Andreessen Horowitz raised $15 billion in 2025. That was a single fundraising cycle, spread across six funds, representing over 18% of all venture capital dollars raised in the U.S. that year. A few weeks ago, Thrive Capital closed its tenth fund at $10 billion, double its predecessor. General Catalyst is reportedly in talks to raise $10 billion of its own. Lightspeed pulled in $9 billion. Founders Fund is circling $6 billion.
Something has clearly shifted. The venture capital industry is consolidating at a speed and scale that has real implications for LPs.
What counts as “mega”?
There's no universally agreed-upon threshold, but general consensus is anything at or above $1 billion. That's the line we'll use here, while acknowledging it's somewhat arbitrary. A $1.2 billion fund and a $15 billion fund are both "mega," but they're playing very different games.
For context, total U.S. venture fundraising in 2025 came in around $66 billion (down 35% year over year and the weakest year since 2017), and the top ten funds alone accounted for 43% of that total.
Capital's concentrating at the top.
Who are the mega-funds?
The names you'd expect. a16z, Thrive, General Catalyst, Lightspeed, Founders Fund, and Sequoia all sit comfortably in this category. These are "pure-play" venture firms (though at this scale, the word "pure" does a lot of heavy lifting) that have expanded from early-stage investing into multi-stage platforms covering seed through growth.
Then there's a related but distinct category: crossover funds. Firms like Tiger Global, Coatue Management, and Dragoneer straddle private and public markets. They often come from a hedge fund background and bring a different investing style: data-driven, high-velocity, less focused on board seats and hands-on company building. These firms are worth understanding because they compete for the same late-stage deals as the mega-fund VCs, and they often set pricing in those rounds.
The thesis behind going multi-stage
The logic is straightforward: if you back a company at Series A and it becomes a breakout winner, why let someone else lead the Series B, C, and D? By running a multi-stage platform, a firm can lead or participate in follow-on rounds, maintain (or even increase) its ownership stake, and capture more of the value creation over a company's lifecycle.
That sounds elegant. In practice, it introduces real tensions.
Fund economics. A $10 billion fund charging a standard 2% management fee generates $200 million per year in fee revenue before a single investment pays off. That's a significant business in its own right, and critics argue it creates incentives to raise larger funds regardless of whether the opportunity set supports it. The counterargument is that these firms need the capital to lead massive rounds. When Databricks raises at a $134 billion valuation or OpenAI raises at $840 billion, writing a meaningful check requires a massive fund.
Deployment pressure. A $10 billion growth fund typically has a 3-5 year investment period. That means deploying $2-3 billion per year. At that pace, you need large deals, and lots of them. The risk is that deployment pressure leads to less discipline on valuation, especially during periods of hype (sound familiar?).
Portfolio construction. Traditional early-stage venture is a power-law business. You make 30 bets, most fail, and one or two return the fund. Mega-fund growth investing is a different animal. The hit rate needs to be higher, the multiples are lower, and the underwriting looks more like growth equity than classic venture.
Not everyone’s going mega, though...
While there’s definitely a trend towards raising larger and larger funds, some VCs have maintained them relatively flat. Benchmark and Union Square Ventures are good examples.
Both funds have generally stuck to modest (using modest very loosely here) fund sizes when compared with their now larger peers.
Benchmark’s latest fund size was flat from the one prior at $425 million, and USV’s sat around $275 million.
What this means for LPs
Here's where it gets interesting.
For large institutional LPs (endowments, sovereign wealth funds, large pensions), mega-funds solve a real operational problem. If you're managing a $5 billion alternatives allocation and need to commit $200 million per year to venture, writing four checks of $50 million to mega-funds is a lot easier than writing forty checks of $5 million to emerging managers. Fewer GP relationships to manage, fewer side letters to negotiate, fewer quarterly reports to review.
That logic has held for a long time, and it's a big reason capital has concentrated the way it has. LPs aren't irrational. Managing a diversified portfolio of 40+ fund relationships is genuinely painful when your team is five people and every GP sends data in a different format.
But here's the thing: that constraint is increasingly a technology problem, not a structural one. The reason LPs default to fewer, larger commitments isn't always because they believe mega-funds will generate better returns. Often, it's because they simply can't handle the operational overhead of tracking more managers. The data extraction, the normalization, the cross-fund comparisons, the reporting. All of it is manual, slow, and expensive.
That's changing. The tools available to LPs today are fundamentally different from what existed five years ago. An LP team that once spent a day per week extracting and normalizing GP data can now automate most of that workflow. When you remove the operational bottleneck, you open the door to a broader, more diversified portfolio: one that can include both a mega-fund allocation for large-cap exposure and a set of smaller, specialized managers where the return potential may actually be higher.
The punchline isn't that mega-funds are bad. They serve a purpose, and many of them are run by genuinely talented investors. The punchline is that "I don't have the bandwidth to manage more relationships" is becoming a weaker and weaker reason to concentrate your portfolio. And for LPs who care about returns, that's worth paying attention to.