Haun Ventures has closed $1 billion across two new funds split evenly between early and later-stage vehicles, with founder Katie Haun expanding the firm's mandate beyond crypto and blockchain into startups that combine financial services, artificial intelligence, and alternative assets, reflecting a deliberate bet that AI agents will become the next meaningful demand driver for the infrastructure her existing portfolio is already building.
The firm's track record going into this raise is the context that makes the mandate expansion credible rather than opportunistic. Haun backed Bridge, the stablecoin infrastructure startup that Stripe acquired for $1.1 billion, before that outcome was obvious. It backed BVNK, which Mastercard agreed to acquire for up to $1.8 billion earlier this year. It backed Fireblocks and Chainalysis through secondaries, and Superstate, which raised $82.5 million in January in a round that also included Bain Capital Crypto and Brevan Howard Digital. These are not speculative early-stage swings. They are positions in financial infrastructure that turned out to be exactly what payment companies, banks, and card networks needed as stablecoins moved from speculative rails into production settlement systems. The portfolio built the case for the next thesis before the next thesis was stated explicitly.
The AI agent argument Haun is now making is a specific one, not a general gesture toward AI. Her position, articulated publicly and to LPs, is that AI agents will eventually transact and exchange value on behalf of their users, and that this requires new financial rails designed for machines. That is a more precise claim than the generic observation that AI and crypto will converge. It points to a specific infrastructure gap: payment systems, identity frameworks, and custody solutions that were designed for humans authenticating actions cannot handle the permission models, spending controls, accountability requirements, and audit trails that autonomous software agents require when they are executing financial transactions. An AI agent booking a flight, paying a contractor, or rebalancing a portfolio needs a different financial primitive than a human logging into a bank app. Haun's crypto infrastructure portfolio happens to sit at the layer that would build those primitives.
The fund structure is identical in design to her 2022 raise, though smaller in total. The first pair of funds split $1.5 billion between $500 million early-stage and $1 billion late-stage. The new pair splits $1 billion evenly, $500 million each, reflecting a deliberate calibration toward current market conditions rather than peak-cycle sizing. Both pairs are likely to be oversubscribed, according to people familiar with the raise. The fact that Haun Ventures is one of only a handful of top-tier crypto VC firms that saw assets under management grow in the first quarter of 2026, rising more than 30% year over year to nearly $2.5 billion while Paradigm, Pantera, and others saw portfolio values decline, gave LP conversations a different tone than the one available to peer funds entering the market at the same time. Timing and track record combined to make this a more straightforward fundraise than the April 2026 VC funding data would suggest is available to most crypto-focused firms.
The question founders should ask about this mandate expansion is whether the AI agent thesis genuinely validates a crypto-plus-AI product or creates pressure to bolt crypto onto an AI agent product that does not need it. Haun's stated rationale is coherent: if AI agents need to hold, spend, and receive value autonomously, the existing financial infrastructure, bank accounts with human login requirements, card rails requiring card-present authentication, ACH with business-day settlement latency, is not adequate for machine-speed, programmable financial execution. Crypto's programmable payment rails, stablecoin settlement, and smart contract-based permissioning are architecturally better suited to what autonomous agents require. That argument is strongest for use cases where an agent genuinely needs to execute financial actions without human approval at each step: autonomous vendor payments, dynamic API cost management, real-time royalty distribution, micro-payment settlement in high-frequency agent workflows.
It is weakest for the majority of consumer AI agent use cases that do not actually require crypto rails to function. An AI scheduling assistant that books calendar appointments, an AI writing tool that manages draft revisions, or an AI research agent that browses the web does not need a blockchain wallet. Attaching crypto infrastructure to those products because a fund has a crypto mandate does not make the product better. It adds complexity, custody risk, and regulatory surface area that the product does not require. Founders pitching to Haun Ventures should expect to be asked specifically why their AI agent product needs decentralised financial rails rather than a conventional API connection to a payment processor, and should have a genuinely specific answer to that question rather than a general convergence narrative. The cases where the answer is compelling are real and interesting. They are also a subset of the AI agent market, not the whole thing.
The broader signal from Haun's raise is the one the market will carry. Of the top-tier crypto-focused funds that raised or attempted to raise in the 2024-2026 window, Haun achieved a clean close at a meaningful size in an environment where crypto VC deal volume hit a two-year low in April. The firm's willingness to expand its stated mandate reflects both genuine conviction about where the AI and crypto stacks intersect and a practical acknowledgment that LPs writing large cheques into specialist crypto funds in mid-2026 want a thesis that has legs across multiple technology cycles rather than one that depends entirely on the next token price appreciation event. Framing the fund around AI agent infrastructure gives the portfolio a second growth driver that is not correlated with crypto market cycles. Whether the startups that benefit most from that thesis have already been founded, or are still being built in someone's garage right now, is the question Haun's team will be answering one investment at a time over the next two to three years.
Also read: Someone Used Morse Code to Trick Grok Into Sending $174,000 and It Has Happened Before • Anthropic Is Handing Wall Street the Keys to Its Enterprise Distribution and That Changes the AI Services Landscape • Japan's $2,000 cardboard drones are not a novelty and the defense startup implications are more serious than the material suggests
