demand signals, which could actually cause them to prolong their investment in long live assets, meaning land and buildings, which they've been engaging in our over the course of the last eighteen months. So how do we marry that up with data center lease cancelations. Well, it's interesting Microsoft in
particular makes a lot of use of leases. They came from behind relative to the strong demand signals they're seen in getting that requisite capacity to be able to fulfill AI and other demand for the next five ten years, So they were building rapidly. Within that, they signed a ton of leases across the globe, which was a brilliant strategy, by the way, because it froze it froze some of
those opportunities to competitors. And then within that, I think they decided that they didn't want to be the sole source of training for open AI. And also within that there were power constraints, and I think they realized that these leased facilities wouldn't be up and running until twenty seven to twenty twenty nine, and in that vein they could actually shift to some degree to own data centers as well as least data centers and you're seeing that
kind of across the board. But in the end, some of the data center lease capacit we were hearing was five x what you were seeing signed on for by an Amazon, for instance, for the rest of the ecosystem. And last time I choked, Microsoft wasn't five x the size of an Amazon. So basically they were giving themselves optionality. And in retrospect, they're still building and they're still demand.
And again a part of the higher capex reflects the higher cost of doing business because of the impact of tariffs to supply chain. I've been thinking a lot, as one does, about year to date performance of the hyperscalers. And it's interesting, right if you just say hyperscalers Microsoft, Amazon alphabet. Microsoft stock is up more than eight percent year to day. The other two names have been under pressure.
We of course kickoff with Microsoft Build, but we also talked last week about some of the difficulty Microsoft has in reaching its own goals because of its relationship with open AI. This is something you've been thinking about as well.
Yes, Well, first of all, with respect to relative to performance, I think you're almost seeing from a cloud perspective in the other two what we saw with Microsoft for the end of last year and early this year, which is that they're behind a getting capacity and that is constraining cloud growth. So it hasn't been a demand issue, it's been a supply issue. Second of all, Microsoft has been very tethered to open Ai. They've been diversifying, but the
two are quite aligned. I'm not sure I view it at all in terms of their not being able to realize their goals. I think their goals are to focus on their own compute capabilities for their customers and to add in open ai and other LLLM vendors, but especially
open Ai as part of that equation. I think it was a deliberate decision not to focus solely on training that would have crowded out a lot of the other exciting and high margin endeavors Microsoft can engage in an as core customer base, and in the meantime it's a
highly symbiotic relationship. Microsoft's infrastructure was originally built in AI to around open Ai and the related infrastructure it has in finnabaum throughout it has some tremendous capabilities and I believe the two companies want to continue to work together.
Hillary Frish, Senior Research Analystic Clearbridge Investments.
Thank you for joining us.
Now coming up, Video announces new AI products as the computing giants descend. On Compute's twenty twenty five, We'll go out to Taiwan next this Spoomberg Technology.
We're outside the venue here where Jensen Wang just kicked off Kombu Techs for twenty twenty five. This is one of the biggest tech events in Asia and Jensen Wang, the founder and CEO of Nvidia, was on stage for ninety minutes at talking through his vision for AI and the products that are going to be needed to sustain it.
Today we're announcing the Fox Kind Taiwan. The Taiwanese government and VideA. We're going to build the first giant AI supercomputer.
The focus was on greater AI applications, so specifically the integration of AI into products like humanoid robots, also autonomous driving vehicles. On the product front, we did get an update on the next generation GB three hundred system that is expected to come in the third quarter of this year. Also, Nvideo is going to be offering a new version of the complete computers that it provides to its data center customers.
Now what that means in a nutshell is that it's going to be opening its AI servers to chips from other companies. This is its names like Microsoft, Amazon, they try to design their own.
Processors and accelerators.
In Taipei, I'm Annabel Drawlers Bloomberg News for more.
On computext and Nvidea's latest revealed Bloombozine. King joins us here in San Francisco. What we're talking about is envy link Fusion. You and I have sat through many briefings about envy link Fusion and read and written about a lot of it.
But it's a high speed into connect.
