To the Xenith community,
This is the eighth annual letter we've written since 2018. Each year we try to do three things in this document. Tell you what we did, tell you what we got wrong, and tell you what we believe matters for the year ahead. The format is unchanged. The conviction underneath it is sharper.
The year in numbers
In 2025, we wrote 22 first checks, deployed $78M, and participated in 14 follow-on rounds. Three of our companies had liquidity events. The portfolio is now 110 active investments across 24 countries, with the team operating from offices in New York, Singapore, and London.
Fund III is now 31% deployed against a $250M committed pool. We expect to be 60% deployed by the end of 2026, with reserves earmarked for follow-ons in 8–10 of our existing companies.
What worked
The clearest signal of the year was that our 2023 vintage is doing well. The companies we backed in the quietest part of the cycle are now the ones with the strongest fundamentals. Blockshift crossed $40M ARR. Chainforge grew to 110,000 weekly active developers. NexaSphere announced a strategic round at five times our entry price.
Our hands-on operating support showed up in the data. Companies where we ran a structured 90-day onboarding closed their next round at a 38% higher pre-money than companies where we didn't. We're now defaulting to that program for every seed-stage check.
What we got wrong
Two things, both worth being explicit about.
First, we underweighted consumer in 2024. We had a thesis about onchain consumer surfaces and we wrote about it (see "Web2 and Web3, in one portfolio") but we deployed only 12% of capital there in 2024, when we should have been at 25%. The companies that broke out in 2025 — wallets, social, AI agents holding wallets — were companies we had conviction about and didn't move fast enough on.
Second, we backed two companies whose go-to-market motion didn't match their product. Both pivoted in 2025, both are now back on track, but the lesson is that our diligence on enterprise sales motions has been weaker than our diligence on technical depth. We've added a second reference call dedicated specifically to GTM for any company selling into the enterprise.
What we believe matters in 2026
Three trends, in order of conviction.
1. The AI-agent economy will pull crypto rails to the front. Agents need wallets, agents need micropayments, agents need stablecoins. The infrastructure that exists today was built for humans transacting with humans. The next wave is built for software transacting with software. We expect 30–40% of our 2026 deployment to be in this seam.
2. Real-world asset tokenization is the quiet giant. Stablecoins are now the most-used onchain primitive. The next layer is tokenized treasuries, money-market funds, and credit. The regulatory unlock has happened in three major jurisdictions in 2025, and the application layer is ready to follow. We've made four investments in this category this year and expect more.
3. The Web2/Web3 line will keep dissolving. The most interesting consumer apps shipping right now don't market themselves as crypto, even though half of them route through onchain rails. Founders building this way are exactly the ones we want to back: technical, distribution-aware, allergic to ideology.
The road ahead
We will keep doing what we've done since 2018. Write first checks into founders building the open internet. Show up after the wire. Stay on the cap table through the full cycle. Tell you in writing what we got wrong and what we got right.
If you're building something on the seam between Web2 and Web3, infrastructure or consumer, we want to hear from you. Pitch us. Always within five business days.
Thank you for being part of this.
Adrian Lee & Sophia Martinez
Founding Partner & Managing Partner
Xenith Ventures
December 2025