a16z's Investment Review: 86% of VC Deals Now Flow to AI, and the Trillion-Dollar Feast Has Only Just Begun

Z Highlights

• The gap between what models can do and how they are actually used is exactly why I'm so excited about the application layer.
• Among companies in the top 1% of AI tool spending, median spend is roughly eight times that of the top 10%.
• For founder-led companies, the founder is itself the asset class.

This article is compiled from the official a16z podcast "AI, Infrastructure, and the Next Investment Cycle." David George, Sarah Wang, Alex Immerman and Santiago "Santi" Rodriguez are four partners at a16z's Growth Fund. Using the 25 key charts in the annual State of Markets report as a thread, this episode discusses the earnings behind the market rally, the hyperscalers' compute arms race, the extremely low adoption rates at the model and application layers, and the latest observations on consumer, the public software market, IPOs and the secondary market. The episode was published on September 30, 2026.

01 Technology Is the "Everything Cycle"

David: Welcome back to the a16z podcast. I'm David George, and today I'm joined by my colleagues Sarah Wang, Alex Immerman and Santiago Rodriguez. In this episode we'll walk through the 25 key charts from the latest State of Markets report: the earnings behind the market rally, the scale of the AI buildout, the evidence of adoption growth, and what this cycle means for hardware, software and the next generation of private companies. We'll explain what's behind the charts, and we'll also talk about what we're seeing on the front lines at companies. Sarah, Alex, Santi, thanks for coming.

On the day this podcast goes live, we'll also publish the full State of Markets report. It's an annual tradition for our growth team, pulling the biggest trends in tech, AI, infrastructure and the broader market into one document, and today we're picking out some of the interesting charts to discuss. If you want the full picture, be sure to read the complete version — there's a lot of substance in there. The areas we'll cover today include the macro — capex, data centers, accelerating demand; and front-line observations — deep dives on models, applications and a few verticals. Technology is driving an investment boom that runs across the entire economy: earnings growth is supporting the market rally, and AI demand is pushing infrastructure spending far beyond previous forecasts. Hyperscalers are pouring nearly all of their near-term operating cash flow into capacity buildout, and demand continues to exceed supply in almost every case we see. That money flows into chips, power, cooling, construction and skilled labor, and much more. This expansion coincides with the larger need to rework our physical infrastructure, and it's creating opportunities across every industry.

Alright, let's dive right in. First page: technology is the "everything cycle." This part may be somewhat obvious, but the numbers on this page look quite astonishing today — high-tech equipment, software and R&D now make up about 55% of US capital spending, which is a staggering number. Technology drives investment cycles in every part of the economy, from software and models to power, construction and industrial capacity. Tech stocks account for about 40% of total US stock market value, and eight of the ten most valuable companies in the world are American tech companies. This is a grand narrative — you can see it in capex, you can see it in data center construction, and you can see it in the capital flowing to model companies, which together have raised more than $350 billion. Put it in longer historical context, because this is the most cited analogy: the scale of this buildout as a share of GDP has already surpassed the railways of their era. We all have a feeling: if you fast-forward five to seven years, or even ten, and look back at these numbers, the cumulative figure will very likely be more than 20 times what it is today. So this really does feel like the next chapter of Marc Andreessen's "software is eating the world."

Alex: Over the past 15 years, software has changed all of these industries, but back then only a small group of people could build software — none of us could. Today, look at us: every one of us has several automated tasks running every night. If you extrapolate software demand along that logic, it means far more compute, chips, power and construction. So it's no surprise that most investment is now concentrated in technology. Our listeners often ask us the same question: this is all great, but is it a bubble? The market has hit new highs, yet the market's current trading multiple has actually come down — stocks are up about 20%, but multiples are down about 20%. That means the gains are driven by fundamental earnings, not multiple expansion. The S&P 500's earnings multiple is below 20x, and most of the companies that make up the S&P 500 are very high-quality businesses. This is completely different from the dot-com bubble — back then, the most valuable companies in the world saw their share prices surge on rising trading multiples, and many traded at 100x earnings. That is not the situation at all now. By contrast, some memory companies — which are inherently cyclical — trade at six to seven times forward earnings. So the difference is enormous.

Santi: There's an important addition here: since GPT arrived nearly four years ago, the market is up 90%, or 17% annualized. When you see four consecutive years of 17% annualized growth, the instinctive reaction is "this has to pull back," and we are indeed in a hot period right now. But when you put that together with what we just described — the market trading at below 20x earnings while earnings are still growing at 15% — this does feel different from 2021 recently, and different from 2000, when multiples and growth weren't really in sync.