You can connect lots of GPUs and in this case also CPUs.
The main point here is in Vidia is.
Saying if you are a hyperscaler, perhaps with custom Silicon, then we can allow you into data centers where we dominate, and it kind of opens up this place where everyone is not only and solely reliant on one single name, which is in video.
I think the other way to look at this would be it's a business decision, right if you have your biggest customers who are essentially creating rival efforts, rival efforts that may one day mean that they need less of you why not make it easier to actually work with you say hey, great, yeah you've got a great CPU, come work with us. Hey you've got a great GPU, Come work with us. This is a way to keep people in in Video's ecosystem and to stop them going elsewhere.
That's the pitch.
Computexts is a really big deal, and in VideA seems to have dominated it in more recent years. What else do we learn about things of substance that they're actually doing at all?
Anything of substance, Well, computext wasn't a big deal up until in Video basically made it a big deal again when they started using it on this worldwide trip that they do.
To keep the.
Air momentum going as a show, it was definitely fading away. So they put a lot of updates of things out there, robotics, new servers, new designs for data centers, and basically just trying to keep the momentum of this massive AI infrastructure build out going.
Is there anything else happening at computext that is not in Video related?
Well, our old friend Christiano Amon actually appeared on stage with Qualcom right the Corecom CEO. As we reported a long time ago, they want to get into the data center with their technology. This is part of his diversification and if you can't beat them, join them. So he's there saying we'll work with in video as a step towards the data center strategy, and that is very pragmatic and we'll see how it works out.
Okay, Bloombozie and King, who leads our coverage of semiconductors, Thank you very much. I want to stay on that news out of Computex in Taiwan and bringing Michelle Geider. She's the CEO of the Crash Institute for Tech Diplomacy at Purdue. She's also served as Assistant Secretary of State for Global Public Affairs under the first Trump administration between
twenty eighteen and twenty twenty. For me away from the technology news, the absolute key piece of importance was Jensen one talking about his relationship with Taiwan and in Vidia's relationship with Taiwan.
The basic plan is to.
Build a supercomputer for Taiwan, but he was kind of at pains to point out that this is in partnership with TSMC and other leading names that are Taiwanese champions.
Purely from a geopolitical perspective.
How crucial was that or how high level was that for Jensen Wong to say such a thing.
It's very important and in video with Jensen Wang at the Helm has been on quite the tech diplomacy tour, the chip diplomacy tour really over the past few in the past week, including in Taiwan, as you reference where they're talking about building a supercomputer. Taiwan is a very strategic partner and trusted partner of the United States.
It's very important for our shared security.
And that's right off the heels of some major deals announced in the Golf over the weekend and late last week with Saudi Arabia with a UAE who also want to be global capitals when it comes to artificial intelligence, and there's going to be more sharing of advanced chips there. So all of that is really important as the United States seeks to secure trusted partnerships with important partners across the globe when it comes to advance semiconductors.
Michelle, those deals with Golf nations, how concerned are you that they can act as sort of backdoor indirect access for China to US technology.
It's definitely a concern among some in the US government. Including in Congress. But I'll say the big picture is we have to do two things. One is maximum proliferation of American technology, including our semiconductors, because if it's not our technology, it's going to be our adversaries technology like China,
So max proliferation is really important. At the same time, we also have to have maximum security controls to make sure that those technologies, including chips, don't get diverted toward adversaries like China.
Those two things have to happen.
It can be done, but it requires that companies and the US government and our allies by the way, work really closely together.
And that's what tech diplomacy is all about.
And I think you see the White House, the Commerce Department, as well as leaders in Congress with the new Chip Security Act making sure that those security controls are in place.
Jensen Wong's text diplomacy, I think it is worth lingering on Taiwan is clearly important to Nvidia, and we remind ourselves that the mainstay of TSMC's lead edge fabrication is in Taiwan. But China is also important to Jensen one whether that market is closed to him or not, is he getting the balance right.
