The Two Venture Markets

Here's a paradox worth sitting with: 2025 was simultaneously venture capital's second-best year on record and its most anemic fundraising environment in a decade.

Global VC deal value hit $512 billion last year, nearly matching 2022's all-time high. But VC fundraising collapsed to $66 billion, down 35% from 2024 and 70% from the 2022 peak. Only 537 new funds closed, the lowest in ten years.

How do you get record deal-making and cratering fundraising at the same time? The answer reveals something fundamental about what's happening to venture capital.

Artificial intelligence accounted for more than half of all deal value and nearly one-third of completed deals globally. In the U.S., AI captured nearly two-thirds of total capital invested. OpenAI and Anthropic alone absorbed 14% of global venture investment. OpenAI's $40 billion raise, Databricks' $4 billion round, Cursor's $2.3 billion: these aren't venture deals in any traditional sense. They're capital formation events that happen to involve equity.

Meanwhile, funding to non-AI startups slipped nearly 10%. The bifurcation is stark: if you're building AI, capital is abundant. If you're not, you're competing for scraps.

While the broader market struggled, Andreessen Horowitz just closed $15 billion in new capital, representing over 18% of all venture dollars allocated in the U.S. last year. That brings a16z to $90 billion in assets under management, neck-and-neck with Sequoia as the largest venture firm on the planet.

The fundraising "collapse" isn't hitting everyone equally. It's a flight to quality. LPs are concentrating capital in firms with track records, platform capabilities, and the scale to compete for the deals that matter. Emerging managers are getting squeezed; established platforms are getting bigger.

Yes, the IPO drought has created real liquidity challenges. Without exits, LPs can't recommit to new funds. Traditional institutional anchors are in "repair mode." But exit value nearly doubled in 2025, climbing to $549 billion globally. Public listings accounted for half of that. The IPO window is cracking open, not a flood, but enough to relieve pressure.

The venture industry is building new liquidity pathways. Secondaries are becoming mainstream, no longer a distress signal but a legitimate tool. M&A activity is accelerating. The ecosystem is adapting.

The biggest AI rounds bypassed traditional VCs entirely, flowing from sovereign wealth funds, family offices, and hedge funds.

But the smart venture firms saw this coming. a16z's LP base reportedly includes Sanabil Investments, the venture arm of Saudi Arabia's Public Investment Fund. They're not competing with sovereign wealth. They're partnering with it. The firms that figured out how to tap these capital pools are raising mega-funds. The ones that didn't are watching their LPs head elsewhere.

The venture market split in two, but that split creates opportunity. For founders building in AI, capital remains abundant. The challenge is differentiation, not fundraising. For founders outside AI, the bar is higher but the competition for attention is lower. The non-AI founders who raised in 2025 had one thing in common: margins that didn't require a story.

For investors, concentration in top-tier assets will persist. The 2026 playbook is simple: pick winners, stay liquid, don't overthink it.

The uncomfortable question a year ago was whether venture capital could survive the AI funding surge. The answer from 2025: the asset class is transforming, not dying. The firms that adapted are bigger than ever. The ones that didn't are facing a clearing event.

2025 was a record year for AI funding and a brutal year for undifferentiated managers. 2026 will accelerate the same dynamics, which means the window for emerging managers to break through is closing fast.