AI Stackin’ Benjamins, But Where are the Exits?
Silicon Valley is swimming in AI capital. The numbers are eye-popping: in the first half of 2025 alone, U.S.-based artificial intelligence startups raised $104.3 billion, nearly matching the entire haul from 2024. From OpenAI’s $40 billion raise in March to Meta’s $14.3 billion bet on Scale AI, the cash is gushing like a broken fire hydrant on a hot night in July.
But here’s the paradox that seems to be defining the AI economy: while the money pours in, the exits are barely trickling out.
Venture investors are placing unprecedented bets on AI, but when it comes time to cash out, the scoreboard looks eerily quiet.
Money is chasing vision, but vision isn’t yet returning money.
The Billion-Dollar Firehose
According to PitchBook, two-thirds of all venture capital deployed in the U.S. so far this year has gone into AI startups. That’s up from 49% in 2024. And while that headline-grabbing figure might suggest a sector at the height of its maturity, the reality is more complicated.
Mega-deals are dominating the landscape:
- OpenAI raised $40 billion, led by SoftBank.
- Meta funneled $14.3 billion into Scale AI, acquiring not just tech — but talent.
- Anthropic grabbed $3.5 billion.
- And Safe Superintelligence, the new venture from OpenAI co-founder Ilya Sutskever, debuted with a $2 billion raise.
A Market Full of Entrances, But Few Exits
Despite the capital tsunami, exits haven’t kept pace. In the same six-month window, there were 281 VC-backed exits totaling just $36 billion — about one-third of what was invested. That might sound like healthy activity until you realize that the value of those exits is disproportionately low.
According to Dimitri Zabelin, senior research analyst at PitchBook, most of these deals are bolt-on acquisitions. They’re not acquiring AI companies to spin them out or take them public; they’re acquiring to subsume, to package, and to posture.
The fuel is abundant, but the fire is small.
Slide Insurance, an AI-driven homeowners insurance company, did go public at a $2.3 billion valuation. EvolutionIQ, which helps with insurance claim processing, was bought for $700 million. Solid outcomes, but hardly the home-run exits VCs typically fantasize about.
These aren’t IPOs with champagne bell-ringing ceremonies. They’re M&A footnotes in broader corporate strategies.
Show me the Return?
There’s a mix of economic gravity and strategic calculation at play here. Zabelin attributes the sluggish IPO scene to current liquidity conditions and high interest rates. In today’s climate, it’s simply safer for AI startups to be acquired quietly than to brave public markets.
Startups are built to be bought, not to break new ground.
And most of those bolt-ons aren’t in foundational infrastructure. They’re in vertical AI applications that plug into existing enterprise systems like insurance, HR, or logistics. These are highly fundable because they’re practical. But they rarely generate the kind of outsized returns that turn a $10 million seed into a $1 billion IPO.
Meanwhile, infrastructure companies have much higher potential to scale. That’s where the rare breakout happens.
Take CoreWeave, the GPU cloud infrastructure firm that IPO’d at the end of Q1. Its stock surged 340% in Q2. The company now sits at a staggering $63 billion valuation. It’s the unicorn among unicorns. But also an outlier, not a bellwether.
Exit Strategy: Rethinking Success in AI
The goalposts are shifting in the AI ecosystem. Where once a successful exit meant a blockbuster IPO, today it might mean becoming part of a larger company’s pitch deck ahead of their IPO.
The goal is no longer to go public, but to go useful.
In other words, the new North Star isn’t always independence, it’s integration. An AI startup that makes a legacy firm look more innovative could be more valuable as a story than as a standalone business.
Final Thought
In AI, the runway is long, the stakes are high, and the scoreboard isn’t just about dollars; it’s about direction. As investors hold their breath for liquidity and founders recalibrate their ambitions, the AI market continues to evolve.
Because in this game, the winner isn’t the loudest entrant, it’s the last one standing at the exit.
weishaupt.ai | Beyond Intelligence.™ Talent for the AI Revolution.
Source: AI startups raised $104 billion in first half of year, but exits tell a different story
