Key Takeaways
- • AI and machine-learning companies accounted for roughly 83% of U.S. venture capital deal value through September 2026, according to data shared by Social Capital, up from 65% in 2025 and the highest share in the firm's historical data.
- • Social Capital founder Chamath Palihapitiya estimates that roughly 70% to 80% of AI-company revenue comes from other AI companies that themselves raised venture capital.
- • Palihapitiya questions whether this revenue reflects durable end-market demand or a venture-funded ecosystem in which startups increasingly sell to one another, drawing a comparison to the SaaS boom of the early and mid-2010s.
- • Investors should monitor customer concentration, cash burn, outside-enterprise adoption, and free cash flow to gauge how durable AI revenue is if fundraising slows.
Summary
Artificial intelligence now absorbs an unprecedented share of U.S. venture capital. According to data shared by Social Capital, AI and machine-learning companies accounted for roughly 83% of U.S. venture deal value through September 2026, up sharply from 65% in 2025 and the highest share in the firm's historical data. The trend has been steep: AI represented only about 10% of U.S. venture deal value in 2015, rising to 26% in 2020, 31% in 2021 and 2022, 39% in 2023, and 51% in 2024.
Social Capital founder Chamath Palihapitiya argues that headline growth may obscure the quality of the revenue underpinning the boom. He said he suspects roughly 70% to 80% of AI-company revenue comes from other AI companies that themselves raised money from venture investors. In his view, the central question is not whether circular revenue exists in isolation, but whether it reflects durable end-market demand or a funding ecosystem in which venture-backed companies increasingly sell products and services to one another. He drew a parallel to patterns he says emerged during the software-as-a-service boom of the early and mid-2010s.
The concentration itself is significant. Capital has coalesced around a single theme to a degree the firm's data has not previously shown, and the question is whether part of the AI economy is being sustained by capital moving from investors into startups and then from one startup to another. The article notes that intra-ecosystem revenue is not automatically low quality, since AI infrastructure providers can generate legitimate revenue from other AI companies.
The critical test, as the analysis frames it, is what happens if fundraising slows. Revenue funded by customers' internally generated cash flow would be considerably more durable than revenue that depends on repeated venture rounds. For investors, the metrics to monitor include customer concentration, cash burn, adoption outside the venture-funded ecosystem, free cash flow, and the share of AI revenue that ultimately comes from customers outside that ecosystem.
Social Capital founder Chamath Palihapitiya argues that headline growth may obscure the quality of the revenue underpinning the boom. He said he suspects roughly 70% to 80% of AI-company revenue comes from other AI companies that themselves raised money from venture investors. In his view, the central question is not whether circular revenue exists in isolation, but whether it reflects durable end-market demand or a funding ecosystem in which venture-backed companies increasingly sell products and services to one another. He drew a parallel to patterns he says emerged during the software-as-a-service boom of the early and mid-2010s.
The concentration itself is significant. Capital has coalesced around a single theme to a degree the firm's data has not previously shown, and the question is whether part of the AI economy is being sustained by capital moving from investors into startups and then from one startup to another. The article notes that intra-ecosystem revenue is not automatically low quality, since AI infrastructure providers can generate legitimate revenue from other AI companies.
The critical test, as the analysis frames it, is what happens if fundraising slows. Revenue funded by customers' internally generated cash flow would be considerably more durable than revenue that depends on repeated venture rounds. For investors, the metrics to monitor include customer concentration, cash burn, adoption outside the venture-funded ecosystem, free cash flow, and the share of AI revenue that ultimately comes from customers outside that ecosystem.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.