Haun Ventures Secures $1B Amid Crypto VC Slowdown—Doubling Down on Stablecoin Rails for AI Agents

Crypto venture is cooling, yet Haun Ventures just expanded its firepower. Katie Haun closed $1 billion across two new funds while April 2026 crypto VC fell to $659 million across 63 deals—down 74% month-over-month and the weakest print since July 2024. Year-to-date, $5.64 billion has been invested, but the run-rate lags 2025’s pace. Since the October 2025 peak of $3.84 billion in monthly funding, each month has slid lower, echoing a 37% contraction in total crypto market cap over the period. Against that tape, raising $1 billion isn’t just headlineworthy—it’s a signal.

Here’s the signal: Haun is scaling counter-cyclically. Firm AUM has climbed from $1 billion to $2.5 billion even as other crypto-native peers—Paradigm, Pantera, and a16z crypto—have seen AUM compress since 2025’s high-water mark. That divergence isn’t narrative; it’s driven by realized outcomes. Haun’s first fund backed Bridge at a $100 million valuation before Stripe bought it for $1.1 billion, and invested in BVNK at $678 million before Mastercard acquired it for $1.8 billion—the largest stablecoin acquisition to date. The throughline across both exits is simple: stablecoin infrastructure became essential financial plumbing, and the buyers were payments incumbents, not crypto consolidators.

The new funds extend that plumbing thesis into the world of AI agents. The premise: autonomous and semi-autonomous software will need regulated financial stack components—wallets, compliant payment rails, and audit-ready identity/compliance layers. The teams that built the stablecoin rails have a head start on distribution, controls, and trust with regulators. Haun is explicit that this isn’t a pivot to “be an AI fund”; it’s an expansion of the same regulated-infrastructure playbook into the next buyer of those services.

The anchor position today is Erebor, Palmer Luckey’s federally insured digital bank valued at $4.35 billion. Unlike consumer neobanks, Erebor is architected for AI and defense technology companies—entities that need programmable treasury, clean counterparty risk, and continuous uptime across domestic and cross-border flows. If AI agents are going to spend, bill, and settle programmatically, they’ll need institutions that can intermediate fiat, stablecoins, and tokenized instruments with traceability and permissioning that pass supervisory scrutiny.

This is where the opportunity—and the tension—sits.

- Technological edge: AI agents require wallet primitives that are safe-by-default, policyable, and recoverable. On-chain settlement offers speed and composability, but production deployments need circuit breakers, role-based controls, and integration into sanctions and fraud tooling. Stablecoin-rail veterans already ship those controls. If Erebor can combine bank-grade balance sheets with on-chain finality, it becomes a native bridge between agent workflows and real-world money.

- Investor psychology: Raising into a downturn often reflects LP conviction in a manager’s distribution to acquirers. Stripe and Mastercard validating prior positions matters more to institutions than a token chart. That track record reframes “crypto exposure” as “infrastructure M&A optionality,” which tends to remain open even when token prices chop.

- Business model reality: The exit path here likely remains skewed to strategic sales. Payments networks, processors, and fintechs will keep buying critical-path components—compliance-forward wallets, settlement middleware, and agent-facing APIs—before launching them at scale. Take rates on pure stablecoin flows may compress over time, but embedded distribution and bank licenses are defensible moats. A regulated bank tailored to AI spend could own the highest-value junction: where programmable agents hit insured deposits, treasuries, and 24/7 settlement.

- Governance trade-offs: The very features that make regulated rails attractive—KYC, reversibility, supervisory visibility—introduce centralization risks and the temptation to over-permission. As tokenized finance grows, design choices will either enshrine user agency or slide toward black-box gatekeeping. Building credible transparency and user protections without replicating the worst frictions of legacy banking is where reputations will be made or lost.

Context also favors this posture. The market’s migration toward compliant tokenization is accelerating: the DTCC, which processed $4.7 quadrillion in 2025 and custodies $114 trillion, plans live tokenized securities trades in July with a commercial rollout in October. The service will tokenize DTC-custodied assets—Russell 1000 stocks, major ETFs, and U.S. Treasuries—under an SEC no-action framework and in collaboration with a deep TradFi-crypto working group. That’s the same “regulated wrapper, digital mobility” thesis Haun is leaning into.

Can this strategy compound? It can if the bottleneck remains where it has been: between open networks and compliant balance sheets. If AI agents become routine economic actors, the winners will likely be those who own the payment, identity, and settlement chokepoints—and can sell them to both startups and incumbents. Haun’s $1 billion gives her portfolio time and runway to build those choke points while the rest of the market waits for price to lead conviction again.

Meanwhile, Bitcoin sits around $81,000 and ETF flows remain strong, but the more durable story is capital rotating toward pipes and permissions. In downcycles, the market often decides what actually matters. This time, it seems to be regulated plumbing for autonomous finance.