When I was a founder, a $1B outcome was pretty crazy. Salesforce was worth $2B when I started as a B2B founder, and when the first wave of B2B software IPOs went out after that (Box, HubSpot, Zendesk), they all were worth around $1B, plus or minus.
Boy that sounds quaint today.
The goal for many large VC firms today? $100B+ exits. With a shot at $1 Trillion.
What VCs Are Hunting Today: $100B in 10 Years
When I was a founder, a $1B outcome was pretty crazy. Salesforce was worth $2B when I started as a B2B founder, and when the first wave of B2B software IPOs went out after that (Box, HubSpot, Zendesk), they all were worth around $1B, plus or minus.
Boy that sounds quaint today.
The goal for many large VC firms today? $100B+ exits. With a shot at $1 Trillion.
That’s not hype. The target a lot of large funds are now underwriting is roughly $100B in roughly ten years, and the biggest winners are blowing past it by 5x to 10x. One outlier is a fluke. Six is a category. And VCs hunt categories.
Here’s the roster.
Each of these proves something slightly different, and that’s what makes the pattern real instead of a single lucky data point.
- Anduril proves it works outside software. An $88M seed in 2017 to $61B in 2026. That’s 693x in nine years, in defense and hardware, with $2.2B in revenue that doubled last year. For most of venture history, “capital-intensive” and “category-defining outcome” were treated as opposites. Anduril killed that assumption.
- Cursor proves the velocity case. Founded in 2022, $8M seed in 2023, and SpaceX exercised its option to buy it for $60B in June, four days after its own record IPO. ARR went from $100M in early 2025 to over $4B by June 2026, roughly $2.6B of it enterprise, the fastest ramp in business software history, with the company projecting $6B by year end. Four years from incorporation to a signed $60B all-stock deal, with closing targeted for Q3 pending regulatory approval. The time-to-$100B is collapsing.
- Anthropic proves the magnitude case. Founded in 2021. Now a $965B company at a $47B revenue run-rate, with a confidential S-1 already filed. Five years from zero to nearly a trillion dollars in value. Nothing in software has ever compounded this fast at this scale.
- OpenAI is the literal embodiment of the thesis. A $122B round at $852B, roughly eleven years in, generating $2B a month and growing. This is the company that re-priced the entire top of the stack. Sequoia reportedly raised its internal valuation bar for any AI app over $1B after OpenAI’s round reset the price of everything above it.
- Databricks proves boring infrastructure compounds too. Thirteen years, $5.4B ARR growing 65%, free cash flow positive, raising into a $165B to $175B range ahead of a likely late-2026 IPO. No frontier-model drama. Just relentless revenue compounding and a category that re-rated. The “unsexy” path still gets you to $175B.
- SpaceX is the OG that proves the model persists. Twenty-four years private, then the largest IPO in history in June at $1.77T, now trading above $2.5T. It validated the whole stay-private-and-compound playbook before anyone called it a playbook. And it wasted no time becoming the acquirer, pulling the trigger on its $60B all-stock buy of Cursor within days of going public. The companies that ran this game first are now the ones consolidating it.
How and Why The Math Inverted
It’s not that giant outcomes exist. Google, Facebook, and Amazon all happened. What changed is the frequency and the speed. We went from one of these every five or ten years to six of them visible at once, with the fastest doing it in four years instead of fifteen.
When that happens, VC math inverts.
The old discipline was about entry price. Pay 10x too much at the seed and you’ve capped your return, so you fight over valuation. When the ceiling is a $1B exit, a 5x entry mistake is fatal.
When the ceiling is $100B-plus, the entry price matters far less than picking the right company. If you’re right about the company, a “too expensive” entry still returns 50x. If you’re wrong, no price was cheap enough. So the smart money stopped optimizing for the cheapest entry and started concentrating harder into the names that could be firm-makers, not just fund-returners.
This is the part I’ve believed for a long time and run SaaStr Fund on. Concentration beats diversification when outcomes are this skewed. You don’t need 30 shots. You need to be 100% conviction on the few that could be enormous, and put real money behind them.
The Funds Doing the Hunting Got Just as Big
There’s a reason the target moved, and a lot of it is sitting on the other side of the table.
In January 2026, a16z raised over $15B in a single haul, the largest VC fundraise in history. That one raise was roughly 18% of all venture dollars deployed in the US in 2025. The firm now manages north of $90B. Sequoia is in the same neighborhood. Thrive closed a $10B fund. General Catalyst sits above $40B and is reportedly raising another $10B.
Run the math on a $15B fund. It has to return roughly $45B to hit a 3x. You don’t get there backing companies that exit at $1B. You get there by owning a real stake in something that becomes worth $100B-plus. The fund size forces the hunt. Once you’ve raised $15B, a $1B outcome doesn’t move the needle, so you stop chasing it.
That’s why the capital is pooling at the top, and the numbers are staggering. In Q1 2026, global VC hit a record north of $300B, and AI took roughly 80% of it (the highest concentration in any single sector in venture history). Four companies, OpenAI’s $122B, Anthropic’s $30.6B, xAI’s $20B, and Waymo’s $16B, pulled in about 65% of all venture dollars on the planet that quarter. Across 2025, AI captured 61% of global VC, and OpenAI and Anthropic alone took 14% of total global venture investment, more than most entire national startup ecosystems.
The deal count is moving the opposite direction. Late-stage rounds were about 82% of all capital deployed in Q1, and fewer than 3% of deals soaked up nearly 80% of the money. More dollars, fewer companies. A small group of mega-funds writing enormous checks into a small group of would-be $100B winners.
So the “$100B in 10 years” hunt isn’t a vibe. It’s arithmetic. The funds got big enough that nothing smaller moves the needle, and LP money concentrated into the handful of managers who can write $500M to $1B checks. The whole system reorganized around producing and capturing a few enormous outcomes.
There’s a flip side where a smaller fund actually wins. A $15B fund needs the $100B outcome to matter. A concentrated $50M or $100M fund just needs to be early in one of them. You don’t need to write the $500M growth check. You need to own a real piece before the mega-funds show up. That’s the entire game for everyone not named a16z.
This Also Means VC May Be a “Worse” Fit For Many
The pattern is real and rare at the same time, and pretending otherwise is how people lose money.
For every Anduril there are 200+ defense startups that will never clear a $1B exit, maybe far more. The base rate underneath this list is brutal. What you’re seeing is survivorship dressed up as a strategy, and it only works if you can pick the survivor.
The multiples are also extreme. SpaceX is trading near 115x sales. Cursor went out at roughly 25x ARR. If the 2026 IPO wave re-rates these names toward where Snowflake trades, closer to 7x forward sales, late-stage marks compress across the whole category, fast. Bill Gurley put it well earlier this year: AI valuations have reached the point where being right about the product doesn’t protect you from being wrong about the price.
And “several of these at once” can be the top of a cycle just as easily as a new normal. Six visible winners in 2026 is not proof of six more in 2030.
Call It Unfair. But Expectations for Growth Are Insane, Too. They Have To Be
Don’t confuse the outcome with the input. None of these companies raised their way to $100B. They compounded real revenue and got an entire category re-rated around them. Anthropic at $47B run-rate. Databricks at $5.4B and cash-flow positive. Anduril at $2.2B doubling. The valuations are downstream of the compounding, not a substitute for it.
What the winners share: they created or redefined a category, the revenue growth outpaced the capital raised, and in the capital-heavy ones, they could deploy billions productively instead of just burning it.
That’s the target. Not “raise at $10B.” Build the kind of compounding that forces the market to re-rate everything around you.
Call it unfair. Call it concentrated. But now that there are several of them, it’s no longer insane to build a fund around hunting the next one. The hard part hasn’t changed. You still have to be right about which company, and the price you pay still has to make sense at exit.