David: Completely agree. I mentioned the scale of the capex buildout earlier with the "railway analogy." Just look at the hyperscalers — Alphabet, Amazon, Meta, Microsoft and Oracle — their 2026 capex is about $780 billion, versus $416 billion in 2025; all expectations point to more than $1 trillion a year from 2027 onward. There's a pattern here: every generation of computing platform has supported far more users and use cases than the one before. This buildout can happen so fast while demand still exceeds supply because the user base is driven by distribution channels that already exist. This trend builds on the internet, cloud computing and mobile phones, so in theory it can reach billions of users right away — a sharp contrast with previous technology cycles.

Sarah: I think that's worth expanding on. We've moved far beyond the "occasionally ask a question" mode. Think about the rise of agents: parallel tasks, long-running tasks — as you mentioned earlier, tasks keep running while you sleep at night or do other things during the day. If agents can autonomously write code, do search and execute these tasks, then the compute these platforms need will far exceed anything we've imagined before.

David: Completely agree. This chart shows what's happening on the capex side: the capex forecasts of the five major hyperscalers have been repeatedly and substantially revised upward in a very short period. Spending that once looked like a ceiling — something that would only be reached several quarters or even years later — has now become a near-term target, because compute demand keeps swelling. Looking at this chart, I think at any given moment the instinctive reaction is "investment will plateau from here." But over the past four years, as an economy, we have consistently underestimated the strength of this trend. As Sarah said, as models evolve, utilization of installed capacity keeps getting maxed out. So the numbers David just mentioned are the actual current estimates.

Alex: To use DG's phrase, we'd all call today's compute demand a "model buster." One very telling example we've all lived through: about a year ago, Sam Altman and Sarah Friar took a lot of heat over a massive compute commitment, and everyone thought they were reckless and aggressive. Now everyone would say that decision was extremely prescient. Even so, just two weeks ago, we all saw OpenAI have to pause subscriptions to its Pro plan — a $2,000-level service, and still demand outstripped supply. What we're experiencing is a kind of demand that is hard to satisfy.

David: Everyone at every link in the supply chain, everyone we've talked to, tells us a different version of the same thing: demand exceeds supply. Some parts of the data center supply chain involve materials or products you can't get until 2028. So this tightness isn't going to ease. At the same time, you can see evidence of high-quality business on the demand side: Microsoft, Google and Amazon together have about $1.7 trillion in cloud backlog. These customer commitments are accumulating fast while the platforms invest heavily to serve that demand. So while free cash flow is under pressure during this buildout period, consensus forecasts show it recovering from 2028 and then growing significantly. You can look at Amazon's most recent earnings call, where they did a good job explaining this J-curve dynamic: the actual useful life of a GPU (or TPU) is in fact quite long. You have to build the data center "shell" first, and that takes time; then you buy chips, but the economic life of those chips is long — longer than any of us expected.

Sarah: Yes, I know some hyperscalers have been criticized for taking on debt to fund capex, but similar to the dynamic with the model labs, those who were more conservative out of hesitation all regretted it later. In that recent podcast you did with Gavin you talked about Microsoft, but I think this is an industry-wide issue.

Santi: Yes, and another signal that gets discussed a lot is the spot market price of existing GPUs: any GPU you can bring online can be rented out at an attractive price, and hyperscalers can earn considerable returns. This is the key point we often discuss: every wave like this creates enormous user and consumer surplus. From what we've seen so far, consumers and users are getting tremendous value out of it — which is exactly why they use it so heavily. The companies serving those users are doing very well. Then you go layer by layer down to the chips — chips cost much more to acquire than they did 12, 18, 24 months ago, but you can still earn high gross margins and create a lot of surplus for users.

