Do investors systematically underestimate how large and how fast AI-era outcomes become?

Adam Mosseri · Jul 24, 2023

Generative AI is arriving much faster than he expected, and he has become much more bullish on it over the past year

Scope: a change of mind within the last twelve months

60:57 20VC: Instagram CEO, Adam Mosseri on Threads: The Journey from 0-100M Users; What Worked, What Didn't and the Plans Ahead | Instagram: Biggest Mistakes, Successes, Misconceptions, TikTok Competition & The Future of Social Media; Interest Graph or Friend G

Harry Stebbings · Jul 22, 2024

Investors always underestimate the eventual size of their biggest winners

Bessemer's memos valued Procore at $300M and Snap at $500M; Shopify went public at $700M and was sold at $2B; nobody would have predicted Airbnb's current scale

28:33 20VC: How I Lost Airbnb at Seed Because of an Exploding Term Sheet | Investing Lessons from Roelof Botha & Peter Thiel | Why VC is Less Collaborative Than Ever and Great Companies Are Being Destroyed by Too Much Cash with Kevin Hartz @ A*

Eiso Kant · Oct 7, 2024

Whether GPT-5 delivers a step function change is the wrong question; in a decade the salient fact will be how much value and abundance we failed to anticipate

It parallels how we look back at the early days of the computer, the internet and Google and see that people didn't internalize the coming unlock

36:02 20VC: Raising $500M To Compete in the Race for AGI | Will Scaling Laws Continue: Is Access to Compute Everything | Will Nvidia Continue To Dominate | The Biggest Bottlenecks in the Race for AGI with Eiso Kant, CTO @ Poolside

Harry Stebbings · Oct 14, 2024

Investors systematically underestimate how big their winners become, so insisting on a $20B outcome up front only costs you great-returning companies like Olo and PillPack

Bessemer's published memos on Snap, Procore and Shopify show that every great company's size was underestimated at entry

54:46 20VC: Investing Lessons from FC Seeding Uber, Airtable and Coupang | Why Pro Rata is the Original Sin in VC | Why Liquidity Has Died in 2024 | Why LPs are Pissed with VCs | The Hard Truth About Seed Fund Economics with David Frankel @ Founder Collective

Harry Stebbings · Mar 10, 2025

Investors and founders consistently underestimate how big a landing wedge can become

56:21 20VC: Lessons from Investing $2BN and Returning $8BN in Cash | Why Most Venture Partnerships are Broken | We Sold Salesforce Early and Lost Out on Billions | Are The Best Deals Always Expensive and Competitive with Jake Saper @ Emergence Capital

Harry Stebbings · Aug 11, 2025

The biggest and most common mistake in venture is underestimating how big your winners will become

50:21 20VC: Figma, Scale, Wiz: Inside Index's Decacorn Factory | Decision-Making, Investment Process, Biggest Lessons, Biggest Misses | Why Gross Margin is a Fallacy at Seed | Never Turn Down a Deal on Price with Martin Mignot, Partner @ Index Ventures

Byron Deeter · Aug 25, 2025

The AI wave will create many trillion-dollar businesses — the scale of outcomes is roughly an order of magnitude larger than previously expected.

He thought he understood how big this next phase would be, but the scale has effectively 'added a zero to everything'.

0:00 20VC: Do Margins Matter in AI? | Is Defensibility Gone For Good? | Is Vertical SaaS Dead in a World of AI | What SaaS Rules Are BS and No Longer Apply in a World of AI | The Future of Venture: Why Chanel vs Walmart is BS with Byron Deeter

Byron Deeter · Aug 25, 2025

He has added a zero to his estimates of the AI opportunity — many trillion-dollar businesses will be created from this wave

Anthropic's actual shared numbers show a path to billions in revenue with compelling unit economics never before seen in the industry, and past memos that treated a billion-dollar outcome as a big success now look embarrassing

75:27 20VC: Do Margins Matter in AI? | Is Defensibility Gone For Good? | Is Vertical SaaS Dead in a World of AI | What SaaS Rules Are BS and No Longer Apply in a World of AI | The Future of Venture: Why Chanel vs Walmart is BS with Byron Deeter

Harry Stebbings · Aug 25, 2025

His dismissal of Thrive's 'if it's a trillion dollar company we'll make money' rationale for the $30B OpenAI round was wrong

At the time it sounded like the weakest investment rationale he'd heard, but the outcome has vindicated it

76:46 20VC: Do Margins Matter in AI? | Is Defensibility Gone For Good? | Is Vertical SaaS Dead in a World of AI | What SaaS Rules Are BS and No Longer Apply in a World of AI | The Future of Venture: Why Chanel vs Walmart is BS with Byron Deeter

Andrew Feldman · Oct 6, 2025

There is a 100% chance he is still underestimating demand, valuations, and the rate of new ideas in AI

He has been repeatedly wrong before — OpenAI's current valuations would have been inconceivable a year or even months ago

9:34 20VC: Cerebras CEO on Why Raise $1BN and Delay the IPO | NVIDIA Showing Signs They Are Worried About Growth | Concentration of Value in Mag7: Will the AI Train Come to a Halt | Can the US Supply the Energy for AI with Andrew Feldman