Well, it's interesting that you mentioned that because on this tech diplomacy tour there were the deals announced in the golf there's the news in Taiwan today. But let's also not forget that he announced a new research center in Shanghai just a few days ago. And it's no coincidence that all of these things are happening in the same time. And look, he's got a global business to run. He
definitely has multiple interests that he's trying to meet. But we have to make sure from the US national security standpoint is that all of that doesn't endanger American national security and give some of our most advanced technologies to China. So the role that he plays in working together with our government, we have to make sure that business objectives meet national security objectives. And it can be done, but we have to be working together in.
Order to do it.
In January of this year, Anderill's co founder Pamelucky told me point blank that they are preparing for scenario where China invades Taiwan at some point in the coming years. It is one severe come take that argument and extrapolate it out to ai infrastructure and semiconductive manufacturing and Taiwan's role in everything that we've just discussed.
Well, it's a reason that we have to make sure Taiwan's resilience and the US resilience when it comes to these critical sectors like semiconductors is really strong, and so partnerships like with Nvidia TSMC is investing another one hundred billion dollars year in the United States to build out its FABS, to build out advanced packaging facilities.
All of that has to scale so.
That we have some more strategic partners and it's also with other allies like Japan, like Korea s Khiiniz is investing in an advanced manufacturing facility for semiconductors at Predue Research part where the CROC Institute is. All of that has to happen more deeply in its scale in order to make sure our shared resilience and security, because.
We have to be prepared for a scenario like that.
But rather than just try to slow China down, we have to make sure that we are turbot our own supply chains, our own resilience in our collaboration with our allies.
Michelle Guided, CEO of the Cracked Institute for Tech Diplomacy at Purdue, Thank you very much. Okay, take a quick look at shares of Ali Barber and Apple. The New York Times reporting that the Trump administration has raised concerns over a potential AI deal between Apple and Ali Barba that would allow the Chinese firm to make it's AI available to use on iPhones in China.
We will track that story as we go, and we stay on Apple.
The tech giant promised a bold leap into AI, but instead Siri has been left stumbling and the company's AI ambitions are lagging behind rivals. It's the subject of the latest Bloomberg Big Take, and its author Mark German joins us. Now, this is a deeply reported piece about the people inside Apple working on AI, the rollout of Siri and Apple Intelligence. But if you could just summ arise the pace, how would you summarize it?
I think the best way to summarize it is that Apple has a history of coming into emerging areas late. You saw that with tablets, smart watches and B three players, even smartphones. But they've come in and then they've destroyed or vanquished the competition. We've seen that with the iPod, the iPhone, the iPad, and.
The Apple Watch with AI.
Not only were they late compared to many of the competition, right, but they're not the best. In fact, it's fair to say all the big technology companies, they're behind more than anyone else. And there are philosophical issues, management issues, prioritization issues, financial issues, perhaps even innovation issues, and many questions that led to this point. Apple was caught off guard when
chat GPT launched in twenty twenty two. There are questions about how a company of Apple scale and expertise could have not seen this coming. Then there are questions of how they released, with all their resources products that appeared so rush to market, And this story takes a deep look at the how and why about all of that.
I remember, not necessarily twenty eleven specifically, but twenty eleven and twelve thirteen early editions of Syria and being like, Wow, Siri, what is this? This is astonishing, and then quickly it just fades away.
At the heart of your reporting.
Is the hiring of a key executive in twenty sixteen. Who is that person and why do you follow that executive's passage through Apple?
Yes, John g Andrea.
He was the head of search and AI at Google until Apple hired him in twenty eighteen, so about seven years ago. And he runs the show when it comes to AI, or at least he did run the show when it came to AI. For the last several years, he was in charge of SIRI, in charge of foundation models LMS, AI testing AI infrastructure teams designed to annotate in train data to determine how well llms and the different AI technologies are working. And him come right to
Apple was very exciting for the company. They thought they would turn into an AI powerhouse. Instead, the one person in charge of AI oversaw as the company completely missed the boat on this technology. But as they get into the piece, it's not him alone. Marketing issues, finance issues and software issues.
