The press release will tell you that crypto VC Haun Ventures raises $1b for blockchain and AI investments. What it won't tell you is that this fund is smaller than the last one — and that's actually the most interesting thing about it.

In 2022, Katie Haun raised $1.5 billion for her debut fund after leaving Andreessen Horowitz to launch Haun Ventures. That was peak crypto exuberance. FTX was still standing. NFT marketplaces were printing billion-dollar valuations. The environment rewarded ambition. Haun raised the most ever for a debut crypto venture fund by a solo GP — and then watched the market crater. The Block

The new fund is slightly smaller than the debut because the team expects less dramatic swings in liquid token prices — a remarkably candid admission. It signals something important: Haun Ventures isn't trying to ride another crypto wave. It's repositioning for a longer, slower, more structural trade. The volatility-extraction phase of crypto investing is over, at least for this firm. What's next is infrastructure — and specifically, the infrastructure that AI agents will need to transact like economic actors. The Next Web

The Bet Underneath the Bet

The $1 billion is split evenly between early and later-stage funds, to be deployed over the next two to three years into crypto and blockchain companies. That's the clean version. The revealing version is what Haun said herself: she wants to invest in startups at "the new frontier" of finance, where financial services, AI, and alternative assets collide. Bloomberg

The specific claim that animates this fund — the non-obvious one — is about AI agents. As AI agents take on a growing share of tasks, they'll need native financial rails: payment systems, credit, identity verification, and fraud prevention designed specifically for how machines transact, not humans. That's not a marginal observation. If you believe it, it has enormous implications for where venture capital should be flowing right now, and what kinds of companies should be getting funded. Decrypt

Consider what "an AI agent with a bank account" actually requires. It can't walk into a branch. It can't pass a KYC interview. It needs programmable, machine-readable financial infrastructure — custody solutions, on-chain settlement, automated compliance rails. Haun Ventures said it will target companies redesigning backend payments and settlement systems, custody and compliance tools for tokenized assets, and credit and risk products for AI-driven transactions. That's a specific enough thesis to actually build a portfolio around.

The Track Record That Got the LPs to Write Checks

A $1 billion raise doesn't happen without something to point to. Haun's first fund wasn't uniformly brilliant — Haun Ventures invested in OpenSea in a round that valued the NFT marketplace at $13.3 billion; a top investor later marked the company down to $1.4 billion. That's a near-total wipeout on paper.

But the wins more than covered it. Haun scored a pair of notable stablecoin-related exits: BVNK, acquired by Mastercard for as much as $1.8 billion, and Stripe's acquisition of Bridge, valued at $1.1 billion. Both were bets on payment infrastructure, not speculative tokens. Both delivered real exits to real acquirers at real prices. The firm also bought distressed crypto assets at discounts during the FTX-era downturn and sold them at peak prices in 2025, returning capital to limited partners.

That combination — infrastructure bets with strategic acquirers, plus counter-cyclical token positioning — is what convinced LPs to come back. Accolade Partners founder Joelle Kayden, a limited partner in the fund, pointed to the staking strategy and token trading as key contributors to investor returns. It's a more sophisticated story than "we picked good crypto companies." It's closer to a hedge fund mentality married to a venture fund structure.

"I've been following the flow of assets my entire career, and this is the most dynamic period in technology and finance I've ever witnessed." — Katie Haun, founder and managing partner, Haun Ventures

That quote could sound like PR boilerplate. But Haun has an unusual vantage point. She's a former DOJ prosecutor who investigated digital asset crime before it was a mainstream concern, then joined Andreessen Horowitz to help build one of the earliest and largest crypto investment franchises in venture, then left to run her own shop through the worst downturn the industry had seen. Her career spans every phase of digital finance — from when Bitcoin was a money-laundering concern to when stablecoins process volumes approaching Visa and Mastercard combined. That context matters when evaluating whether her AI agent thesis is genuine conviction or fundraising narrative.

The Flagship Bet: A Bank That Didn't Exist Two Years Ago

The most concrete signal of where this fund is going is already deployed. Haun Ventures is one of the largest investors in Erebor, a digital bank co-founded by Anduril Industries founder Palmer Luckey and backed by Peter Thiel's Founders Fund. Erebor received FDIC deposit insurance approval in late 2025 and raised $350 million at a $4.35 billion valuation. It's designed to serve technology companies working with digital assets, artificial intelligence, defence, and manufacturing.

What Erebor represents is a federally regulated bank built not for consumers but for companies — specifically, companies building AI agents, processing digital assets, or working in sectors that existing bank infrastructure wasn't designed to serve post-SVB. Stripe has already launched a machine payments preview integrating stablecoin settlement for agent-to-agent transactions, which tells you the demand signal is real. The question isn't whether AI agents will need financial plumbing. It's who builds it first and who gets regulatory approval to operate it at scale.

