Aram Verdiyan | The a16z Show Summary

1. Key Themes

The Power Law Has Become More Extreme Than Any Prior Tech Cycle

David George argues that for the first time, capital itself compounds a startup's competitive advantage rather than creating organizational dysfunction. "Right now, clearly the power law is more extreme than it has been in the last 10 to 20 years of technology investing... For the first time in my career, you can take capital and throw it at a company, and it compounds their advantage" [00:00:06]. Historically, over-funding a startup created "coordination issues and overhead issues and dueling priorities" [00:03:07] because you can't hire fast enough — but with AI, "you can throw dollars at compute, and compute can make products and the businesses better" [00:03:34].

AI's Addressable Market Dwarfs Software Because It's Attacking Labor, Not Just IT Spend

Jen Kha reframes the TAM conversation entirely: AI isn't competing with SaaS budgets, it's competing with labor budgets. "Healthcare spends $60 to $100 billion on healthcare IT per year. But AI is hitting on actual labor and the value of tasks that are being performed in healthcare. That's claims, billing, administration. That's a trillion dollar industry. So the TAM of AI can be 10x plus bigger than traditional SaaS or healthcare IT" [00:06:38]. David George adds the macro framing: "if you just look at like how much of dollars are spent in the U.S. economy on labor versus software, it's something like 40 times more" [00:07:27].

Consistency in Venture Returns Is Astonishingly Rare — and Concentration Is the Answer

Jen Kha shares proprietary data that reframes how allocators should think about manager selection: "We've looked at the data of 3,000 venture capital firms in the U.S. Only 20 have achieved consistent 3x net returns over the last two decades" [00:00:00], later clarifying it's "less than 1%" [00:11:46] and that it only requires "three to four 3x net TPI funds over a 20-year period" [00:11:51]. Her conclusion: "if you have an asset class where 20 firms out of 3000 do well, you should concentrate in those 15, 20 firms pretty consistently... when I see a portfolio with 50, 60, 70 venture capital firms, it's very hard for me to imagine that the overall portfolio can generate better than the average" [00:28:51].

"Death of the Middle" — Only Barbell Strategies Win

Both speakers converge on a structural claim about fund strategy: hyper-specialized small funds and full-stack mega-funds both work, but generalist mid-sized firms are being squeezed out. David George explains the founder's revealed preference: "the founder cares about... taking capital from a partner that they think can de-risk the outcome for themselves" [00:15:01], which favors either deep specialists or platforms with "700 employees" [00:15:01] worth of resources. Jen Kha adds a theory on why small pre-seed/seed funds survive alongside giants: they "can carve out a niche for themselves a clip or two earlier than the big firms and actually have a right to win" [00:17:19] before larger firms wait for "more relative certainty" [00:17:19].

Late-Stage Investing Now Has Genuine Venture-Like Return Potential — But Only With an Early-Stage Engine Behind It

Jen Kha notes this is structurally new: "Fund returning math in late stage didn't exist before. It now does" [00:12:16], contingent on concentrated sizing: "your best company should be 5%, 10% plus of your fund. That way, you can actually return the fund on a single company" [00:12:16]. But both agree this capability is downstream of a strong early-stage franchise — "It's really hard to come in as a de novo late stage firm and write a $500 million check" [00:20:22], to which David George replies, "I've lived that world" [00:20:45].

Signal Extraction Is Getting Harder Even As Deal Volume and Speed Increase

Jen Kha describes a market where traditional diligence heuristics break down: "Company comes out of pick your accelerator. I went from zero to five million ARR in a month... There's no renewal cycle yet on that company... for every nine companies like that, there's one really special one" [00:22:01]. David George's answer is that financial analysis alone can't resolve this at this stage — you need product-usage-level texture: "you're not going to be able to do it with financial analysis. You'll have to do it by really understanding the customers, talking to the customers" [00:23:50].

Private Equity and Legacy SaaS Are Being Repriced by AI Resilience, Not Just Growth Rate

Jen Kha lays out a stark valuation collapse: "you look at the pre-chat GPT vintages in private equity. You would have paid... 15 to 20 times EBITDA for a software asset that's growing 10, 20% max. If you look at the public markets today, that asset is trading at two times revenue" [00:32:16]. She also quantifies the new growth premium: "one percentage of growth in the public markets is equivalent to three percentages of EBITDA" [00:32:35], inverting the COVID-era premium on profitability.

The Real AI Bottleneck Has Shifted From Demand to Physical Supply (Energy, Power, Chips)

Jen Kha identifies where the next big venture opportunity sits: "the bottleneck in AI today is not demand. It's on the supply side. So you got energy, the grid data center, then you got chips, then you got frontier models and apps... the U.S. doesn't have a problem with energy generation. It has a problem with speed to power... permissioning transmission... regulatory" [00:46:52]. She notes global competitive pressure: "Other countries are putting out 10x more renewable capacity a year" [00:47:21].