Blue most Mark German with the latest Bloomberg Big Take. It is an absolutely must read on what's happening inside Apple with AI. Thank you very much, Welcome back to Bloomberg Technology Ed Lovelow here in San Francisco. So stocks have pared some of the declines, but the story in the session has very much been AI infrastructure, semiconductor names
under pressure. There was the impact of Moody's credit rating downgrade for America from Friday Night but actually I think a lot of it is a reality check from all of the headlines of last week about all of the billions of dollars of investment in AI infrastructure Microsoft, the exceptions of the rule of up half of percent, There is a lot going on right now in the technology industry. There are several events either underway or due to take
place this week. Computext we talked about in video focus of that Google iOS coming up, Dell Technology World coming up, and then there's Microsoft Build.
And I think that's a good place to start.
Let's get to what we can expect from Microsoft Build and bring in anarag Rana of Bloomberg Intelligence. You've done a little preview piece of research for Microsoft Build, and I think that you, like everyone, are kind of expecting Microsoft to talk about how we're actually going to be using some of the AI tools that they charge us for.
Yeah, you know, when you look at the air infrastructure space, Microsoft has really made a name for themselves to saying, you know, we will focus a lot more on inferencing rather than then training, and this is where a lot of the innovation I think it's going to come from. They're going to come from new products that are going to be launched throughout the ecosystem, more focus on small language mores and that you know may or may not use that much of GPU use out there, And I
think that's the way they want to focus on. Is the company that you know is more focused on inferencing than then training.
Well, so important about Bloomberg technology this show and RAGU is like lots of people that watch it are either already at one of those events on their way to it, or they'll be tracking it closely their developers or their workers in the technology industry. And part of your research is focusing on the lower cost large language model. Why is that important? Why does it impact all of the people across the world of technology that are watching this program right now.
So one of the most important things for us is how do you get this technology to be used, you know, throughout the masses, whether it's on the consumer applications or on the enterprise applications. So when you look at even Microsoft co Pilot, you know, subscription it's at thirty dollars per user per month, and with that high rate, we think the adoption rate is going to be a lot lower than if they bring it down. You can extrapolate
that on the enterprise use cases. Also. Currently we're in a phase where running a large language model costs a lot of money. Now, over time the cost of computing will go down or broken usage, and I think this is where one of the ways you can do that is by using a small language model that does one specific task and one specific task very well. And for that you really don't need that sophisticated orf A hardware to run it.
All, right, Ran or Bloomberg Intelligence.
Let's catch up at the end of the week, when all of these events are out the way, we can assess what was said. There is more AI news out of the Middle East and Europe in video and Abadabi Investment Vehicle MNGX partnering with French firms to build Europe's largest AI data center campus. The moves follow a week
of AI announcements. We've been over them during the hour during President Trump's visit to the Middle East, and that includes plans by Saudi investment fund STV to launch a one hundred million dollar like AI Focus fund.
We're backing from Google.
STV General partner and COO Luca Barbie joins us now from Riad. Congrats on the fund. You're one of the biggest sort of private growth equity venture firms out there in terms of money under management that's focused on like a very specific domain. What do you want to do with this fund and what is the unlock of everything that happened last week because your region is so under the microscope with all of the infrastructure investment that's taking place.
Thanks for letting me. STV Indeed is a povate technology company. We invest in technology in the Middle East, so we have a one point five billion of asset under the management and last week we announced.
A thematic fund for AI.
So our thesis is that this time to invest in the application layer. The Middle East has tool in itself being able to build a technology company. Today we have fifteen unicorns in the region and more in the pipeline. In our portfolio, we have nine companies that are preparing for IPO and so there is enough knowledge, there is enough momentus, there are enough infrastructure to compete and play a role in the AI arena. The fund will be
focused on the application area. As I mentioned, there are a number of use caes that are immediate relevant and can find tangible application in the market, especially with the customer service or enterprise, automension or legal This is where we are focusing today.
We talked about Google backing this fund. Which are the notable LPs and investors? Have you managed to get behind you?
Most of the things that are still in the making, So we disclosed name in the new course, But we have a number of institutional investors that have been with us before, and we'll come also in the iFun and I think the time is right to attract international investors in our fund and in general in the vision.