Stablecoin transaction volumes reached the double-digit trillions in 2025, approaching the combined processing volume of Visa and Mastercard — the figure Haun Ventures cited to justify deploying a billion dollars into this infrastructure gap. That number is striking. If accurate, it means stablecoin rails are already carrying systemic volumes; they're just not carrying them through regulated, insured, bank-chartered institutions yet. Erebor is the bet that they will.

The Global Dimension: Who's Actually Building This Infrastructure

Haun Ventures is competing for deals in a category that barely existed 24 months ago. Paradigm, one of the largest crypto-focused venture firms, raised $1.5 billion in February for a new fund that will invest in AI and robotics alongside its core blockchain portfolio. Matt Huang, Paradigm's co-founder, said developments in AI are too interesting to ignore, though the firm intends to focus specifically on the intersection of AI and crypto rather than chasing general-purpose AI deals. The Next Web

Globally, the regulatory picture is shifting fast — and unevenly. In the United States, the GENIUS Act, moving through Congress in 2025, would create the first federal stablecoin framework, which would directly benefit companies in Haun's portfolio building payment infrastructure on stablecoin rails. The EU's MiCA regulation, which came into full effect in late 2024, has already driven several crypto infrastructure companies to establish European entities — creating a bifurcated regulatory environment that favours well-capitalised, compliance-first operators over small-scale token projects.

In Asia, Singapore's MAS has quietly become the regional home of choice for stablecoin infrastructure companies seeking a credible regulatory anchor. Firms like Paxos have established significant Singapore operations. Hong Kong has re-entered the conversation as a crypto-friendly hub under its 2023 virtual asset regulatory framework, and several stablecoin issuers are actively pursuing Hong Kong licences as a gateway to institutional Asian capital. For a firm like Haun Ventures investing globally, the regulatory map is as important as the technology map — and right now, it favours operators who can navigate multiple jurisdictions simultaneously.

A founder building at the intersection of AI and financial infrastructure right now faces a strange choice: raise from a traditional fintech VC who doesn't really understand crypto, or raise from a crypto VC who's just starting to understand AI. Haun's argument — implicitly — is that her firm is the one that's spent years at the junction of both.

The Skeptic's Corner

There's a version of this thesis that doesn't work. AI agents transacting autonomously raises genuinely hard questions about liability, fraud, and regulatory ownership that no one has answered yet. Who's responsible when an AI agent executes a bad trade, sends stablecoins to the wrong address, or gets exploited? The financial infrastructure Haun wants to fund will need compliance tooling, identity layers, and fraud detection systems designed from scratch — and regulators haven't shown they're ready to approve them at speed. Erebor took just four months to get conditional OCC approval, which is unusually fast. That pace probably reflects the current administration's pro-crypto posture as much as it reflects Erebor's quality. A different administration could slow that runway considerably.

There's also the question of whether AI agents will actually need crypto-native financial rails, or whether they'll just transact through existing bank APIs. Stripe, JPMorgan, and SWIFT are not standing still. If traditional finance absorbs the AI agent payment layer before crypto infrastructure matures, Haun's portfolio could end up stranded at an interesting but unbankable intersection.

What to Watch

• The GENIUS Act's progress in the U.S. Senate. Federal stablecoin legislation would validate Haun's entire payment infrastructure thesis in one stroke. A stall or failure would create significant headwinds for the regulatory moat her portfolio companies are building toward Erebor's commercial launch in 2026. It's the most tangible test of whether the "bank built for AI companies" thesis converts from a compelling pitch into actual deposits, actual clients, and actual revenue. If it struggles to attract customers beyond the crypto-adjacent founder community, the AI agent banking narrative needs revision.

• Whether Paradigm's AI-and-crypto fund attracts the same deals as Haun. Two of the most credible crypto-native firms are now competing for the same category. Valuations in that gap will compress, and the differentiation will come down to portfolio construction quality and founder relationships — not thesis.

The honest read on this fund:

The crypto VC Haun Ventures raises $1b story is a bet on financial plumbing at a moment when the pipes are being redesigned. Haun's track record on infrastructure — Bridge, BVNK, Chainalysis, Erebor — is better than her track record on pure-play speculation. The smaller fund size relative to 2022 is discipline, not weakness. The AI agent thesis is genuinely non-obvious in its specifics, even if the broad "AI and crypto will converge" framing has become a fundraising cliché. The global regulatory tailwind is real but fragile. And the OpenSea write-down is a reminder that even well-positioned firms make expensive mistakes when sentiment runs hot.

Founders building in this space should pay attention — not because Haun has $1 billion to spend, but because the thesis she's deploying behind it is one of the cleaner frameworks for where machine-to-machine finance is actually heading. That's worth understanding regardless of whether you're raising money.