The Next Wave of Massive Outcomes Will Be in Categories That Barely Exist Today

David George explicitly rejects the idea that current AI market cap creation is close to peaking: "robotics is going to be bigger than the language stuff... We are almost nowhere on autonomy... fewer than 10,000 Waymos live in the U.S.... healthcare is 18% of GDP... we've done nothing to scratch the surface either on care delivery or on drug discovery yet" [00:45:37].

2. Contrarian Perspectives

LPs Are Structurally Incentivized to Avoid Career Risk, Not Maximize Returns

Aram Verdiyan makes an explicit incentive-asymmetry argument that most allocators would not say publicly: "a GP can get fired for missing out [on] the next Facebook, the next Uber... that is error of omission and like that is fireable. But LPs on the flip side only get fired if you invest into a [mentor/loser]... you don't get fired for investing in IBM if you're an LP" [00:26:33]. This means, per Jen Kha, "if you miss the frontier models... you kind of were along the benchmark... you're keeping your job" [00:27:24] — a direct claim that LP career incentives actively work against capturing the power law.

Everything Might Work — The Zero-Sum "Which Layer Wins" Framing Is Wrong

Against conventional stack-layer debates (labs vs. infra vs. apps), David George argues the market is simply too large for winner-take-all logic across layers: "I don't know, the market is going to be so big. Like I think everything might work... it's far too limiting to think, oh, if open source does a good job, it's bad for the labs and vice versa" [00:08:20].

A 60% Loss Rate Is a Feature, Not a Bug

David George frames high failure rates as evidence of correct risk-taking rather than poor judgment: "if we're not losing money in a given fund on a given amount of investments, we're not taking enough risk... if you look at our best performing venture funds over time, I think the loss rate is 60% or so" [00:09:38] at early stage.

Coding's AI Success May Be a "Head Fake" for Broader Enterprise Diffusion

Despite being an AI bull, David George offers a real caveat that undercuts the "AI is already transforming everything" narrative: "coding hit, but that's kind of a head fake, right? Coding is perfectly documented... verifiable and... simulatable... most tasks in business do not share those three attributes... maybe the diffusion into other knowledge work beyond coding will take a lot longer" [00:35:11].

Private Equity's "AI Version" Is Not a Shortcut — It Can Make Things Worse

Aram Verdiyan directly challenges the assumption that bolting AI onto legacy PE roll-ups is a safe strategy: "just because you put Sears on a website didn't make it Amazon... turns out if you don't actually build on the workflow, you start to turn customers very quickly if they're used to talking to a human. And for every dollar, every drop in NPS is like a direct correlation with drop in revenue. And then you start to spiral, especially if you have debt laid on top of it" [00:39:46].

3. Companies Identified

SpaceX, OpenAI, Anthropic — Frontier AI labs. Cited as the epicenter of the new power law: "particularly the three frontier model companies... represent somewhere between $3.5 to $5 trillion of potential enterprise value" [00:01:42], and "a lot of our LPs... didn't have a lot of exposure to it" before SpaceX went public [00:01:42].

Cursor — AI coding tool, acquired. Cited as a case study in mispriced signal and market skepticism right up to the end: "even the morning of the acquisition announcement, people were still saying that cursor is dead" [00:22:55]; the deal was reportedly around "$300 million in ARR, $400 million round" territory [00:23:04] before being acquired by SpaceX. Noted as an example of extremely fast path from first financing to acquisition [00:42:41].

Harvey — Legal AI. Used as the clearest example of usage-based signal beating financial analysis: early on "the usage was not very good... it looked like mediocre compared to some other software firms" but "post reasoning models, that totally flipped... absolute takeoff of adoption... every client is actually demanding the law firms use the product" [00:24:21]. Aram also notes personally: "My billable hours have only gone up with the advent of the usage of AI" [00:07:53].

Stripe, Databricks — Cited as generational compounders that LPs were right not to sell early: "Would you have wanted to sell Stripe, Databricks or any of those other companies three, four years ago? Like the answer is unanimously no" [00:42:26]. Stripe specifically noted as an a16z Fund I seed investment held 16+ years before a partial exit [00:43:04].

Intercom — Cited as the rare example of a legacy SaaS company successfully self-disrupting with AI. "What Intercom did is a great example. Bring the founder back, revamp the whole business, create an AI native product, scale it, and then sell. Like, it's almost like you're suiciding your existing business, which in private equity is really hard to do" [00:33:06]. David George: "I just spent a little time with the founder... I just gave him a big high five... You did it, man" [00:33:31].

Airtable — Referenced as an example of a company finding new value through business model transformation, specifically "spin off the hyper agent piece of the business" [00:37:39].

Workday — Referenced in context of Silver Lake's rumored buyout interest as an example of PE "doing more provocative things... to potentially infuse and bring them into the future" [00:31:38].

EA (Electronic Arts) and Medline — Cited as this year's largest traditional PE buyouts (~$50 billion each), used as a contrast point showing venture-stage M&A (Cursor) now exceeds classic PE deal sizes [00:32:07].