Harry Stebbings · Oct 13, 2025 · hedged

He has changed his mind about OpenAI and Anthropic's magnitude — he didn't think they could be $3-4 trillion companies and now thinks it's plausible

Scope: 'if you squint'

54:44 20VC: Atlassian CEO on Why Everything is Overvalued & Are We in an AI Bubble | Do Margins Matter & Does Defensibility Exist in an AI World | Is Per Seat Pricing Dead & The Future of Vibe Coding with Mike Cannon-Brookes

Miles Clements · Mar 9, 2026

There is still time to make generational AI investments — he was wrong a year ago to think all the generational AI bets had already been made

The winning companies will be bigger than he envisioned, there is still time to participate in some, and the innovation flywheel is barely getting started

Scope: explicit reversal of his prior view

39:56 20VC: Inside Accel's $4BN Growth Investing Machine | Cursor is Dead is Total BS: Here is Why | What Missing Rippling and ElevenLabs Taught Us | Are $2BN-$10BN IPOs Dead | Why Now is a Great Time to be Thoma Bravo with Miles Clements

Gokul Rajaram · Mar 16, 2026

The biggest investing misses are failures of imagination about terminal value — when these companies work they compound to a scale that is unimaginable at the time

In 2010 he argued with Facebook corp dev over whether terminal market cap was $20B or $40B and it became a trillion-dollar company in about a decade; at Google's IPO PMs thought $30B was already too expensive

Scope: about the rare winners, not typical companies

71:17 20VC: The 8 Moats of Enduring Software Companies: How to Analyse for Durability and Defensibility in a World of AI | Why Dropouts are "AI Maxing" the World & Remote Early-Stage Companies are Dying with Gokul Rajaram

Andrew Feldman · May 26, 2026

The industry's mental frame for scale has shifted extraordinarily fast — multi-gigawatt buildouts that would have sounded delusional five years ago are now treated as routine

Thresholds moved from 20MW to 100MW to a gigawatt to multi-gigawatt in a couple of years, mirroring how users instantly demand more once they experience something good

23:12 20VC: Cerebras CEO on the Future of Data Centres, Token Costs and Memory | We are Not in an Infra Bubble & Dario Got a Bad Deal with Elon for Compute | Should US Companies Sell to China & Why Most Layoffs are AI Washed with Andrew Feldman

Matt Murphy · Jul 27, 2026

He has changed his mind on how big companies can get, and Menlo therefore needs to be bolder and staff free thinkers willing to take those risks

The scale achievable by companies now is larger than he previously believed

Scope: change over the last twelve months

52:56 20VC: Leading Anthropic's First Ever Round | Will Open Source Threaten Anthropic's Business | Do Margins Matter in a World of AI | Why Triple, Triple, Double, Double is Not Good Enough Today | Why Series A is Hard Today with Matt Murphy @ Menlo

Harry Stebbings · Aug 24, 2026

Nobody anticipated outcome sizes like Anthropic and OpenAI reaching trillion-dollar scale as fast as they have

10:46 20VC: Inside Sequoia's Investment Committee: Lessons from Don Valentine, Doug Leone and Alfred Lin | How the SpaceX and Citadel Deals Went Down | What Sequoia Specifically Looks for in Founders with Julien Bek

Avichal Garg · Apr 11, 2022

The correct thing to have done at every moment on the technology curve was to extrapolate exponentially rather than linearly, even though exponential extrapolation feels ludicrous in the moment

Looking backwards, PCs, internet, mobile and cloud were clearly exponential; forward extrapolation yields conclusions that feel silly, so instinct pulls toward linear, but at 1980, 1990, 2005 or 2015 the exponential extrapolation was right

Scope: credits the framing to an article he believes was by Packy McCormick

43:00 20VC: Why Crypto is Software Eating Money, Why Crypto Firms Will Outcompete Traditional Venture Firms, How To Price Tokens and When To Have Them, DAOs: How Are They Structured and What Makes One Successful with Avichal Garg, Co-Founder @ Electric Capital

Tomasz Tunguz · Apr 21, 2023

The pace of AI progress snuck up on everyone because humans reason well about linear change and badly about geometric change; we are on the steep part of the curve and it will eventually flatten into an S

Quality of answers, breadth of knowledge and emergent behaviours all compounded in a way that is difficult to appreciate, which is why it's now hard to stay on top of daily model releases

Scope: expects an eventual S-curve plateau

40:59 20VC: Who Wins in AI; Startup vs Incumbent, Infrastructure vs Application Layer, Bundled vs Unbundled Providers | From 150 LP Meetings to Closing $230M for Fund I; The Fundraising Process, What Worked, What Didn't and Lessons Learned with Tomasz Tunguz

Andrew Feldman · Oct 6, 2025

We are underestimating the size of the AI compute market.

Because demand is the product of three compounding variables, intuition systematically undershoots.