A big factor in the Gulf is the activity of the sovereign wealth funds. So you think about Saudi Arabia for example, we covered a lot last week on the show Humane.
How does that impact your ability to do deals?
Do you kind of get muscled out by either the public investment fund or another vehicle like Humane?
Or are you looking at.
Different company sizes and targets for your investments?
What's the foundership puting prim and public sectors so far has worked quite well, even in the most recent announcement that we had last week.
Most of these money and capital will.
Be deployed in the infrastructure layer, so large project costs billions of dollars to develop data center, to develop also and localize and models.
So this is.
Actually favoring the initiative of the private sector. And as I mentioned before, we focus more on the application layers.
So this is actually enforcing our teases that investing in AI in the application layer.
In this vision we make more and more sense as the infrastructure will be comparable to what we see in the most developed markets.
President Trump's visit kind of put Saudi's name out there, even if you forget the infrastructure layer right now, are you worried about evaluations a little bit? Given how much interest and energy there is into looking at Saudi Arabia as an investment opportunity, There's.
Always a challenging job.
There is a lot of volatility that we have observed across different cycles, so we need to deal with it. We need to put forward our reputation as a long term credible investors that can add on a number of value adding activity, in particular to integrate company in our ecosystem, talking to government, reguletto large layers.
And make things happen. So we believe that smart and experienced entrepreneur we value that.
Actually, we welcome the Trump visit I think the world is more and more noticing what's.
Happening in our region, in the Middle East. So far things have been under the RADA.
But I think, you know, especially in our sector, in technology, the growth is exponential and I would expect in the next three five years to become you know, the success case of the Middle East intact to become more visible, and so we'll be more more more global capital and more global investors. I see we enter the market and play a role next to us.
Luca, you talked about some of the unicorns in the region and some of what's in your portfolio. You must be about excited about some exits this year. Any case studies where you see an IPO or even M and A giving you a nice exit.
Absolutely so.
Look, I mentioned fifteen unicorns so far, right, five of those fifteen.
Being materialized the next eighteen months.
This is why I was talking about the exponential shape of how things happen in technology. You know, our portfolio, we have other four candidates.
One of it is actually already unicorner. It's Toby by Now paying later.
They raised recently around that three point two billion's dollar valuation and as a credible path to be listed in the Middle East Socket Change in the next eighteen months. So we need more of this success case to hove ourself and the rest of the global audience that the market is big enough, entrepreneurs are innovative enough, and the capital can be you know, put a play with attractive.
Y Luca Barbie, general partner and COO of STV. Great to have you here on Bloomberg Technology. Thank you very much. Now coming up, we're going to talk to Luxe Capitalist Josh Wolf about the firm's helpline for academics and scientists and the state of research in this country.
This is Bloomberg Technology.
As the US public sector pulls back on funding for research and development, some big name private sector names are stepping in. That includes Lux Capital Managing partner Josh Wolf is here. Lux Capital's portfolio includes names like Anderil and
drug discovery company Icon Therapeutics. What you were doing, Josh, is a one hundred million dollar commitment for basically a helpline for American scientists, American researchers whose work might otherwise be taken from underneath them because of the lack of funding.
That's absolutely right.
You go back eighty years American competitiveness started with that of the Bush. Yeah, this great thing, the Endless Frontier. That is what gave us competitive advantage. We attracted the best and the brightest from German Jews that came here and made sure that we had the atomic bond, went
to the Institute for Advanced Studies at Princeton. You go through the Cold War in the eighties, attracting Russian emigree so that we can have the best in the brightest sciences in the world, all of our academic institutions being funded.
Then you fast forward to two thousand and five.
You've got Normal Augustine, previously head of major defense company, writing the Gathering Storm, which was basically like the clouds are gathering and we are starting to lose, particularly to China. Today, I would call it a perfect storm. You've got politicization at universities. You've got massive federal cuts and that is not a political thing between the Republicans and the Democrats. That is the past twenty two years we've had major
funding cuts in federal government. We are at the lowest level in science funding from the government since nineteen ninety seven. Why does this all matter? We are competing with China, This is not a left or right, or red or blue. It is a past verse future thing. And with those funding cuts and the politicization on the problems that scientists are facing, the lifeblood of venture capital, the life blood of innovation is this.