4. People Identified

David George (a16z) — General Partner leading a16z's growth business. Identified for his track record and thesis that late-stage venture-like returns are now achievable: Jen Kha notes, "you had that insight in 2019 when you left GA. It's venture-like outcomes in late stage, which is now happening" [00:06:00]. Also credited by Aram for the firm's flywheel model built on 700 employees of operating resources [00:15:01].

Jen Kha (Accolade Partners) — LP/allocator with proprietary data on manager performance across 3,000 VC firms; described by Aram as "the most prolific fundraiser I know" [00:25:51]. Central voice on portfolio construction discipline and the labor-vs-software TAM reframing.

Aram Verdiyan (a16z) — Host/moderator, former a16z employee ten years prior, now works closely with LPs including Accolade. Surfaces the LP incentive-asymmetry insight and the PE/AI "Sears vs. Amazon" analogy.

Elon Musk — Referenced regarding public commentary on "BrockBot on Sam's side... Astra and some of the long-running capabilities that are going to come out soon" [00:00:28] (note: attribution/content here is garbled in the transcript but referenced as a signal of frontier lab roadmap visibility).

"Endowment Eddie" — Referenced via a tweet cited by Aram Verdiyan: "interest in big VC funds has been driven by founders, not LPs. Founders, more often than not, want the brand that can scale, be a lifecycle investor, and help land customers slash hires" [00:13:07], used to substantiate the "death of the middle" thesis.

Gavin (unspecified, referenced by Aram) — Mentioned as having a point of view on the "next $20 trillion market cap company" in a separate conversation with David George, prompting the closing question of the episode [00:44:16].

5. Operating Insights

Judge Product-Market Fit Through Usage Texture, Not Cohort Math, When Financial History Doesn't Exist Yet

David George's explicit diligence framework for pre-traction AI companies: "Everyone can do cohort analysis. Everyone can look at renewal data. But like understanding the texture of the market and what the customers actually want, need and their alternatives... that's how you make the decision" [00:25:17]. This is why he argues early-stage teams — closest to the technology and the product — are indispensable even for growth-stage decision-making: "That's why it's so important to have the early stage business. Because they're the deepest in the technology and the products" [00:25:17].

Concentrated Position Sizing Is What Actually Converts Access Into Returns

Jen Kha's operating rule for allocators (equally applicable to fund GPs sizing positions): "if you can at scale, put five to 10% of your fund in one of the category defining companies... that way, you can actually return the fund on a single company" [00:20:22]. The counter-example is instructive: "I've seen too many times an LP or an allocator find an interesting fund, actually get it right and put 1% of their fund into it... You 10X it, it returns 10% of your fund. It does not move the needle at all" [00:29:19].

Manufacture Early Liquidity Deliberately Rather Than Waiting Passively for IPOs

Aram Verdiyan highlights a specific behavioral marker of top-tier GPs: taking money off the table during frothy markets. "In 2021, when a lot of folks didn't take money off the table, I think that was a good sign of the first indicator. And now in this next cycle, it's can you actually get some early liquidity out through M&A and then let... the winner's IPO over time" [00:43:51].

Watch Per-Employee AI Spend as a Leading Indicator of Real Enterprise Diffusion

David George offers a concrete, replicable benchmark for evaluating how "AI native" a company or industry truly is: "the median company in the U.S. is spending $12 per employee on AI per month. The top 1% of the dataset... is spending $7,000 per employee on AI per month... the most cutting edge banks are probably doing 1% of headcount cost on AI tools" [00:36:37]. This metric can be used to sort real adoption from hype in diligence.

6. Overlooked Insights

The "Song to Each Other in a Cohort" Problem — Accelerator Metrics May Be Structurally Gamed

Buried in a broader point about signal extraction, Jen Kha makes a striking, almost throwaway claim that deserves more attention: some fast-scaling accelerator companies' revenue traction may be inflated by round-tripping demand within their own cohort — "a lot of times they're [selling] to each other in a cohort potentially. And it's not even ARR, but they're multiplying by 12" [00:22:01]. This implies a meaningful fraction of the "hot deal" signal driving up seed/Series A valuations across the ecosystem right now could be artifactual rather than real market demand — a systemic mispricing risk that neither speaker returns to or quantifies further, but which has direct implications for anyone underwriting fast-growing AI seed rounds today.

Growth Rate Has Become a Direct, Quantified Proxy for Risk Pricing in Public Markets — Not Just a Growth Metric

Jen Kha's statistic that "one percentage of growth in the public markets is equivalent to three percentages of EBITDA" [00:32:35], paired with her observation that this is the exact inverse of the 2021 profitability obsession, is understated in the conversation but represents a fully reversed capital markets regime. The implication: growth deceleration is no longer just a valuation multiple problem for software companies, it is being read by public markets as a direct signal of AI obsolescence risk — meaning any company (public or soon-to-be-public) that cannot show acceleration is effectively being priced as a melting ice cube regardless of current profitability, cash flow, or absolute revenue scale. This reframes "growth at all costs" not as a 2021-style vanity metric but as the single cleanest signal of AI-native survival — a much higher-stakes read than either speaker explicitly draws out.