Listen: Next in Tech Episode 157: AI Investing

View Full Transcript

Eric Hanselman

Welcome to Next in Tech, an S&P Global Market Intelligence podcast, where the world of emerging tech lives. I'm your host, Eric Hanselman, Chief Analyst for Technology, Media and Telecom at S&P Global Market Intelligence. And today, we're going to be looking at an aspect of generative AI. It's a little different than the tech deep dives we've been doing. And to head into the money flows, the investing, all of the aspects that are the other side of this coin on AI is returning guest to Melissa Incera. Melissa, welcome back to the podcast.

Melissa Incera

Thanks, Eric. Great to be here.

Eric Hanselman

And this is a lively area. We've heard the joke that LLM actually means lots and lots of money. Just looking at the investing that's taking place in generative AI, it seems to have a whole different character to it to the enthusiasm that's really been behind this, on the financial side as well. And you've been doing a lot of research into what that looks like and how is this happening.

I'm curious. We're coming off a year in which tech investing has been down. But if you take a look at what's happening with generative AI, man, there's a lot that's taking place and enthusiasm in a bunch of new ways.

Melissa Incera

Yes, it was the story of the year. We talked on the podcast before about how 2023 was such a weird year for capital markets in general, a lot of factors behind that. Like interest rates, are they going to go up? Will they go down? Investors, we're all still a little shocky after watching valuations come down in the last few years.

But AI and specifically, generative AI really just gave them something to get excited about specifically because it's pretty clear this isn't another incremental advancement, but it's entirely new wave of technology that we're pretty certain is going to change the world. And what's more, it has impact across the entire tech stack. There really aren't many pieces of the tech landscape that aren't impacted by this.

So yes, just to put some color to what you were saying, we ran the numbers. In 2023, investors spent about 30% less in investment than the average over the 5 years prior, [indiscernible] the data from S&P Capital IQ, whereas AI companies, meanwhile, claimed almost 2 in every $10 with generative AI, specifically making up most of the year's biggest rounds.

Eric Hanselman

Wow. Yes. It's a little different than the nature of investment in years past, wow.

Melissa Incera

Yes. So a significant story to tell in terms of the share of the market captured by this new wave of technology.

Eric Hanselman

Well -- and I think the thing that you're pointing out is really the important part, which is that this is a horizontal technology that can touch so many different pieces of the tech stack broadly that it does seem to be something that, certainly from an investment perspective, there is an expectation that it has such great potential. If you take a look at total markets that could potentially touch and this is pretty massive, and I guess maybe meriting it.

But I was curious, you've been doing some research around strategic investors as opposed to the venture capital markets, private equity investment bank, banking sides of things. And it really seems like there are differences in terms of how the strategic investors are approaching this because, again, it gets back to being sort of must-have technology, and the strategics are clearly enthused. Yes?

Melissa Incera

Yes, absolutely. So yes, as you said, we're kind of drilling down specifically into strategic investors and how they're really participating in the investment landscape in ways that they have never seen before. So just to kind of put into perspective, I think probably helpful to get into the mindset of a strategic investor at this point in time. So -- and I won't belabor the impact of generative AI, but really high level, the narrative is we've seen this birth of a new generation of technology whose impact on the world we can't really understand.

It's coming incredibly fast. I was on a call just yesterday with a vendor who was saying that as soon as the end of this decade, we'll have cognitive agents that basically understand how you work and can execute autonomously all of your workflow, so massive societal impact. We also have the financial impact. Our projection is that this will be a $40 billion market within the next 5 years, and that's actually quite conservative. It doesn't account for any infrastructure.

Eric Hanselman

There are some estimates out there that...

Melissa Incera

That are pretty wild.

Eric Hanselman

Yes.

Melissa Incera

This is a relatively conservative one.

Eric Hanselman

Well, it depends upon the -- what's the value you put on destroying the planet. It's like -- yes. Or some interesting scenarios, let's say, out there, but...

