Speaker
Harry Stebbings
Appearances over time
6 episodes
Episodes
6
20VC: Will OpenRouter Sell for $10BN to Stripe? | Why Chinese Open Models Are Beating America—and What Happens Next | Why Enterprises Are More Fearful of Anthropic and OpenAI Than China | Is the Routing Layer Becoming a Commodity with Alex Atallah
20VC: The AI Boom Will Create Enormous Roadkill: Who Wins & Loses | Why Founders Should Never Take Multi-Stage Money at Seed | Why Triple, Triple, Double, Double is Good Enough
20VC: Airtable Sold for $1.285BN | Leo Achenbrenner's Situational Awareness Blows Up | Moonshot AI Raises $3.5B at $35B | Anthropic Model Breaches Three Companies' Security | Big Tech Earnings: Why Palantir Beat The Rest
20VC: Are OpenAI and Anthropic Overvalued? The Open-Source AI Reality | How Token Costs Will Fall 10x And Usage Will Explode 100x | The Future Is Not One AGI; It's Millions of Specialised Models with Lin Qiao, Founder and CEO @ Fireworks
20VC: Wix's Founder on What Wall St Gets Wrong About AI and Wix | Will Base44 Win the Vibe Coding Wars | The Truth About the Economics of Vibe-Coding | The Buyback Disaster: Lessons Learned with Avishai Abrahami
20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean
Podcasts
Quotes & moments
Fireworks AI reached $1 billion in annual recurring revenue with only 200 employees, scaling in roughly 4 years.
Airtable was acquired by Bending Spoons for $1.285B, a steep discount from its 2021 peak valuation of $11B, on $485M in revenue growing 20% YoY.
Frankel agrees that the worst-performing funds of this vintage will be the $50M–$100M seed funds — too big to be a collaborative friend, too small to lead an $8–$10M seed round.
Leo Aschenbrenner's AI hedge fund, which once had $45B in assets using 4x leverage, collapsed in a week, with Citadel buying the public book for a reported $16B.
Token prices have fallen approximately 90% over the last 18 months, raising questions about whether lower prices help or hurt OpenRouter's revenue model.
Suno, the AI music company backed by Founder Collective at seed, is now worth $5 billion.
Base44 has scaled past $150M in ARR at record speed since acquisition, likely more at time of recording.
OpenRouter has raised over $153M in funding and is valued at approximately $1.3–1.5 billion, with reports of a $10B acquisition offer from Stripe.
OpenRouter charges a 5.5% take rate on its pay-as-you-go plan, with a separate enterprise plan based on committed spend with no additional fee.
Chinese AI lab Moonshot AI raised $3.5 billion at a $35 billion valuation, underscoring how Chinese open-weight models are acting as a drag on US closed-source model prices.
Valar Atomics, a 3-year-old small modular reactor company, tripled its valuation to $6B in a Sequoia-led round, driven by explosive demand for AI compute power.
Harry Stebbings noted that on OpenRouter, the top six models by usage were Chinese, with the first US model (Anthropic) ranking seventh.
Fireworks AI raised $1.5 billion at a $17 billion valuation, a remarkable outcome for a 200-person company.
Navan claims the industry average for booking a business trip is 45 minutes, versus 7 minutes on their platform.
UK university curriculum takes 3 years to update, leaving computer science graduates learning pre-ChatGPT material.
Frontier models are becoming a commodity. Arvind Jain says 90%+ of enterprise use cases can already be handled by open-source models, and Glean now uses them to cut customer costs. The real question isn't open vs. closed — it's whether enterprises will accept Chinese models.
While every major CEO is cutting headcount, Arvind Jain is hiring aggressively — from 1,000 to 5,000 people. His argument: if two competitors have the same AI tools, the one with more people can build a 10x better product, not just the same product more cheaply.
When Glean raised its Series C, the company had sub-$5M in revenue but the round valued it above $1 billion. Jain's framing: the extreme valuation wasn't about the numbers — it was a statement to prospective employees that something special was being built.
Glean built an AI triage agent to handle 95% of engineering production alerts, replacing a 15-person on-call team. It worked — but cost $1 million per month, more than the humans it replaced. This is the real AI ROI problem nobody talks about.
As AI agents repeatedly perform business tasks, they accumulate institutional knowledge. If enterprises don't own and control those agents, all of that learning accrues to OpenAI or Anthropic — creating an operational dependency far deeper than any previous technology relationship.
Harry Stebbings argues that AI-driven headcount cuts free up budget for the best frontier models and best engineers, creating faster-shipping smaller teams. Arvind fires back: your competitors have the same AI tools and more people — they'll build something 10x better while you're optimizing for lean. A genuinely unresolved debate.
Microsoft's bundling strategy works — but only in a seat-license world. Once AI moves to consumption-based pricing, enterprises pay per unit of work done, giving them a natural reason to let best-of-breed tools compete. Bundling advantage disappears when you're paying for outcomes, not seats.
On OpenRouter, the top six models by usage are Chinese. The first US model — Anthropic — ranks seventh. Arvind Jain and Harry Stebbings debate whether regulatory barriers or open-source innovation from the US can reverse this trend, or whether it's already too late to close the gap.
Open source vs. closed source is a settled debate for most enterprises — open source wins on cost. The new, unresolved question is whether CIOs will accept Chinese open-source models despite backdoor fears and competitive optics. Early movers who accept them gain a massive cost advantage.
Glean now has essentially 100% of its initial code written by AI — but enforces mandatory human code review before any commit. The paradox: review is now the bottleneck, not writing. Some companies eliminate reviews entirely, but Glean believes maintaining oversight is worth the cost while the industry is still in a learning phase.
Every nation wanted sovereign AI models a year ago. Almost none succeeded. The reason: training a frontier model requires billions of dollars of upfront investment, making it fundamentally incompatible with the skunkworks, volunteer ethos of traditional open-source software development.
The overabundance of venture capital is damaging the startups it's meant to fund. When seed-stage companies pay half-million-dollar engineer salaries and investors shrug, they build structures that are neither sustainable nor competitive with Google — which simply doesn't need to match those salaries.
In 3-5 years, the most valuable employees won't be specialists. They'll be composites: part engineer, part product manager, part designer. Or part salesperson, part solutions engineer, part post-sales. AI makes it possible; competitive pressure makes it necessary.
Most enterprise AI deployments fail not because models are bad but because agents spend most of their time and tokens searching for the right context. Treating AI like a brute-force tool burns money and delivers slow results. The solution is investing in context infrastructure before deploying agents.
Most enterprises connect AI to their systems and let it brute-force its way through context assembly — burning tokens, slowing down, and delivering poor results. The real ROI unlock is investing in context infrastructure so agents start with the right information rather than searching for it.
Analysis
What they talk about
- Business 100%
Connections
Shows they appear on and people they share episodes with. Drag to explore.
