This is the first of what I hope will be an annual letter. Each January I’ll write one of these — the ventures that shipped, the numbers that moved, and the lessons I want to remember next year. Tone-wise, somewhere between a Berkshire Hathaway annual letter (clean, candid, accountable) and a personal essay (specific, willing to say what didn’t work).
Faktorist
We shipped Faktorist v3 — the credit module — in February. By year-end we had embedded trade credit live across nineteen markets in five countries, with $48M in cumulative facilitated receivables and a 0.8% net default rate. The credit module is now the fastest-growing line of the business; it took 18 months from spec to first dollar and another 11 to scale. Slower than I’d hoped, roughly twice the rate of the previous module at the same stage.
Total GMV crossed $400M for the year (versus $260M in 2025). Cities live grew from 32 to 41. Net retention sat at 122% across the platform, with the credit module pulling it up. The customer count grew 1.8×; the revenue per customer grew 1.5×. The numbers are good. The bottleneck is hiring credit risk staff in Kyiv at the rate the product is demanding.
Made-in-Ukraine
Two more quarterly issues out (N°08 and the autumn issue). The podcast crossed half a million cumulative episode downloads, with a 65% repeat-listener rate, which is the only podcast number I actually care about. We did not hit the “launch a daily newsletter” goal from last year’s plan. We launched a weekly one instead, and I’m okay with that — daily is a commitment we couldn’t make to the editorial standard.
Notable feature: the agri-export cover story in N°08, which got picked up by three European trade publications and led to eleven new advertiser conversations. The advertising side of MiU crossed break-even in October. Six months ahead of plan.
FTR
Testnet through the whole year. Mainnet pushed to 2027. The reason is boring and right: we wanted three quarters of clean settlement data on testnet before committing to mainnet parameters. We got that data. We’ll launch on the back of it. Anyone who was hoping for a mainnet pump narrative is reading the wrong project.
Allowlist closed at 1,840 wallets across 23 countries, weighted heavily toward Faktorist counterparties (which is what we wanted). Whitepaper v2 dropped in November with revised tokenomics — specifically, lowering the network fee from 0.5% to 0.3% after modeling told us the smaller take rate would not materially change the rewards pool but would materially help adoption.
The book
#EverythingWillBeOK is now out in three languages. The audiobook (recorded in October) accounted for a higher percentage of sales than the publishing partner forecast, which is the kind of mismatch that probably explains why most of the trade still hasn’t figured out what actually works for first-time business books in 2026.
I’m drafting a second book — working title The Operating Layer — about the B2B emerging-markets thesis. Draft one is at chapter 4 of 11. Slow going, by design. I’m not in a rush.
Refereeing
Twenty-eight UPL matches. One match I’d like to have back (a controversial late penalty call that the post-match review confirmed was correct, but I’d have read the build-up differently with two seconds more time). A full UPL season plus international assignments through Referee Abroad. Knee held up. Still no plans to step away.
Three lessons I want to remember
One: the right pace for high-stakes infrastructure is slower than the market wants. We pushed FTR mainnet back a year and the world did not end; in fact, almost no one noticed. The people who would have been hurt by a rushed launch — counterparties with real receivables — were the ones who were most relieved by the delay. Build slow for the people you most want to serve.
Two: the second time you do a thing, the cycle time is shorter but the quality bar is higher. The credit module on Faktorist took half as long to ship as the original catalog did. But the bar customers held it to — uptime, reconciliation accuracy, dispute resolution — was three times higher because catalog had already taught them what to expect. Re-up your quality target every time you ship a sequel.
Three: the team you trained in 2022 is the team that ships everything from 2024 onwards. We did not lose a single senior person to attrition again this year. The compounding of five years of the same team — same humor, same shorthand, same scar tissue — is the moat I’m proudest of. It’s also the moat I almost lost in March 2022 and didn’t, because we paid people for three months without expecting work. The dividend on that decision is now in its fifth year and shows no sign of slowing.
2027 commitments
Three commitments I’m putting on the record so I can be held to them next January:
FTR mainnet launches in Q3 2027 with at least two real counterparties live at launch. Faktorist credit hits $200M in cumulative facilitated receivables. The second book’s draft one is complete by year-end. I do twelve more UPL matches and at least one signature UPL run. The magazine reaches issue N°12.
That’s the plan. Thanks for reading. See you in 2028 with the verdict.