Jun 9, 2026 · 28m · allin
Bill Maris: How Google Could Crush AI Competitors, Why Small Funds Win, and AI's Atari Stage
Bill Maris · 20m spoken David Friedberg · 2m spoken Jason Calacanis · 1m spoken Chamath Palihapitiya · 1m spoken David Sacks · 45s spoken
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
At the All-In Liquidity Summit 2026, Section 32 founder and former Google Ventures CEO Bill Maris delivers a keynote and panel discussion on venture capital fund dynamics, artificial intelligence, and frontier technology. He argues for disciplined small-fund economics while sharing strategic insights on AI's 'Atari stage', foundation model market realities, and computational biology.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 22% of the talking time here. How this is scored →
The hosts as informed peer 3.6 Guest teaching 4.6 Guest disagreement 2.5 The hosts pushing back 2.5
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 26:45 Exposing perverse GP management fee incentives
Maris forcefully dismantles the large fund paradigm by asserting that a $5B fund returning 1.01x makes the partner more money than a $500M fund returning 3x, calling the ecosystem incentives fundamentally broken.
Hardest push from the hosts ▶ 12:34 Challenging small fund dogma with late-stage compounding dataFriedberg explicitly rejects Maris's premise that small funds are superior by introducing data on high-probability late-stage valuation growth and proposing a barbell fund approach.
Biggest teaching moment ▶ 10:20 The mathematical exit reality of multi-billion fundsMaris walks through explicit exit math showing a $7B fund requires $210B in exits to return 3x, which exceeds total historical venture exit values in almost all years.
The host holds their own ▶ 23:30 Friedberg on global biotech scientist migrationFriedberg displays deep geopolitical expertise regarding China's talent recruitment programs pulling European and Indian scientists, while offering a swift vocabulary correction to the guest.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Lesson #1: Glimpsing the Future in Vermont | 0 | 2 | 0 | 0 | Bill Maris presents a keynote monologue about founding his early internet hosting company in a cold Vermont apartment. Because there is no host participation during this monologue, host expertise and pushback scores are zero. | |
| Lesson #2: Seeing the Future Requires Insanity | 0 | 3 | 1 | 0 | Maris continues his presentation on visionary thinking and founding Google Ventures with machine learning optimization. The segment is entirely monologue with no host involvement. | |
| Lessons #3 & #4: Computer Science and Small Fund Math | 0 | 5 | 2 | 0 | Maris presents data-driven financial math demonstrating why funds under $750M consistently outperform massive multi-billion dollar venture funds. Host scores remain zero due to the monologue format. | |
| Panel Q&A: Climate Corp and Early GV Investments | 6 | 4 | 3 | 5 | David Friedberg challenges Maris's thesis on small fund superiority by citing late-stage compounding data and proposing a barbell fund strategy. Maris counters by questioning whether late-stage asset gathering constitutes true venture practice. | |
| How Google Could Crush AI Competitors on Price | 6 | 5 | 5 | 5 | Maris argues that Google could decimate AI competitors by undercutting token prices by 80 percent and forcefully criticizes private companies staying private to dump overvalued equity onto 401k holders. Calacanis and Palihapitiya press Maris to flesh out market dynamics and bimodal venture return realities. | |
| AI's 'Atari Stage' and Infrastructure Bets | 4 | 6 | 2 | 2 | Maris uses a detailed historical analogy comparing current AI models to 1980s text adventure game Zork to explain why he invests in underlying platforms and infrastructure rather than base models. Palihapitiya interactively synthesizes the lesson into investing in the machinery. | |
| Life Sciences, Longevity, and In Silico Biology | 7 | 5 | 3 | 4 | Friedberg demonstrates strong industry domain knowledge regarding global scientist migration patterns and gently corrects Maris's terminology from neurological reserves to brain trust. Maris explains regulatory bottlenecks in therapeutic clinical trials versus computational biology opportunities. | |
| Deep Tech Tractability and AI Acceleration | 6 | 7 | 4 | 4 | David Sacks argues for writing large late-stage venture checks to avoid early-stage noise. Maris reframes the argument by breaking down how misaligned LP and GP fee structures incentivize massive funds to distort startup valuations. |
Statements from this episode (15)
Google forbade using the term 'AI' for years after GV launched
“At that time, Google would not let us use the term AI. And this persisted for many years.”
Maris estimates Google Ventures achieved 4.1x returns from 2009 to 2018
“We would estimate Google Ventures returns at about 4.1 X”
Bill Maris: All six Section 32 funds are performing in top decile
“And over the course of my time at Section one 32, we've had six funds. We've invested in companies like CrowdStrike and Cohere and Coinbase and all six of those funds have averaged about four hundred million in size, and all are performing in their top decile,…”
Sub-$750M VC funds return 4.76x on average versus 2.42x for megafunds
“Top decile performance of DPI funds smaller than seven hundred and fifty million average return of 4.76 X and funds larger than a billion 2.42 X. Funds below seven hundred and fifty million across that time period represented 95% of top decile performers with …”
A $7B fund returning 3x exceeds total annual venture exit values
“Now, if you have a seven billion dollar fund, and we do the same math through, you know, you've got to return two hundred and ten billion. Seven billion to 70 times three X is two hundred and ten billion, which, ah, exceeds the total venture-backed M&A and IPO…”
Bill Maris: Large fund asset accumulation is not true venture capital
“If you're an RIA and you're, you know, collecting assets, that is not venture. You know, venture, as I practice it at least, is a different craft where you are Making concentrated bets of your time and capital on entrepreneurs and helping them build a business…”
Tech companies claiming public-benefit missions should IPO earlier to share value
“I also have a, an observation that A bit of an objection to companies that wrap themselves up in public benefit language, and then keep the value creation to themselves and an elite group of investors through a big part of the curve, and then say, well, we're …”
Maris: Google slashing Gemini prices by 80% would crush AI competitors
“Well, if you're a company, and you can go to Google and Gemini, and you can pay 80% less for that Basically identical product. Why wouldn't you do that? And then the compression and the pressure on those other businesses goes super critical.”
Section 32 does not plan to invest in large foundation models
“I don't plan on investing in, kind of, larger models, right?”
AI will evolve from Atari to PlayStation 10 in five years
“I think we're at the Atari command line stage of AI, and we're going to get to the, you know, PlayStation 10 stage in the next five years.”
Maris: Finding a drug compound is only 5% of the work
“Because of the human biology and the FDA, if you find a compound, and you think you've got something, that's like five percent of the work.”
Mandatory FDA safety testing means biotech progress will not be exponential
“Well, there's still all kinds of titrating and safety testing that needs to go on, and so I don't think it's gonna go quite as exponential as we would all like it to.”
Friedberg: China is recruiting top scientists from Europe and India
“China's got their own paperclip model now. They're recruiting some of the best scientists from Europe and India, and they're all immigrating to China.”
Maris: Late-stage VC sniping strategy won't work long term
“And so the incentives are broken in all those ways, and the pendulum will swing back, so I don't think just staying late stage and waiting to sniper at larger companies will be a long term. The data would suggest that's not going to work in the long term.”