Melissa Incera

Totally. Totally. But I would say, from a strategic investor perspective, this is really important for a couple of reasons. So one, these things tend to come in cycles, so for instance, cloud computing, the rise of SaaS, e-commerce even. And there are a lot of very big established vendors that have missed the boat on some of these and may never really come back from it because it just completely overhauled their way of doing things.

So I think a lot of these players understand you either become generative AI, you accommodate generative AI, you adopt generative AI faster than others or you might not make it, but also adding to that pressure is this is a space that's incredibly nascent. So many of these vendors are still trying to figure out where the real opportunities are. I would argue that the current funding distribution doesn't map to where we expect the value to be even 5, 10 years from now.

So we have all of these companies, big and small, trying to figure out how those impact me, how do I get on the right side of this, what players do I need to know, and they're participating in this at unprecedented rates. And to contextualize that a bit further, they're thinking about their investment footprints completely differently than we've seen historically.

So last year was a really wacky year for M&A generally, as I mentioned before. So a lot of these guys replaced the spending that they would have probably been doing on acquisitions with shifting a lot more cash into minority investment opportunities. For instance, NVIDIA, they're exploding stock was the story of the year and it still is. And I was looking back at our MAKB, which tracks tech transactions since 2002. They have done at least 1 or 2 deals every single year since 2019. And last year, they didn't do a single acquisition.

They did, however, make 38 capital placements between what they did as a corporate entity and the other venture arm. Microsoft, same deal, not a single majority investment but notched its second highest count of corporate investments ever, 1 of which was its astounding $10 billion commitment to OpenAI in January. So they're playing this field a lot differently than we've seen them take it.

Eric Hanselman

Yes, traditionally in the past.

Melissa Incera

Right. Right.

Eric Hanselman

So I think the important point you're making there is that these are actually a significant shift in their investment patterns that they're taking minority stakes. They're -- clearly, they know they have to be there, but it's gone from wholesale acquisitions to kind of placing bets in various places. What does that do in general to a lot of the start-up environment? Is that having an effect on that?

Melissa Incera

Yes, great question. It's very interesting. So just to kind of put some color on what the impact is on these start-ups, so one example, Hugging Face is a generative AI unicorn. It's a platform for sourcing and collaborating around open source models, data sets, et cetera, raised a ton of money. But what's unique about Hugging face but also generally in start-ups in the space, tech giants make up virtually half of its cap table, right? So I think there are 8 strategic backers in Hugging Face, and this includes NVIDIA, IBM, Intel, AMD. By comparison, a late-stage player like Databricks, which just raised its Series I this past December, counts only 5 corporate venture stakeholders out of a total 60. So this can be a great thing for a start-up, especially if these partners are subsidizing hardware or compute costs, which is obviously one of the main challenges in scaling.

Eric Hanselman

Yes. We saw that with Microsoft and OpenAI and that they're giving them a lot of cloud credits to do the computational work.

Melissa Incera

Yes. Exactly. But on the other side, and perhaps this is more of a challenge for the strategic investors themselves, but this can create quite a tangle of interests.

Eric Hanselman

Yes. I was looking at -- so for our listeners, I will point them to research on this. There's a great chart that you put together that is looking at this tangled web of all of these various investments from the various strategics in all of these different players.

And as you're saying, they're placing bets with multiple players. But a lot of cases, you potentially got an individual start-up that may have competitors as part of that investment portfolio that creates an interesting dynamic and certainly something that's, I guess, out of the ordinary for that regular start-up environment in which normally you've got venture investors, maybe some private equity. You've got folks who are investing for the purpose of return as opposed to investing for the purpose of some level of outcome benefit, access to technology. That sounds like that really shifts the normal operating environment.

Melissa Incera

Yes, exactly. And the challenge, as I said before, from the perspective of the strategic investors, is that this is such an impactful technology. There's still a lot of moving pieces and a lot of them also have interests at many different parts of the stack.