02 A New Era of Atoms

David: We should think of hyperscaler capex as somebody else's order book. These historically most profitable platforms are steadily putting cash into AI infrastructure. As just said, that depresses their free cash flow in the short term, but it's a huge boon for chip orders, power and construction — which is why this broader tech boom has become an industrial boom. We spend more and more time looking at the companies serving this industrial boom, along every link of the data center supply chain. What's interesting is that our world traditionally was more about monetizing existing IP — build once, sell infinitely; but these businesses have far more complex things to manage: financing, managing supplier relationships, forecasting capacity, forecasting demand. We see some teams that are good at this, and others that struggle or even pause. It's a bit different from the expertise we usually spend time on. This brings us to "a new era of atoms." This page carries the big numbers: by 2040, global infrastructure investment demand is estimated at $90 trillion. But importantly, that goes far beyond AI and data centers — it also includes power, water, roads, transportation. We see this everywhere in our portfolio: Anduril's enormous manufacturing plant, the size of 87 football fields; Waabi aggressively expanding its transport hubs; and SpaceX's hundred-billion-dollar-scale investment in Louisiana. Betting on the future today, from our perspective, looks different than it did in the past. As Santi said, it means building factories, expanding infrastructure, and adding skilled jobs across the country.

Santi: Speaking of data centers, investing in physical infrastructure is very different from investing in software. Elon put forward the idea years ago that "the factory is the product." Many of the great founders we back come from Tesla or SpaceX, which isn't surprising, because they know that at the execution level, the factory ultimately becomes a huge competitive advantage in scaling. We often hear a lot of myths about data centers: they'll drain America's water; rich people don't want to live near them; and our favorite — my electricity bill is going to skyrocket. But perhaps counterintuitively, data centers can actually help lower your electricity bill. A recent US study showed that for every 10% increase in data center capacity, residential electricity prices fall by 40 basis points. Think of it this way: the grid is a shared fixed-cost space — poles, wires, substations. A large, stable customer like a data center can spread those costs across more units of electricity. Simply put, more demand on a shared system is a good thing.

David: Yes, I completely agree with what you're saying. Investment in shared infrastructure can also improve the economics of the homes and businesses connected to it.

Sarah: This doesn't get enough media attention. Dina Powell recently mentioned on a podcast Meta's Louisiana project, where they work with the community to lower electricity prices. There are real examples today — this isn't just talk.

03 The Model Layer and the Application Layer: Substantial Revenue, Extremely Low Adoption

Sarah: Completely agree. Now let's talk about the trends at the model layer and the application layer. On one hand, as we previewed, AI is generating substantial revenue and savings; on the other hand, adoption is still extremely early. On that basis, the trend is very clear: costs are plummeting. We previewed agents, and now agents are becoming economically viable across a wider range of work. And even as we speak, the landscape is still shifting: new innovations such as Jev, from new labs like TypeSafe, are pushing this even further. You'll see that a two-orders-of-magnitude cost difference really does affect the number of use cases. It's a bit like the classic Jevons paradox, so the name is quite apt. Overall, this is a fantastic setup for both the model layer and the application layer. These improvements are creating real opportunities for software companies in both growth and profitability. It's not that everyone will win, but unsurprisingly, AI is drawing venture capital into an ever-wider range of industries. We're also seeing more companies reach massive scale while still in the private markets. This isn't a new chart, but it's one of my favorites: the combined revenue scale of OpenAI and Anthropic. Over the past few years we've put this chart up in every internal GP meeting, updating it every month because it goes out of date so fast. The key point is right there in the numbers: the combined annualized revenue of OpenAI and Anthropic has grown to an extraordinary degree. We like to compare it with the greatest software companies in history — on the right you can see that even by estimates of the net new revenue of the best software companies in history, the leading labs' net new revenue has already far surpassed them. It's not just the scale — it's the striking combination with speed.

Santi: But we're still early. Compared with other platform migrations, what's interesting about the AI migration is that we're early not only in the share of the population using AI, but also in those users' share of wallet for AI. This is different from 2008 — back then we modeled on "the share of people who own an iPhone"; now we have to consider both how many people will use AI products and how intensively they use them.

Sarah: Yes, the "still early" point is very important, and here's a set of metrics to back it up. First, actual deployment among S&P 500 companies is 69% — if you're an "AI addict" like us, that's probably close to what you'd expect. But when you look at quantifiable impact — which may be a good indicator of their deployment progress — it's 30%. And then the ultimate metric, one tracked over time, is actually only 2%. So we believe AI is delivering results, but there is enormous room to deepen usage inside organizations and to track those results over time. Moving from these scattered deployments to workflows with real, deep impact that recur — frankly, that's the next phase. I'd say that for most of the enterprises we've spoken with, their exposure to AI still mostly stops at Microsoft Copilot, which in itself shows how far they have to go.

Alex: Yes, the gap between what models can do and how they're actually used is exactly why I'm so excited about the application layer. Sarah, that conversation you had with Ali Ghodsi and Martin of Databricks was great — he talked about how AI can know the world very well but almost nothing about the company you work for. We're seeing exactly that in the application-layer companies in our portfolio — they're closing that gap. One example in the past few months struck me: Revolut has a very mature engineering team, but they still worked with 11 Labs to plug 11 Labs' best-in-class voice model into a secure connection to customer accounts and banking processes, so that when I call as a customer, my issue gets resolved smoothly, efficiently and securely. The opportunity today is largely about harnessing these capabilities and building reliable services.

04 Heavy Users Are Pulling Ahead

Sarah: Beyond enterprise adoption still being early, another notable trend we want to illustrate with data is: heavy users are pulling ahead across the board. You could say AI spending is rising broadly across companies, but the heaviest users we're talking about are spending far more. We looked at some YipitData data: among companies in the top 1% of AI tool spending, median spend is roughly eight times that of the top 10%.

What's striking is that the top 1% is almost equivalent to the total of the 2nd through 10th percentiles.

Even within our own portfolio companies — which you could say are almost entirely AI-native — top users spend roughly $7,500 to $9,000 a month. The median user — again at these most AI-native companies — spends closer to the price of a monthly subscription, about $200, maybe plus another subscription, $200 to $400. So if you look at the gap between median and top users, it's more than a 20x difference in spending.

David: Yes, one question we discuss with portfolio companies is: how much adoption do they actually have, and how do you measure it. For example, look at what they spend on AI tools as a share of employee costs: relative to forward-looking large Fortune 500 companies, it's about 1% today.

Whereas the most forward-leaning AI-native companies in our portfolio can be as high as 10%. So it's all about how early diffusion is and how deep it will go. We've seen another example within the portfolio: many enterprises are very excited the first time they buy Cursor or Cognition, but the reality is that that's just the start of the journey. From buying and trialing these tools to full adoption, there's a long runway.

05 Quantifiable Case Studies: The Cost Side and the Revenue Side

Santi: Sarah, you mentioned earlier that we're starting to see quantifiable case studies. In the report we named several public companies that actually disclose these metrics and are seeing real improvement from AI. On the cost side: Chime disclosed that over the past four years they've cut their cost to serve by more than 10% a year, which compounds to an almost 50% reduction in cost to serve — with support from our portfolio company Decagon. On the revenue side, I think the Shopify example is interesting: they launched an AI assistant (sidekick) to help merchants get up to speed faster. As a result, the share of customers reaching five orders within 15 days of onboarding rose 8%. Shopify tracks this metric because once merchants hit five orders, they stay and keep using Shopify. So that's also a meaningful tailwind for their business.

Sarah: Absolutely. And on the advantage that a lower cost to serve gives you: you have more room to compete — better pricing, potentially broader service, and the ability to reinvest in growth. These benefits don't just reach customers, they reach margins, and then the money gets reinvested. Another very fascinating phenomenon: of course every company is different and you can't generalize, but incumbents have also done quite well in this cycle. ServiceNow is a good example — they disclosed AI annual contract value (ACV) of more than $1 billion, with agentic deployments up 9x. So you see it across the whole stack, from incumbents to new companies.

David: We often discuss an interesting question — our most forward-leaning portfolio companies, like Stripe, have talked with us about: where exactly are they spending incremental AI investment? Is it on building new products for customers to drive higher revenue? Or on cost-side optimization and efficiency gains? It's a bit of a touchstone for us: where do these founders and CEOs actually see the biggest opportunity? The upside of driving revenue growth is unbounded, whereas the opportunity on the cost side...

David: Cost-side improvements — sure, you can reinvest that, but it's always just a potential opportunity that will always be there. If you think the best use of capital today is optimizing your cost structure, what does that say about your new revenue opportunities?

Santi: Even within cost optimization, there are different flavors. One is lowering cost to serve, which lets you reinvest and become a better company. I think six months ago our whole industry was really focused on "rebuilding systems of record." If your engineers are busy rebuilding a system of record just to save a few thousand or tens of thousands of dollars, then I think those resources have better uses.

Sarah: Yes, and if you add it all up, for application businesses the relevant unit is really the cost of "completing a customer task." If you can decouple that — you complete tasks for customers and charge for them, and you can make the product more economical with better routing — that's remarkable, because now you're not using the most expensive model at every step, but the task still gets done. I think we'll see more margin improvement from the best application companies.

