At a glance
WHAT IT’S REALLY ABOUT
a16z’s David George explains explosive AI growth, efficiency, and risks
- a16z’s 2025 dataset shows a sharp re-acceleration in revenue growth, with top AI performers exhibiting extreme YoY expansion (e.g., ~693%) and reaching $100M revenue faster than prior SaaS cohorts.
- The fastest-growing AI companies appear to scale with less sales-and-marketing spend than SaaS peers, implying demand strength and product pull—not go-to-market spend—is the primary growth driver.
- AI company gross margins are often lower due to inference costs, which a16z views as a “badge of honor” when it signals real AI usage and a belief that costs will fall over time.
- ARR per employee (ARR/FTE) is highlighted as a new efficiency lens, with elite AI-native companies reportedly at ~$500K–$1M ARR/FTE versus ~-$400K as a prior SaaS rule of thumb, though much of the advantage may reflect unusually strong demand and post-2021 cost discipline.
- On the supply side, the AI CapEx buildout is massive and increasingly involves debt in some cases, but is still largely funded by highly profitable hyperscalers; meanwhile, private markets exhibit strong power-law concentration with the top unicorns capturing a growing share of total value.
IDEAS WORTH REMEMBERING
5 ideasAI-native demand is the dominant driver of growth—more than spend.
George argues the best AI companies are not outgrowing SaaS by “buying” growth via sales and marketing; they’re growing faster while spending less, suggesting product pull and urgent customer demand.
Lower AI gross margins can be a positive signal when it reflects real usage.
Inference costs can depress gross margin, but a16z interprets that as evidence customers are actually using AI features; they expect inference costs to decline, improving margins over time.
ARR per employee is becoming a core KPI for the AI era, but it’s not pure automation yet.
Top AI companies show ~$500K–$1M ARR/FTE versus ~-$400K historically, yet George cautions this is largely “best-of-best + demand strength” and early post-2021 efficiency, not fully reimagined AI-run organizations across the board.
Pre-AI companies face an “adapt or die” mandate on both product and operations.
The prescription is two-sided: rebuild products as AI-native experiences (not bolted-on chat) and aggressively deploy coding models and AI tools internally to change speed, cost structure, and team design.
Coding is the leading edge of internal AI adoption and may force org redesign.
Anecdotes cite 10–20x faster rebuild cycles using tools like Codex/Cursor, with tool spend high enough to prompt rethinking how product, engineering, and design boundaries work over the next 12 months.
WORDS WORTH SAVING
5 quotesAI demand side is crazy.
— David George
The fastest growing AI companies are reaching 100 million bucks of revenue significantly faster than the fastest growing SaaS companies in their era.
— David George
The best AI companies that are growing the fastest are not the ones spending the most amount of money on sales and marketing, and they're spending less money on sales and marketing than their SaaS counterparts, and yet they're growing much, much faster.
— David George
You need to adapt to the AI era or die.
— David George
I now ask the question, um, for, for every task that we now need to complete, uh, can I do it with electricity or do I need to do it with blood?
— David George
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome