20VC: General Catalyst CEO Hemant Taneja on The Future of Venture Capital: Chanel vs Walmart | Lessons Scaling GC to $40BN in AUM | Investing $5BN+ Into Stripe Over 14 Rounds | Investing Hundreds of Millions into Anthropic at $60BN Valuation

Welcome to 20VC, I’m Harry Stebbings, and today I’m energized to sit down with Hemant Taneja, the CEO of General Catalyst, the firm he’s grown to more than forty billion in assets and a portfolio that includes Stripe, Snap, Samsara, Grammarly, Canva, and a landmark Livongo deal.

It’s been seven years since our last chat and this one’s in person, so let’s dive straight in.

Do you see yourself first as a venture capitalist or as a CEO?

Both, deliberately. GC is a business, but our core is early-stage venture, so I carry the CEO role while working as a partner, because that duality is what it takes to build an enduring firm.

Can a platform your size still earn the right to do seed with the same intensity?

Yes, if you optimize for ownership and deep founder relationships rather than check size, and we’ve doubled down on seed by adding specialist teams so that early work remains our center of gravity.

Has venture slid from boutique craft to a commoditized, low-margin game?

It does if you only innovate in fund size, sector, or geography; it doesn’t if you reinvent the value you deliver to founders, because scale without a better proposition just drags performance.

So can you grow and still perform for LPs?

We believe classic venture funds shouldn’t get bigger; keep them at sizes that can deliver elite outcomes and create separate products for M&A, go-to-market, or rollups so founders get more tools without diluting venture returns.

Why not just deploy huge checks late and ride the giants?

Our best outcomes come from incubation and seed, then resolutely supporting winners through every round, as we’ve done with Stripe and Anduril, because compounding with conviction is where the real multiple lives.

Do you map must-own categories and force an entry?

At seed we back talent over theses, but we pair that humility with a macro lens we call global resilience, which is why we’ve supported defense leaders across the US, Europe, and India.

What macro shift do people underestimate most right now?

Jobs, specifically how AI changes white-collar work; real transformation needs clean data, company-tuned models, a redesigned org where humans and agents manage each other, and CEOs with the courage to drive it.

Offshored tasks will be re-shored to AI, as we’re seeing in our service rollups, which means reskilling at national scale, and the impact looks like a five-year arc, not a twelve-month blip.

Are governments ready for that labor shock?

Not yet; unless countries capture productivity onshore, service economies risk getting hollowed out, and while Singapore shows impressive planning, most places still treat this as tomorrow’s problem.

How do you square this with inequality and capitalism’s rough edges?

Capitalism is a privilege, so we should build for abundance, not concentration; create rules that keep the platform layer open enough for a broad startup ecosystem rather than a handful of controllers, especially in areas like healthcare and education.

How well positioned is the US against a turbulent global backdrop?

Exceptionally strong on market size, energy, AI, and entrepreneurship, though rising trade friction could make it harder for US-born winners to become global leaders.

If you back regional leaders in the same category, do you accept overlap or seek alliances?

We aim to win in each home market and then compete globally, and I’m intrigued by new partnership models among regional champions that could expand everyone’s share.

Is the AI race a US–China zero-sum game, and who’s ahead?

We’re in a bipolar world with the US, Europe, and India aligned on values, China comparable on capability, and adoption driven by which models are better, which is why capturing compute productivity locally matters.

Does being a second mover actually help in this cycle?

Starting on stronger models can outweigh go-to-market head starts, as long as you avoid technical debt and re-architect quickly when the stack leaps forward.

Walk me through your Anthropic bet and the thesis for entry.

We first invested at approximately sixty billion when coding became a breakout enterprise use case, and the revenue traction and pricing relative to scale made the risk-adjusted return compelling.

If execution stays sharp, we’d keep leaning in, because the developer payroll and broader knowledge work market is enormous, and margins in coding agents are attractive against engineering salaries.

How do you view Microsoft’s dance with OpenAI and now broader model access?

That deal was a masterstroke for Microsoft’s AI halo and Azure pull, and now they’re adding optionality with other models, which is rational once ambitions collide.