06 When Costs Plummet, New Markets Open Up

Alex: Completely agree. Sarah mentioned that ServiceNow is seeing massive agentic usage. Agents are already here, and they're executing tasks. Tasks take multiple steps, and multiple steps mean multiple model calls — which explains the 14x growth in agent token usage on OpenRouter. But these steps are getting cheaper. We're seeing caching in use — systems can reuse background information rather than processing it from scratch every time. I saw Hebbia take full advantage of this: their financial chat workloads now run at 10x lower cost. The same budget can now support more work.

Sarah: Yes, and I think your broader point is: lower costs let agents now check for errors and retry, and suddenly a huge range of tasks opens up — where previously inference or tool calls might have been prohibitively expensive. This matters especially in the scenarios where reliability is precisely the deciding factor in whether to use agents at all. Also, we talked earlier about TypeSafe — imagine when costs drop by an order of magnitude, by two orders of magnitude, the number of use cases that opens up is truly hard to imagine.

David: Plus latency improves dramatically. Absolutely.

Alex: Yes, so companies are now thinking more and more seriously about how to optimize latency, or performance and cost, at scale. Take Databricks: they use routing — picking the right model for each specific task — to improve performance and cost. Their intelligent router performs better: it solves more problems at 35% lower cost than the strongest single model. Another approach we're seeing a lot now is fine-tuning: Harvey and Elise AI are examples. Elise fine-tuned a smaller model, cutting costs by 60% with much lower latency, which made some real-time use cases in audio viable. These engineering improvements make AI more useful and cheaper.

07 The Consumer Side: 2% Subscription Penetration and the "Smile Curve"

David: We've mostly talked about usage inside enterprises, but the consumer side matters too — after all, that's where we've had some of our biggest returns and spent a lot of time. We think the consumer side is still early. A recent survey showed that only slightly more than 2% of US households are actually paying for an AI subscription. Maybe subscriptions aren't the best way to monetize most households, but so far that's what it is. One interesting thing about subscription revenue is that it has some of the best retention we've seen on the consumer side. Alex likes to talk about the "smile curve" — there are really very few retention curves that not only flatten out, but actually "smile" upward as the product improves and people come back. But that's what we're seeing with AI, and it speaks exactly to the value subscriptions can deliver to payers.

Sarah: Speaking of subscriptions, what shocks me in these numbers — especially looking at the chart in the bottom right — is how small they are. Think about Amazon Prime, that's more than 200 million households. Netflix is a good example too — 70 million households. These numbers made me deliberately double-check, and I thought, "did we get this wrong?" So I completely agree, this is just the beginning.

Alex: Yes. Our partner Josh Elman has a saying I like, and it's maybe a bit counterintuitive: consumer AI is something I use in my daily life, not necessarily something I expensed. So in his view, it's perhaps not surprising that 97% of households use AI but haven't paid for it. You're right, Santi, I do like smiling retention curves — we like to see users coming back more and more often. Another common feature of big consumer platforms is time spent: active users on Facebook, Instagram, TikTok and Snap spend 30 to 60 minutes a day. And today's best AI assistants — many of which are rising right now — all want to be my go-to app. I want them to be extremely persistent, always on, super proactive, and in that case, in the future they may not be where I spend the most time. True, our jobs get harder when agents work in the background and we no longer look at engagement through external data the way we look at screen time on social media, but it matters. I remember when the latest model was released in December, even in our own work we'd kick off a deep research task or an agent task and then drive into town — that doesn't show up in traditional screen-time metrics, and it's different from the engagement previous generations of consumer products tracked.

Santi: Yes, completely agree. Still early, but the consumer side is truly a brand-new and exciting moment.

Alex: Muse and Instinct are certainly new players, but they're growing fast. Together with ChatGPT, they're increasingly taking my queries away from traditional search.

This is a very dynamic moment, so if you're an existing consumer discovery platform, an existing marketplace, you really need to think about the next step right now. I'd ask a few questions. First, how much incremental demand can I get from AI agents? How many additional orders can they bring? Second, what share of my business's profit pool comes from ownership of the customer relationship and discovery? There was a lot of news last week: Amazon told Muse "no, thank you," but Instacart said "yes, please." Think about Amazon (and AWS) and Instacart: their ad revenue exceeds all of their operating profit. So ad revenue, ownership of the customer relationship, is extremely important. But for Amazon, how many incremental new orders or new customers would adding Muse bring? Not that many. Look at Instacart: online penetration in groceries is still relatively early, and there are a lot more orders to be won. The optimistic possibility is that more orders will come from these apps. In the past I had to click my way through in various ways to place my next order; if I hand that to an agent, suddenly there are no clicks, but hopefully there's more GMV. Trips I thought about booking but didn't will get booked; the dinner order we want tonight will get placed. So I'm optimistic that there will be more GMV, offsetting part of the lost ad revenue.

David: But there's an open question here: how does it get compensated? Take Amazon, as you said — it has a $70 billion-plus advertising business, extremely high-margin flowing monetization, built entirely on the premise that consumers visit the site and click on ads. What if they stop doing that? Today, Meta and Google are famously better at advertising than any internet platform, making more than $200 per user in the US and developed markets. Meta is basically an entertainment app; let's see what happens, it may be relatively safe. On Google's side — interestingly, two years ago when this all first happened, the whole narrative around Google was "oh my god, what will happen to Google's search business," and it turned out to be very resilient. Partly because their high-monetization ad words are relatively safe — those words are more like "I need to buy insurance" or "find me a hotel in this city."

David: Or something like that — you can monetize it directly, but AI can't yet do those things for you. If that changes, it could be a very different dynamic.

David: Yes, I'd add one point: last week's narrative had a bit of a "negative-sum" flavor, very negative — "look, these marketplaces are going to be hurt." But I do think the positive-sum view is: first, that $70 billion spent on Amazon ads might just move elsewhere through a different channel — maybe spent directly on agents in different forms, or finding better ways to reach people; second, as friction falls, we might see more consumption — people might buy more, which is a positive flywheel driving economic growth. I think we're a bit trapped in the "this is bad for profit pools" framing, when actually it might just be good. ... Yes, both can be true at the same time.

David: Maybe profit pools will shift, but overall it's incremental to the economy. I think if we look back in three years and there hasn't been meaningful efficiency improvement at the macro level — which would push consumer spending in advertising higher — we'd all be disappointed. But it may also just be that profit pools in some places disappear and then get reallocated.

Alex: The immediate reaction is that the ecosystem's net market value is positive. Meta's gain far exceeds the marketplaces' losses.

08 The Divergence in the Public Software Market

Sarah: Yes, coming back to our discussion of software, maybe shifting more toward the public markets. The most striking trend in these charts is: the composition has clearly shifted toward slower-growing, more profitable companies. About 75% of this public software sample is profitable, while only 30% is growing at 20% or more.

David: This is a staggering number: only 30% of public software companies are growing at more than 20% — and not just 20%, that's 30%.

Alex: If you draw the line at 30%, as we've discussed, that's fewer than five companies in absolute terms. In the private markets, every company we see and spend time with is growing far faster than 30%.

David: And honestly it's a bit obvious: go talk to IT managers and CIOs — look at their growth and the money they're spending on AI, and the easiest source of that money is to stop spending incremental dollars on new SaaS software projects.

David: We've invested in some SaaS software companies, including in the public markets. Our conversations with their founders are very focused: how do I use my existing distribution — which in many cases is very strong, with locked-in customers who, as you said to Santi earlier, don't churn — to launch genuinely interesting new AI products for them and push revenue growth higher? I think to dispel the biggest fears around the "SaaS apocalypse," We need to see a period in which software companies keep posting numbers like 98% gross retention — which has always been the key to their appeal as an investment; keep driving the efficiency gains we talked about; but most importantly drive revenue growth acceleration. That would show both that they're safe on defense and that their offensive investments actually make the business better.

Sarah: I think the point in that blog you wrote a few months ago about "two paths" is very interesting as a rule of thumb. Maybe you could share a bit more detail on the specific percentages.

David: Yes, the more I write and the more I talk to founders, the more I feel: almost everyone we've talked to says "yes, I'm trying to push revenue growth higher." That's clearly the main path for the companies we invest in and support. This is very different from a year ago — when the fear was "everyone will rewrite their software systems with vibe coding," and that clearly isn't happening in the market. But there's a lot of pressure to deliver revenue growth acceleration. So we said: target 10%-plus acceleration, which is a high number, but with magical products plus the budget available for AI, it seems feasible. We'll see. We have a few companies that are close, and we'll see results over the next 12 to 18 months.

Sarah: Yes, absolutely. As with everything, you can't put every company into the same "software" bucket and talk about multiple compression and slowing growth. In fact, software has already become highly differentiated: look at this chart, cybersecurity and observability really stand out, and vertical software overall is holding up far better than horizontal applications. Our framework for thinking about "what holds up, or even does well, and what is in a longer-term decline" is: really think about how AI changes customer demand for this product. The cybersecurity risks that AI brings are already well publicized; but as more software and agents create new security and monitoring needs, that actually creates greater demand for incumbents in these markets — you can see the CrowdStrike annotation on the right. Conversely, applications are facing varying degrees of workflow change: one of our best CEOs, Ali Ghodsi, likes to talk about "the AI guillotine" and what gets cut first. You can also think about it from the other side: what does AI actually need? Putting these two dynamics together, I think, explains most of what we see in the public markets.

Alex: Yes, I think this is very consistent with what we see in the private markets. As you implied, Sarah, in security, agents are accessing more and more systems and taking more and more actions — the OpenAI and Hugging Face incidents were eye-opening for many people. So security is top of mind. Also, look at vertical AI companies: they're the fastest-growing companies in our private markets — Harvey, Abridge, Elise — growing faster than any predecessor in their respective industries. That reflects customers recognizing that the key isn't just the model — as we discussed earlier — but how you orchestrate it, how you build applications around it, and the workflows relevant to the vertical that require these.

Santi: We spend a lot of time looking at the public markets, not to invest but to price exit multiples, so we track them very closely. If we sum up the year: the first few months were the "SaaS apocalypse." We wrote a blog about what kinds of businesses would perform well. Fast-forward to today, and the software index has actually returned to where it was at the start of the year, but it has really split into perceived AI losers and winners. The public markets sometimes oversimplify things too much. But to your point, AI winners aren't just those who can improve their cost structure, but those who are actually accelerating and capturing this new money; if you don't capture it, you may not be riding this wave.

Sarah: Conversely, we wanted to show some operating data from private companies, to see what customers are actually doing. In Stripe's SaaS customer data there's actually a "counter-narrative" finding: growth accelerated entering 2026, both for younger and more mature businesses.

David: Stripe calls it "RenaSaaS."

Santi: Oh, "Renaissance."

David: I actually quite like the term "RenaSaaS."

Sarah: Yes, I like "RenaSaaS" too.

09 Between Public and Private: $2.4 Trillion of Private Champions and IPOs

Santi: We've talked a bit about the public markets, and Sarah talked about SaaS acceleration in the private markets. We're often asked: will companies stay private longer? Look at the top six companies ranked by latest-round valuation today: Anthropic, OpenAI, Databricks, Stripe, Waymo, Revolut — together about $2.4 trillion. That's more than the total market value of IPOs over the past decade (excluding SpaceX), which cumulatively is about $1.7 trillion. So just the volume of activity in our market — again, six companies, $2.4 trillion — is almost as big as the Russell 2000, a massive public market index where hundreds of public fund managers spend most of their time. So spending time on these late-stage champions is increasingly exciting, because they can keep investing in growth rather than maximizing short-term profit. But David, you can talk about how we advise companies on IPOs, and when they should stay private...

David: Yes, look, on IPOs, especially for founder-led companies, one thing we keep discussing, and that I've written about, is: at this point, the founder is itself the asset class. The reason we make these bets, and part of why it benefits some companies to stay private, is that they can typically make bigger bets in the private markets, and those bets have longer payoff periods. Zuck and Elon are obviously exceptions in the public markets. But you can see how it shows up in the numbers at companies like Databricks and Stripe: huge new bets in new product areas, followed by visible revenue acceleration. You can do that in a public company too, but it comes under much tighter scrutiny — obviously Meta's stock fell below $100 at its lowest point, when there was extreme skepticism about their AR/VR investments. For us, the way we talk about it with founders is: an IPO is just another financing event. What do you get as a public company relative to a private one? Obviously you can reach great scale in the private markets. There are also companies that already operate like public companies: highly focused on efficiency, with basically every metric tracked. And some companies may benefit from publicly disclosing their financials to win customer trust at both the enterprise and consumer level — for example, after Navan went public we saw it accelerate again, partly because some larger enterprises trusted them more because they were a public company.

David: Yes, and of course there are also cases where capital needs may become very large over time; the private capital pool is very deep and can serve many such companies, but at some point the capital need may even become too large for the private markets to hold.