So they're looking for 2 things that contributes to this very tangled web. One, they're looking for influence; and two, they're also looking to diversify. So this is a very expensive dance that not a lot of people can afford. And it's putting particularly some of the biggest investors, biggest cloud service providers really in contention with each other. And there are some really interesting things happening. Where are they placing their bets? What are their strategies there?

Eric Hanselman

Well, absolutely, because when you think about what their motivations are, you've got -- you mentioned Microsoft, certainly, all the big software players in various forms. Microsoft, Google, of course, has all sorts of capacity themselves. Pick your particular environment, Amazon, of course. And then you've got a whole bunch of the semiconductor players, NVIDIA, AMD, Intel. This is a wild mix.

Melissa Incera

Yes, absolutely. So I would say the biggest investors in the space right now are -- it's kind of where the drama is accumulating. So we see...

Eric Hanselman

Drama? Drama?

Melissa Incera

So we have the big hyperscalers you would expect. You've already basically named Microsoft, Google, Amazon and NVIDIA, too. These are the most prolific investors in the space right now. Actually, to put into context, one of my colleagues did a great deep dive into NVIDIA. They more than 10x-ed their investment in nonaffiliates in the first 9 months of 2023.

Eric Hanselman

Wow.

Melissa Incera

I think they went from [ $8 million ] a year to over $100 million. So...

Eric Hanselman

On order of magnitude?

Melissa Incera

Yes. Yes, they are in it. But their strategies are a bit divergent. I'll also caveat this with, these guys, they are the best situated to be making investments in the space, one, because they have the cash. They also are super attractive partners and that they help these start-ups to codevelop. They help subsidize GPUs, cloud resources, et cetera. So a lot of these deals, very mutually beneficial for investor, investee, right?

But let's dig into what their different strategies are. So Microsoft, very unique in that it's an exclusive strategic backup of OpenAI. They have committed, as we mentioned, $10 billion for that, and they've paid handsomely. But as of now, they have no strategic competition behind that company, which gives them a ton of control, and that's something that they are appearing to seek out.

So they've done the same with another start-up called Builder.ai, and also another company is Inflection AI there, -- in there -- with NVIDIA. So they've paid through the nose in these select opportunities for this degree of influence. We'll see if that stands as many of these guys have to continue raising.

Eric Hanselman

Yes, as they'll go for additional rounds, their potential is for other investors to pile in and dilute some of Microsoft's influence.

Melissa Incera

Exactly. On the other side of the equation, Google and Amazon have gotten themselves into a bit of what we would call a proxy work for another foundation model for the company Anthropic, so really interesting. Initially, Google was in there. Allegedly, the rumors are that they had some trouble meeting Anthropic's needs on the cloud services front. So next round, Amazon came in committing $4 billion. Theoretically, it was going to be a preferred cloud provider. And then literally a month later, Google ponied up with another $2 billion. So these guys are very actively trying to meet each other and you see a degree of competitive nature, right?

Eric Hanselman

Yes. A back and forth here. Well -- but you pointed out this interesting sort of situation in which you've got OpenAI that's exclusively Microsoft. So now you've got other players that are, I guess, I wouldn't say battling yet, but at least starting to compete for influence in others in which there's an angle. And yes, clearly, they think the stakes are high enough to put a lot of investment dollars behind this because, yes, that's -- tossing a few billion here, a few billion there that just seems to have become normalized in this environment. Yes.

Melissa Incera

Table stakes.

Eric Hanselman

Crazy.

Melissa Incera

So these guys, they have lots of investments, I'll say, but they're really kind of selecting a horse in the race, whereas somebody like NVIDIA has steered clear of these top 2, Anthropic, OpenAI. Those are the 2 most highly valued generative AI start-ups in current time. But NVIDIA has made up for it in volume of placements.

So when we looked earlier this year, almost half of all generative AI unicorns have NVIDIA as a backer. And they are from, what we are hearing, an incredibly coveted partner, both because of advantages from an infrastructure perspective, but they also do a lot of co-development. So they have paved themselves a nice path in this space with lots of different significant bets happening.

