Framework Ventures has raised another $400 million, but the more interesting part is where the old DeFi investor now wants to spend it.
Framework Ventures is no longer trying to be understood as just a crypto fund. According to Fortune, the San Francisco firm cofounded by Vance Spencer and Michael Anderson closed a $400 million fourth fund on Friday and is presenting it as a bet on frontier technology: AI, robotics, energy, and Web3 under the same roof.
That phrase can get mushy fast, so stay with the concrete pieces. Fortune reported that roughly half of the new fund has already been deployed. Spencer and Anderson declined to name the fund's limited partners, but described a base that includes funds of funds, an Ivy League endowment, sovereign wealth funds, and nonprofits. That isn't the investor list you expect for a small crypto specialist trying to survive on old DeFi glory.
Framework has earned the right to make the argument, but it still has to prove the argument works. The firm built its name on early crypto infrastructure and DeFi bets, including Chainlink, Synthetix, and Axie Infinity. It raised a $100 million second fund in 2021, then a $400 million third fund in 2022. Framework's own press page still points to that 2022 Bloomberg item as the moment it launched a $400 million vehicle, which matters here because the new fund matches that number rather than leaping past it.
That's the honest read. This is a strong fund close in a bruised category, not a victory parade.
The firm's latest SEC filings put assets under management at $1.28 billion as of December 2025, according to the article. The fourth fund is already showing the broader mandate in practice. Framework is citing a stake in Mecka AI, a robotics data startup, and a position in Better.com, the public mortgage company, as signals of where the capital is going. If you came to Framework through DeFi, Better.com is the name that should make you pause. It tells you the firm is willing to leave the old lane, not just decorate the deck with AI language.
Spencer and Anderson's explanation is that they're following founders rather than chasing a market slogan. That's a plausible story. Crypto-native builders have spent the past two years moving into AI agents, decentralized compute, physical infrastructure, and energy-linked networks. A firm that knows those operators well can see a shift before the wider market packages it into a clean trend.
But here's the thing: founder access is not the same as category edge. Framework had a clearer advantage in DeFi around 2019, when many institutional investors still treated smart contracts as a strange corner of finance. AI and robotics are different. Andreessen Horowitz, Sequoia, and Khosla Ventures have been living in those markets for years, and they don't need a tutorial before they can evaluate a robotics data company. Framework is betting that its network, technical taste, and experience with tokenized systems will travel. Maybe they will. You shouldn't assume they automatically do.
The timing also deserves a colder look than the announcement language allows. The broader crypto venture market spent 2023 and much of 2024 in retreat. Venture investors put more than $25 billion into crypto companies in 2025, a 73% increase from the year before, according to the figures cited in the article, but deal count fell to around 1,200 transactions from more than 2,900 in 2024. That tells you the recovery has been selective. Fewer companies got funded, and the checks were bigger.
That selectiveness cuts both ways for Framework. LPs are not writing easy checks to every manager that raised at the 2021 and 2022 peak. Many funds from that period struggled to beat bitcoin itself, which is a brutal comparison when you're asking institutions to pay venture fees. So a $400 million close, with endowment and sovereign wealth participation, says something real. It says Framework is one of the crypto-affiliated firms that still has enough credibility to raise institutional capital after the washout.
It also says LPs want optionality. They may still want exposure to Web3, but they don't want a manager trapped inside one cycle's vocabulary. DePIN, decentralized compute, AI agents, robotics data, and energy infrastructure are now close enough that investors can sell the package as one broad systems bet. You don't have to love the branding to see why it works. The same founder who once built token incentives for a protocol may now be building machine data rails or compute markets for AI workloads.
Still, the fund will be judged on outcomes, not on whether the thesis sounds current. Mecka AI and Better.com are useful markers because they make the pivot visible. They also raise the right question. Is Framework expanding from a genuine strength, or stretching away from the market where it had the most unusual insight?
You won't know that this year. Venture funds take time, and this one is already half spent. For now, the important fact is simpler: Framework closed another $400 million vehicle at a moment when crypto managers still have to explain what they learned from the last cycle. Spencer and Anderson's answer is to keep Web3, add AI and robotics, and follow the founders who are moving between them. That's a clear bet. Now it has to earn the size of the check.
Also read: Ripple's RLUSD stablecoin goes live in Japan through SBI, setting a regulatory template for dollar coins in Asia • Solana's price crash to multi-year lows tells only half the story of what the network is actually doing • North Korea stole $1.5 billion from Bybit and Iran's central bank quietly spent it