Do you worry about Sam’s capital scale talk and mega-infrastructure plans?

Never bet against Sam, but Anthropic’s focus and compute discipline are a feature, and I expect its enterprise business to be one of the biggest.

Growth eventually slows; then unit economics rule. Are margins real here?

Even at more modest growth, the scale supports substantial value, and coding agents have strong pricing power and healthy margins when compared to engineering labor.

With so many model players, does this consolidate to a few?

Expect a handful of global platforms plus some sovereign models; not everyone survives, but the addressable labor pool is so vast that a few winners can be huge.

Is sovereignty enough to make Mistral a winner in Europe?

Arthur has grown from scientist to CEO, they’ve caught up despite earlier compute constraints, and open source with enterprise credibility is resonating, particularly in Europe.

Name a success story where sovereignty was the decisive driver.

US defense primes are the obvious example, and AI is now strategic at a similar level.

Do classic growth playbooks still apply, or is the cadence different now?

The old triple-triple-double cadence is over; today’s category leaders jump in larger steps, and durability is the real unknown.

What do you do with steady, slower-growth SaaS that the market ignores?

We built a Customer Value fund to help those durable, often profitable businesses scale sales and marketing so they can endure outside the hype cycle.

This era feels like peak uncertainty; how do you navigate it?

You lean on values and a true north; ours is building deep relationships, backing enduring companies, and transforming industries with access to talent, policy, distribution, and differentiated capital.

In markets this big, does price even matter?

Price stings once but missing a generational company hurts forever, and too many investors hide behind price because they lack conviction about what the company can become.

Where do you most regret not doubling down, and how concentrated should you be?

The biggest mistakes are under-ownership in winners; concentration is how you drive returns, size it properly within a fund, then cross funds if conviction stays high.

How do you decide when to distribute stock post-IPO?

We ask if our continued time compounds value and pace distributions so we don’t flood the market, while recognizing many LPs sell programmatically.

Private markets are stretching; how do you think about secondaries and liquidity?

The very best privates already behave like publics with secondary liquidity, credit access, and usable stock, while mid-tier companies will still need the public markets to scale, and smaller steady growers need new capital solutions.

Does the longer private window worsen wealth concentration, and should retail get access?

Retail access is coming and it should, but carefully staged into the highest-quality opportunities rather than the riskiest parts of the stack.

Do fee structures need to change, and how do you keep culture performance-first?

We reinvest fees into the business and orient economics around carry; people who want big salaries over performance aren’t a fit for GC.

What product do you still need to add to serve founders in the AI era?

Energy, because AI at scale depends on it, and there’s a path to expand supply while moving toward sustainability, but it requires real infrastructure thinking.

How are you evolving your LP base across strategies and geographies?

We’ve broadened from endowments to pensions, states, sovereigns, and soon retail, aligning each pool with the right part of our platform so founders get flexible, strategic capital.

What have you changed your mind on in the past year, and what’s been hardest about evolving as a leader?

I’m more open to indexing a clear macro wave when picking is impossible, and shifting from practitioner to teacher is still my biggest leadership challenge.

What would be your single biggest piece of advice to an LP navigating venture today?

For founders, the ground is shifting; choose venture partners who build with you and push the model forward.

What worries you most right now?

Keeping quick wins aligned with long-term, broadly shared prosperity.

What would you do if fear was not a factor?

I do not lead from fear; we are already taking big swings and innovating across the board.

Does money make you happy?

No; it is a byproduct of the impact I am trying to make.

Top parenting advice?

Nurture their individuality, and in an AI world teach them to ask better questions rather than chase rote answers.

On college, my sixteen year old is definitely going; for my eleven year old, I said the path might evolve as skills training changes, which made him happy.

Final one: what are you most excited about?

Tech is neutral. Over the next two decades we expect to deploy several hundred billion dollars to help guide how AI serves society, and I want to look back and feel we did right by people.

Conversations like this remind me why I love investing; thanks for making it such a joy.

Thanks for having me; this was a lot of fun.

Watch the episode on YouTube; search for 20VC. Share feedback at Harry at 20VC dot com, and big thanks to the team that makes this happen.