So we're stepping forward and seeing at.
Least one hundred million dollars, we're going to be doubling down on early stage venture creation around scientists.
Josh I discussed this exact issue with Michael Kratzios, recently Director of the White House Office for the Science Technology Right just listen to how he summarized funding right.
Now, if you spend a lot of money on the wrong things, it's worse for spending a little money on the right things. So we're making sure of that the biggest priorities for the nation, things like leading artificial intelligence, leading in quantum information science. Can we continue to fund those areas and make sure that our universities are a big part of that agenda.
How mister Kratzios would summarize it is that it's how and where you spend money, spend but.
Research funds rather than the volume.
Your reaction to that, and I guess that the lack of knowledge right now and the direction that this administration is going in terms of committing to growing.
R and D budgets.
Well, first of all, I have a lot of respect for Kratzios. I think he's doing a great job across many things. I would say as it relates to funding the things that are evident and obvious science and the fundoing for it is not a slot machine for strategy returns. This is about the slow burn of serendipity. We never knew that COVID was coming, but thank god we were able to have messenger RNA mRNA vaccine that was there. We would have been late to the game if we
were trying to fund it when the moment happened. So when we were talking about the things today that mattered, AI and quantum, which I'm super skeptical about, and some of the cutting edge areas and robotics and aerospace and defense, those kinds of things are the things that we've been funding for ten years. So the venture industry has been funding it for ten years, which means that the breakthroughs have come from science and academia, and that's.
Fifteen years ago.
So the things that we should be funding now are this slow burn of serendipity that five, ten, fifteen years.
Hence folks like us from the private sector come in, identify and take the risk and put in risk capital behind the human capital, behind the human capital and the intellectual capital of intellectual property and patents owned by the US and owned by our government and licensed to the universities and to the researchers, and then ultimately to the startups that can create multi billion dollar companies doing exact
actly what great cratios are saying. So I do not agree that we should be funding the here and now. We want to own the future. We should be funding the future, and you never know where it's going to come from. If we did, it wouldn't be science.
Josh, you this conversation pretty clear. This is about competition with China, right, And I record conversation I had with Jensen Wong in March where he basically said, look, fifty percent of AI research is being done either in China or by Chinese researchers in other domains. It is simple and clear cut as the Chinese government just being willing to commit capital and do what it takes kind of
attitude for them to come out on top. Or is it different to that, what is the domain for the competition here?
It is a geopolitical domean. It is a cultural competition. Cultures get what they celebrate. When we're celebrating celebrities like Kardashians and Paris Hillmen, TikTok and nonsense, and they are directing their youth to go after the cutting edge that will define competitive advantage, to let them dominate in biotech, in aerospace, in breakthrough materials, in space itself, in life saving drugs. Today most of the world depends on the
American biotech industry. You go back to Conan Boyer on the West Coast that launched Genetech and helped to create this vibrant industry unleashing cancer diagnostics and drugs, immunotherapies and JLP ones for weight loss. All of that came from basic, undirected science. It's absolutely critical. You go to aerospace and again the atomic bomb, thank god it was built here and not.
Somewhere else through our Manhattan Project.
That was massive directed scientific research with lots of unknowns, where we had Germans and Austrians and British all coming here to work so that the Allies could have superior advantage. So we are fighting and losing right now in nearly forty to forty four critical technology areas where China is dominating, and Jensen from Nvidia one hundred percent correct, fifty percent of all AI graduates today are coming from China. That's not a twenty twenty three or twenty twenty four decision.
We're in twenty twenty five. This was a decision from fifteen years ago. So the decisions we're making today about long horizon science is what is going to give venture capitalists like us the ability to go in to early stage science and do what we do. We've done over twenty companies Denovo from university research, but that wasn't research that was done a year ago. It was research that
was funded ten to fifteen plus years ago. Being able to back Nobel Prize winners like Richard Axel, Josh where we in Caliobe, or or all kinds of other companies critically important that the science research funding is there now so that US venture capitalists in the private sector can step in tomorrow.