Eric Hanselman

Yes, because they're basically backing as many horses as they can. They're not going for that single winning bet. They're backing a lot of horses and hoping they win. And of course, as you're pointing out, it's not just that. Yes, they're the manufacturer of like the most popular GPU today, so wouldn't it be nice to have a strong connection there. But the fact is that they're also helping to drive a lot of the use cases. And that's actually Nick Patience and Alex Johnson are going to be on in a couple of episodes to actually talk about some of the use cases.

But I think it's interesting that you're pointing out this fact that it's being an active participant in that ecosystem that's actually creating demand that they're actually helping to build the use cases. They've got businesses that were coming to them who are looking to go solve some of the thornier problems, which is great for start-ups because they can say, oh, we've got somebody who can tackle that kind of problem here, and being plugged in to that ecosystem has a lot of value to it for the start-ups who want to get connected to NVIDIA to be part of that pipeline of potential opportunities.

Melissa Incera

Yes, absolutely. I'll also call out, there are 2 other groups of strategic companies that are playing in this space, maybe not to the same degree as mentioned before. But on the one hand, you have the other cloud infrastructure semiconductor companies, which also have a ton at stake, perhaps not as free wheeling or cash rich to be investing as prolifically as those 4 at the top of the market. So many of these guys are betting on 1 or 2 unicorns.

One example is Oracle, who has basically tied its wagon to Cohere, HPE in Europe, SAP as well. They've invested in a European-based foundation model company called Aleph Alpha. So it's definitely a lot of corporate investment happening.

Another avenue that I haven't mentioned, where virtually all of these players are leveraging, is the corporate venture capital arm, where they're investing in earlier-stage opportunities. And application software providers, this has really been the avenue of choice for them because, yes, they're excited about integrating with these foundation model providers, but they're also watching the [ secret ] unfold at the application layer.

For instance, Salesforce Ventures announced a $500 million generative AI fund this past August, which is then used to lead Hugging Face's Series D round. So these guys have a massive amount of firepower behind them, but they are investing a little bit more at an arm's length currently.

Eric Hanselman

So this is, again, moving up the stack when you got someone like a Salesforce as a SaaS provider starting to look at this. Now we've moved out of the infrastructure, the core semiconductor providers, now starting to get into that next stage of the actual end users want to be able to put this -- the capabilities to work in various forms.

Melissa Incera

Yes, definitely.

Eric Hanselman

A lot of different players. This is a complex space. I'm curious -- I mean, the other thing that's sort of hovering out there are the discussions about regulations and what's going to happen and is this going to have an impact on it or are we going to shut down certain pursuits with generative AI. I mean we've already got activities to regulate and constrain Europe. Think about data security and GDPR regulations. They happened in Europe first. Then, eventually, they rolled out globally. Are regulators going to start to throw a little downer on this party? Or what else is going on?

Melissa Incera

Yes, this is a big X factor in the coming year. So to help contextualize this, the amount of funding that's been happening is incredible in terms of the size of some of these commitments, and a lot of these guys are tossing out dollars that match some of their largest acquisitions ever in the past.

So Microsoft has only ever done 3 deals equivalent or larger than the $10 billion it committed to OpenAI earlier this year. So it follows that there's been an increasing amount of antitrust action happening initially out of Europe and U.K., now the U.S., and they're probing not just the relationship between Microsoft and OpenAI, but now I think FTC is looking at Google and Amazon and Anthropic. And really, the question is should it be subject to merger rules. The concerns here.

Eric Hanselman

Wow.

Melissa Incera

Yes.

Eric Hanselman

Because the transactions are so big that they're effectively an acquisition, and oh, boy.

Melissa Incera

Yes. And one of our analysts wrote that Microsoft is believed to own 49% of the cap profits portion of OpenAI and OpenAI's generative AI work is done exclusively on their cloud platform. So their 49% is bridging on [indiscernible], right?

Eric Hanselman

Yes, not quite a controlling interest, but yes, we're pretty gosh darn close.