Santi: That's where the public markets come in. Another area where we spend a lot of time in the private markets is the secondary market. There are two dynamics here. First, we're seeing more and more companies run tenders for employees — if they choose to stay private, letting employees get liquidity outside the public markets. One interesting data point here: Carta shows a tender participation rate of only 58% — employees choose not to take liquidity because they have enormous confidence in their business's performance. This is exactly the opposite of the sometimes-told narrative that "employees are cashing out" — what we're seeing is employees choosing not to cash out because they believe so much in their companies.

David: Imagine if that were 100%, what would happen to San Francisco housing prices.

David: That would be tough. Indeed. We obviously often participate in and lead these tender offers. We think they're good things. They serve two purposes. First, for employees and prospective employees, it's sometimes hard to compete with the liquidity appeal of the public markets — RSUs hitting your account on a net-of-tax basis every quarter. So it's a competitive weapon for private companies competing for talent, or retaining employees, against public companies. The other side: We advocate resetting valuations more frequently. That keeps your private share price fresh, for a few reasons: one, it's easier to talk about with employees and prospective employees; two, if you want to do M&A — and many of our companies have — you have fresh currency to use.

Santi: I think secondary market discounts get discussed a lot, because that was the dynamic of the past five years: coming out of 2021, 2022, 2023, 2024, the median discount of the secondary market to the latest round price was meaningful, because some rounds may have been priced too high or gone stale. But what we're actually seeing today is that when there is a secondary market trade, the discount to the latest round price is basically zero. That shows the prices companies raised at are fresh, and there are always new investors willing to buy at the same price — which was not the case in the past year or two.

10 Everything Is AI: Where the Next Opportunity Lies

David: Yes, good point. This interesting slide is called "Everything is an AI computer." In this 2026 snapshot, AI-related companies account for 86% of US VC deal activity, up from 65% in 2025. Clearly, AI has become the dominant destination for venture capital.

Alex: Yes, but under the AI label, the opportunity has actually broadened enormously: enterprise applications, consumer applications, and also services, semiconductors, power.

Alex: And defense. A lot of attention is focused on OpenAI, xAI and Anthropic, but beyond that, our opportunity set has never been as deep or as broad as it is now.

David: Finally, we thought it would be fun to talk about some areas we're very excited about right now. Obviously we touched on some of the consumer work that will be done by agents — you could call it long-running agents, autonomous agents, heavy-use consumer. If you think about it this way: consumer agents can finish all the tasks you don't want to do, giving you things you otherwise wouldn't have spent time or money on — that's very compelling, and I think diffusion will be fast.

Robotics is an area we've spent a lot of time and attention on, and we happen to think it could even be bigger than LLMs — but might arrive three to five years earlier. So over the next five years, this will be a huge area of investment and excitement.

Autonomous driving is already here, driverless is already feasible, and this is a very exciting moment. Step back and look at the auto and transportation industry: it's one of the biggest industries in the world, but we probably don't talk about it much because we spend so much time talking about AI now. If you look at Uber and Lyft's miles, they're only about 1% of total US miles driven. I think with autonomous networks that are 14 times safer than human drivers, that number will expand by at least an order of magnitude in the next few years. Also, the US sells 17 million new cars a year, and I think in the next decade they'll all be autonomous.

David: AI times biology: this is a super exciting area, and people in the labs are obviously talking about it — new drug discovery, solving some of the world's most debilitating diseases. We all hold enormous hope for that promise, and we'll see a lot of progress in the next decade.

David: Personal health is another area I'm personally very excited about — as a health maximalist, I've always felt there's no good place to put all your information about yourself and get highly personalized advice.

David: Finally, diffusion into the enterprise and beyond coding: this is what you, Sarah, just talked about — actual technology diffusion in the enterprise is still so early, and I think this will be super, super exciting.

David: And finally there's one area that isn't in the AI bucket, which we call American Dynamism — the re-arming of the whole American Dynamism tech stack. We've invested in this area for a long time and are big investors. It's still a small slice, less than 5% of total defense spending. Newer suppliers like Anduril, Saronic and Castelion — we expect this to grow substantially as demand changes.

David: There are so many exciting areas. Obviously we're very active in AI, and we're optimistic about the impact it will have on the broader economy — the scale of the buildout is enormous, but we think it will be a massive efficiency gain in the US. It was great recording with you all. Thank you.

Join the official Coincamps community:

X: https://x.com/coincamps

Telegram: https://t.me/coin_camps