Data: A Handful Of Big Startups Are Sucking Up A Third Of All VC Dollars
New quarterly funding data from Crunchbase shows that just 11 mostly-AI companies raising $1 billion-plus rounds account for 1/3 of all startup funding so far this year. Here's what VCs say about it.
The Upshot
A recurring theme in early Upstarts reporting, and at our events, has centered around a sense of bifurcation in the startup ecosystem right now.
There are the red hot AI startups that can raise $20 million before they’ve announced their existence, and the survivors that pivot their way to success. The ones getting acquired off the board for relatively modest sums, and the ones embracing an arms race with few winners. There are VC mega-funds that look more like private equity shops, and the emerging managers who fight and scrap their way to a $22 million fund.
Most recently, we wrote about the competition for elite researcher talent at AI startups: what talent and hiring experts make of the market, and how OpenAI’s pay bands can promise millions to PhD caliber hires, data Upstarts exclusively revealed.
So when Upstarts looked at startup data provider Crunchbase’s info on Q2 venture capital investment across startups, we expected to see, as some more literary tech leaders have described it in a nod to Dickens, another “tale of two cities”.
At face value, the topline numbers show signs of cautious optimism. Overall funding reached $91 billion, per the Crunchbase report, up from $82 billion a year ago. Seed-stage deals showed modest growth, and while early-stage rounds were down year-over-year, later-stage funding soared to $55 billion, up 53% from 2024. Not bad!
But the stat that really stood out to Upstarts – and made this data dump the subject of today’s edition, when we promise to skip most quarterly macro updates in the future – is this: Just 11 companies raising $1 billion-plus accounted for more than one-third of all venture dollars deployed in the first half of 2025. In Q2, 16 companies raising $500 million-plus accounted for nearly one-third of the quarter’s deployed capital.
Most of those companies are, as you probably guessed, operating in AI. Whether they qualify as startups still is more debatable. Q1’s data is dominated by OpenAI raising $40 billion and Databricks $10 billion. More recently, Meta pouring $14.3 billion into Scale AI led the quarter’s deals, though many industry insiders consider the transaction more a pseudo-sale than an investment in a traditional sense.
A strong quarter for seed-stage deals, meanwhile, is skewed by Thinking Machines Lab, the unlaunched new startup from former OpenAI executive Mira Murati, which accounts for 20% of all ‘seed’ dollars announced. Remove it, as Crunchbase’s report admitted one reasonably might, and the quarter would be tracking below a year before.
How does the data compare to what investors are seeing among their portfolios and in their pitch meetings? Upstarts pinged a few of our subscriber VC leaders for feedback; their tips and takeaways — positive, negative, and scornful — are below.
Presented by Notion.
Perfect meeting notes. Every time. Focus on the conversation while Notion AI captures everything. Get instant summaries, action items, and decisions—all searchable in your workspace. Never lose another insight.
A ‘flight to consensus’
That’s how investor Ethan Kurzweil, a co-founder and managing partner at VC firm Chemistry, reacts to the data. Firms are anointing select startups as the consensus picks, then pushing them into bigger early-stage funding rounds to get their allocation of shares, he says. And while VCs might pride themselves on contrarian viewpoints, what they might consider to be such isn’t so differentiated anymore.
“The pace feels more frenetic than ever, because there are still a lot of interesting companies out there,” Kurzweil says. “But the dispersion is so much lower, driven by the market converging on certain things.”
Those “things,” by the dollars deployed, include businesses led by OpenAI veterans, especially senior ones; big infrastructure plays that skip straight to large-sized financing that would have historically counted as a growth-stage deal; or AI application businesses demonstrating rapid early adoption, or a claim to win in a specific vertical, multiple investors say.
In big-spending AI companies, multi-billion-dollar venture capital firms needing to make outsized bets in order to generate fund-moving returns – or at least have the ability to call more capital – have found a perfect match, says one prominent investor and Founding Upstart who asked to remain anonymous: “It’s a perfect storm for investing.”
Founding Upstart Krishna Gupta, CEO of REMUS Capital and chairman of voice AI business Presto, puts it more starkly: “It’s a massive bubble where a few will reap all the gains, but unlikely the few that everyone thinks it will be.”
Illusion, or abundance
For early-stage founders who plan to raise funding later this year, investors share mixed signals. At the top, big headline numbers are “totally misleading,” says Jenny Fielding, co-founder and managing partner of Everywhere Ventures. “There’s an illusion of abundance. Founders going to raise a seed round feel the crunch right now.”
Still, Fielding says plenty of funds are looking to back the more drawn-out, non-OpenAI alumni type startups – they’re just increasingly emerging funds or smaller ones, she argues.
And investor Eric Bahn, co-founder and general partner of Hustle Fund, notes that some of the money going to early-stage startups is simply not getting announced and reflected in the data. “There’s actually money going to the tail-end of this market that is being underreported,” he claims.
He has a point: In its report, Crunchbase notes that some deals, especially at earlier stages, are typically delayed from appearing in its database; such companies are more likely to want to wait to announce a funding round for strategic reasons, or to peg it to a later product release, meaning noisier data than the hundred million-plus rounds that tend to get more formally announced, or scooped by press.
Competitive cash
For Aaron Jacobson, a partner at NEA, tech’s biggest funding rounds reflect an efficient solution of a market demand for an equivalent to an initial public offering. The sort-of startup unicorns solve for liquidity, get to remain private, and as seen with Databricks, raise war chests to make acquisitions without worrying about analysts or earnings results.
For founders in the busiest categories like AI, defense and cybersecurity, that means what might have been a $10 million Series A can now swell to $20 million if desired, Jacobson says; it also means more companies with money to spend as buyers.
On the flip side, any startups that find themselves competing for customer acquisition with the mega round companies might find themselves outgunned. VCs who sense a startup might brush up against one of those bigger fish might be less likely to invest, as well. “I think it’s a much more competitive environment,” he says.
But Bloomberg Beta leader Roy Bahat scoffs at founders reading much into the macro environment. Ultimately, they’re running their own race — and raising funds when they need to.
He responds to Upstarts with a rhetorical question: “Does a soccer player decide not to play based on the weather?”
Just a note this is the sort of article I want to read more of on Upstarts. The Big picture, the macro trends in startups and VC.
The "flight to consensus" isn't just VC groupthink. There's perceived safety in it. No one loses their job for backing the same company everyone else is investing in (unless it's FTX, ofc...)