Melissa Incera

Right. And the concern here is that there's such a high cost to getting these models up and running. They take enormous amounts of computing resources, and these hyperscalers are best positioned to provide that. So as they inch closer to these start-ups, the fear is that leads to a market where the highest quality pretrained models are controlled by a really small number of players. And they also control the compute, which is the biggest factor in being able to scale up, support customers and impacts the start-up landscape as well.

Eric Hanselman

This starts becoming a barrier to entry and potentially and excluding competitive pressure that you have to be a hyperscaler in order to simply have the compute resources necessary to do the training. We're not in the typical sort of, hey, any start-up can dive in and do this stuff. You actually have to have massive computational resources or the ability to pay for it to be competitive.

Melissa Incera

Yes, absolutely.

Eric Hanselman

So how do you see this shaking out? It seems like there are so many different forces that are in play, so many different aspects that some conflicting, some supporting. Where do you see this headed?

Melissa Incera

I think, initially, we're going to see the cracks that we're seeing now as we were chatting about regulators and the tensions that we're seeing between strategic backers. These are going to steepen because these guys are not done raising. They need more cash. I think both Cohere and Anthropic are expected to announce a round of funding in the next couple of months.

This could be a rumor. I was reading that Anthropic is raising money not because they were in desperate need of it, specifically because one of their VC partners wants to up their stake before a future round. So there are just so many interests behind these guys. They are literally being thrown cash.

But another dynamic that's important, I think, to call out is the economics of this space are going to change pretty dramatically in the next few years. So right now, the majority of investment is going to these foundation model providers mostly because the cost of developing models are so high. But these products are or rapidly will be at a point where they can match each other in terms of performance, and as an end user, there's not really a perceptible difference between them.

And what's more, they're starting to face more competition from all of these smaller specialized models coming out of the start-up ecosystem that are smaller, cheaper to run. And what that means is that some of these generative capabilities like text to text, for instance, will commoditize, not to mention that compute costs are going down rapidly.

So I think we're going to see some of these guys are going to have to pivot. They're going to have to focus on these bigger challenges in the AI space that they alone are well financed enough to pursue. Like AGI, artificial general intelligence, is something that only the biggest and best-funded start-ups will be oriented towards.

I can see a lot of these guys shifting towards infrastructure, right? I can envision OpenAI building out an infrastructure play, just to own more of that stack, but a lot more focus from an investment perspective will be given to companies that have the specialized infrastructure services that focus on implementation and kind of productionizing this stuff. I mean who knows what's going to be born tomorrow?

Eric Hanselman

Yes. So you start looking at maybe vertical industry specialization, although I think Nick had talked about the shift from a lot of the upfront training to inference. And to your point, once you get an inference, you're looking to get models, typically smaller models that can operate efficiently, as you're saying, that don't cost an arm and a leg to run. And now you're getting into much smaller specialized environments. And so the days of the great massive, does-it-all kind of model maybe start to wane at some point.

But clearly, as you're saying also, there still is a lot of money that's sloshing around here because, again, in the near term, the expectation is that the computational costs are so high that you need to -- at least for any of the major players, you need to keep feeding that, and that's going to take more cash. So I guess still a lot of cash to burn through first.

Melissa Incera

For some, yes. Another projection I'll call out, so this past year or 2, most of the discussion has been on funding because that's where the space is right now. It's very nascent. You need cash. But actually, I mean, in the last 3 months, we're starting to see more tendrils of M&A, specifically because the space has just exploded, and there's no way the market can sustain the number of start-ups that have popped up in the last few months, especially with incumbents starting to integrate this stuff too.

So we expect to see consolidation happening in the next few months. It started already. Late last year, Adobe bought a video generator. Jasper AI just yesterday, as text generation company, bought image capabilities from one of its partners, Stability AI. So a lot of these vendors are already reaching to add more multimodal capabilities. They're trying to cement their positions early. The winners and losers will start to emerge. So the next 6, 12 months are going to be really interesting, specifically because there's going to be a lot more to talk about than just funding.

Eric Hanselman

As we see with many things, we start to get into that platforming movement. Everybody wants to be the place that has all of the things that you need to do the thing. And again, that leads to consolidation and especially in something that's as frothy as everything that we've seen, clearly, there are opportunities because some folks are going to get funding. Some are going to have the opportunities and see where that drives the market.

Melissa Incera

Yes, absolutely. It's moving quickly. So this is definitely an accelerated curve we're seeing for generative AI at the moment.

Eric Hanselman

But at least for now, the party keeps on rocking and rolling. Well, this has been great, Melissa. I guess we'll have to see where this goes and check back in as we start to develop. But thank you. This has been wonderful.

Melissa Incera

Yes. Thanks for having me, Eric. This was so much fun.

Eric Hanselman

Well -- and especially when there's so much that we're looking at. But for at least today, we are at time. So that is it for this episode of Next in Tech. Thanks to our audience for staying with us, and thanks to our production team, including Caroline Wright and Kaitlin Buckley on the Marketing and Events teams and our agency partner, the 199.

I hope you'll join us for our next episode, where we're going to be talking about what we saw at Mobile World Congress, which is actually just getting underway because there's so much happening there. We've got a whole crew that's going to be actually out in Barcelona, and we'll be reviewing all of that. I hope you'll join us then because there is always something next in tech.

Copyright © 2024 by S&P Global Market Intelligence, a division of S&P Global Inc. All rights reserved.

These materials have been prepared solely for information purposes based upon information generally available to the public and from sources believed to be reliable. No content (including index data, ratings, credit-related analyses and data, research, model, software or other application or output therefrom) or any part thereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of S&P Global Market Intelligence or its affiliates (collectively, S&P Global). The Content shall not be used for any unlawful or unauthorized purposes. S&P Global and any third-party providers, (collectively S&P Global Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Global Parties are not responsible for any errors or omissions, regardless of the cause, for the results obtained from the use of the Content. THE CONTENT IS PROVIDED ON "AS IS" BASIS. S&P GLOBAL PARTIES DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT'S FUNCTIONING WILL BE UNINTERRUPTED OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no event shall S&P Global Parties be liable to any party for any direct, indirect, incidental, exemplary, compensatory, punitive, special or consequential damages, costs, expenses, legal fees, or losses (including, without limitation, lost income or lost profits and opportunity costs or losses caused by negligence) in connection with any use of the Content even if advised of the possibility of such damages. S&P Global Market Intelligence's opinions, quotes and credit-related and other analyses are statements of opinion as of the date they are expressed and not statements of fact or recommendations to purchase, hold, or sell any securities or to make any investment decisions, and do not address the suitability of any security. S&P Global Market Intelligence may provide index data. Direct investment in an index is not possible. Exposure to an asset class represented by an index is available through investable instruments based on that index. S&P Global Market Intelligence assumes no obligation to update the Content following publication in any form or format. The Content should not be relied on and is not a substitute for the skill, judgment and experience of the user, its management, employees, advisors and/or clients when making investment and other business decisions. S&P Global Market Intelligence does not act as a fiduciary or an investment advisor except where registered as such. S&P Global keeps certain activities of its divisions separate from each other in order to preserve the independence and objectivity of their respective activities. As a result, certain divisions of S&P Global may have information that is not available to other S&P Global divisions. S&P Global has established policies and procedures to maintain the confidentiality of certain nonpublic information received in connection with each analytical process.

S&P Global may receive compensation for its ratings and certain analyses, normally from issuers or underwriters of securities or from obligors. S&P Global reserves the right to disseminate its opinions and analyses. S&P Global's public ratings and analyses are made available on its Web sites, www.standardandpoors.com  (free of charge), and www.ratingsdirect.com  and www.globalcreditportal.com (subscription), and may be distributed through other means, including via S&P Global publications and third-party redistributors. Additional information about our ratings fees is available at www.standardandpoors.com/usratingsfees.

© 2024 S&P Global Market Intelligence.