23:47 20VC: Cerebras CEO on Why Raise $1BN and Delay the IPO | NVIDIA Showing Signs They Are Worried About Growth | Concentration of Value in Mag7: Will the AI Train Come to a Halt | Can the US Supply the Energy for AI with Andrew Feldman

Harry Stebbings · Nov 21, 2025

You cannot do this job well while being pessimistic about the future; claims that AI has already saturated its user base echo the same errors made about early internet adoption

Historical newspaper worries about internet user limits proved absurd, and similar limits are being asserted about OpenAI today

73:27 20VC: Max Altman on The New Seed War: Can Anyone Compete with Sequoia and a16z | Leaving $2BN on the Table with Reddit | Lessons from Backing Rippling at $25M Post | Why Climate Tech is a Mirage and Disaster

Brendan Foody · Jun 1, 2026

Investors who called Mercor's 100x-revenue Series B expensive were wrong because they failed to price in that 50% month-over-month growth would persist for another year-plus

The company had already grown 50% month over month for six months and continued for twelve more, reaching ~$400M run rate, making the round an incredible investment in hindsight

Scope: judged in hindsight

30:48 20VC: Mercor CEO on Why Application Layer Companies Have No Defensibility, The Model is the Product | Token Spend Will Exceed Headcount Spend in 5 Years | The True Cost of Hiring AI Researchers in the Valley Today with Brendan Foody

Julien Bek · Aug 24, 2026

Investors should revisit and update their priors on companies they passed on, because the human brain reasons linearly and systematically underestimates early-stage companies riding an exponential

AI is transformative and we're only on the foothill of the exponential; three years in, the exponential is visibly playing out and early valuations look cheap in hindsight

10:14 20VC: Inside Sequoia's Investment Committee: Lessons from Don Valentine, Doug Leone and Alfred Lin | How the SpaceX and Citadel Deals Went Down | What Sequoia Specifically Looks for in Founders with Julien Bek

Byron Deeter · Aug 25, 2025

Fortune favors the bold in this business and there will be many more trillion-dollar-scale outcomes ahead

77:09 20VC: Do Margins Matter in AI? | Is Defensibility Gone For Good? | Is Vertical SaaS Dead in a World of AI | What SaaS Rules Are BS and No Longer Apply in a World of AI | The Future of Venture: Why Chanel vs Walmart is BS with Byron Deeter

Andrew Feldman · May 26, 2026

OpenAI's key advantage was Sam Altman's willingness to believe exponential demand data and contract years ahead for power, data centres and hardware, while others couldn't believe the numbers

Believing your own data in an exponential growth environment two or three years out is a superpower; competitors saw the same curve but couldn't accept it

7:46 20VC: Cerebras CEO on the Future of Data Centres, Token Costs and Memory | We are Not in an Infra Bubble & Dario Got a Bad Deal with Elon for Compute | Should US Companies Sell to China & Why Most Layoffs are AI Washed with Andrew Feldman

Andrew Feldman · May 26, 2026

Sam Altman and Elon Musk are uniquely capable of reasoning about 100- to 500-gigawatt scale where everyone else's thinking shuts down, and that unconstrained thinking is an extraordinary competitive power

They are not constrained by the way the world has always been

24:37 20VC: Cerebras CEO on the Future of Data Centres, Token Costs and Memory | We are Not in an Infra Bubble & Dario Got a Bad Deal with Elon for Compute | Should US Companies Sell to China & Why Most Layoffs are AI Washed with Andrew Feldman

Anish Acharya · Feb 9, 2026

Investors consistently underestimate how big markets are and overestimate how easy it is to go from zero to one

When you squint you can look at something that isn't working and convince yourself you see how it will work; but once a company genuinely gets something working it tends to compound into greater versions of itself

36:07 Markets are underestimated while zero to one difficulty is underrated

David George · Dec 15, 2025 · hedged

This AI wave will likely produce companies much larger than the previous software generation, and it will not require AI to consume all human labor — most of the value shows up as surplus delivered to end customers

He never imagined the big software companies could get as large as they did, so the next wave should scale further; in practice massive surplus flows to customers rather than pure labor replacement

25:00 Bigger outcomes driven by customer surplus not labor capture

Lucas Swisher · Feb 23, 2026

The outcomes in this generation of technology will be far bigger than in the last generation, because AI will genuinely address labor — human inputs becoming machine inputs — rather than just being assistants.

He became convinced by using agentic tools like Claude Code himself and feeling the shift from an assistant world to an agent world.

54:48 Labor replacement makes this generations outcomes far larger than last

Avichal Garg · Apr 11, 2022

Traditional LPs should do the work to understand crypto deeply before deploying, and cost-average in over a couple of cycles rather than FOMO-ing into new crypto managers

the space is still very early with enormous runway and a lot of strategy nuance, so there is nothing to miss out on by being deliberate

48:17 Human imagination about market size not selectivity is the real constraint so allocate broadly to tech

Harry Stebbings · May 23, 2025

AI company valuations that look expensive are cheap by the time rounds close, because revenue grows so fast during the process

Lovable trebled revenues while the round was completing, making it effectively a 10x revenue multiple

57:52 Fast in round revenue growth makes seemingly expensive prices retroactively cheap