I sorry to interrupt you, Josh, I just want to go back to the lux Science Helpline before we run out of time. And it used to be that you would require those researchers to have their work ninety ninety five percent scientifically proven before you'd commit capital. You've lowered the bar to fifty percent scientifically proven. What are the risks with that and the motivation for doing so? We have about a minute left.
Lowering the bar for the science risk is raising the stakes. Before we would want to only take market risk, technology risk, people risk, product risk, financing risk. Science risk we wanted funded by the US taxpayer and by the university. Now
we're saying, if it's not fully baked, it's okay. If it's half baked, if it's down to fifty percent ready, and it still needs money from the private sector, in a private lab, in a venture supported company, We're willing to take that risk because the stakes are that high, the profits for the scientists, for us, for our limited partners are that high. It's worth doing. We're not talking about a huge amount of money. We can't do it alone.
We're leading this effort. We're inviting our venture peers who can be scientifically minded to create syndicates of American greatness here to come along for the ride.
Lux Capital Managing partner Josh Wolf, thank you very much coming back on the program, and thropt pic has cultivated a reputation for taking issues such as safety and responsibility seriously, more seriously than the company from which it is sprung, open Ai. But can it keep up while remaining so focused on safety? Bloomberg Shering Gafari joins us for more
on her Business Week's story. I look at Anithropic and as you know, like you spend a lot of time talking with Anthropic as well over several years, I feel like they have managed to move quickly, but it's this kind of repetition of safety. What are you talking about in the Business Week piece?
Right?
If you think about just even a few years ago, right, Dario Miday, the CEO of Anthropic, was someone who really had an academic background.
He had worked at open Ai and buy New Google, but largely spent his time as.
A biophysics researcher right running experiments. And so now he's in this position where he's actually running a sixty one billion dollar startup while trying to balance that sort of academic commitment that he has to his ideals around safety.
In the piece, what do you report on how they're getting that balance?
Right?
Running a business versus the safety responsibility.
Right, So they have some impressive numbers.
I mean the latest annualized revenue projections they have are two billion, and that's double what it was at the beginning of the year. So it's nothing to you beat an eyelash at. And at the same time they are going to have to keep racing it this very fast feed to keep up with deep Seek, to keep up with open Ai, with Google, and so how do you sort of balance it when sometimes you do have to take a step back.
That's going to be the challenge I think in the years ahead.
When I speak to computer scientists or even people in the enterprise software space, I think what and frop It gets a lot of credit for, and what they're good at is coding agents and those sorts of domains. How have they done well there? And what are their products? What is it that the anthropic cells to the world.
Yeah, So another nugget in this piece is it one of the biggest sources of growth for them recently has been through their coding agents and sort of coding technology, and that's essentially software that helps engineers write their code. And that is, on the one hand, a huge boon to them because if they have the best software for that, they're going to get big contracts. On the other hand, it also provides a complication in terms of what does that mean for the future of work?
What does that mean for employees?
Even with an anthropic whose jobs are going to change, you know, and A told me, we don't want to slow down. We don't want to have to fire our employees because of our product claud code, but that means we may have to slow down hiring in the future, right, So I think he's being quite open about the trade offs.
Here very quickly, just twenty seconds. Where are Anthropic? Were's home for them? And who are their biggest investors?
So they are based here in San Francisco actually, where Drio Meda is from. And their biggest investors are folks like light Speed.
You know. Eric Schmidt was.
An early investor who I talked to Google and Amazon of course, who will also partner with them on the Infrat.
By Sharen Gafari with Frankly must read in BusinessWeek.
Thank you very much.
That does it for this edition of Bloomberg Technology. Don't forget check out our podcast. So many of you listen to this show as a podcast. You know where to find it. It's on all of the Bloomberg platforms, as well as on Apple, Spotify, and online on iHeart from San Francisco.
This is Bloomberg Technology.