Oct 27, 2025
Is AI Investing a House of Cards?
Is AI investing a house of cards? ♠️ ♦️ ♣️ ♥️
$100B has been deployed into 8 private companies…. over the past 12 months.
To put that in perspective, VCs have raised $500B in total… over the past 5 years.
Funding the AI ecosystem has quickly become the biggest collective bet in VC history, and it’s so far the right one for those who got in early. 3-year price performance according to Caplight MarketPrice:
📈 OpenAI: +21x in value.
📈 Anthropic: +44x in value.
📈 xAI: +260x in value.
Despite the capital inflow, is it too early to call forever winners? Companies are being valued for flawless execution; any slips could unravel a house of cards. Here, we present the *for* and *against* cases:
💣 Yes, it's a house of cards 💣
⭕ AI Circular economy:
-Heavy interdependence between OpenAI and major tech suppliers (NVIDIA, Oracle, AMD, etc.) ties much of the AI economy to a single platform.
- Compute commitments (~$650B spend vs ~$13B of revenue), amplify systemic risk if growth slows.
- Much of the chip, data center, and energy spend comes from model developers funded by the same suppliers, blurring the line between real demand and capital recycling.
⚖️ GPs may be optimizing for the short term:
- VCs are raising massive SPVs with healthy upfront fees, effectively syndicating some of the long-term risk.
🔎 VC capital is ultra-concentrated:
- 25 VC firms control 40% of total capital raised in the past 5 years, and much of it is flowing into AI mega rounds.
- Bursting of an AI ‘bubble’ would send a shock through the entire VC ecosystem.
💪 No, this one’s built different 💪
📈 Revenue is going up, fast.
- We’re seeing record user adoption, with nearly all top AI companies growing ARR triple digits. From 2024 to 2025 estimates:
----> Anthropic +900%
----> OpenAI +218%
----> Anysphere (Cursor) +2,032%.
📉 Multiples are going down.
- Example: Anthropic trades at 44x 2025 revenue, down from 72x last year, a trend seen across top AI companies.
- One can argue these multiples look attractive vs. Dot Com.
🏦 Model companies have been deemed “too big to fail”.
- Solidifying differentiated sources of capital beyond VC gives deemed winners enough runway to get to AGI.
What do you think? Has the hype of AI diminished VC diligence, or are we entering the ‘age of abundance’ where old risks don’t translate?
Based on Caplight's data.
DISCLAIMER: Past performance is not indicative of future returns. Investments in private companies are illiquid, as there is a limited secondary market for private company shares.
Sources: