Applications for Cohort 2 of Antidote are now open. The latest episode of Antidote Inside is a conversation between Andy and Ben about what the programme involves, what we’ve learned from running Cohort 1, and the three areas we’re particularly interested in funding this time around. If you’re a founder considering applying — or thinking about starting something in the Bitcoin space more broadly — this post is the written version of that conversation.
The programme. Six months in total, based at our office in Hatton Garden. The first three months are structured: weekly sessions covering product, roadmap and sprint planning, go-to-market, sales, growth, and fundraising. The curriculum draws on the frameworks used by accelerators like Y Combinator, Techstars, and Hexa, adapted for the realities of building in 2026 — particularly the impact of AI tooling on what a small team can ship. The second three months are less structured: founders work out of the office, apply what they’ve learned, get traction, and prepare for demo day, where they meet relevant seed investors in London. We invest capital, and we work closely with founders throughout — both Andy and Ben have founded and scaled early-stage businesses, so the support is hands-on rather than advisory at a distance.
We have three investment theses for Cohort 2.
The creator economy. Creators on platforms like Patreon, Substack, YouTube, Twitch, and OnlyFans run global businesses, but the underlying payments infrastructure was not built for cross-border peer-to-peer transfer at internet scale. Bitcoin rails — combined with mature on- and off-ramp APIs from providers like Strike, Lightspark, ZBD, and Coinbase — solve much of the technical problem. The opportunity is in building products that abstract Bitcoin away from the end user, focus on the actual creator-to-consumer experience, and address real network effect challenges in specific niches. The unsolved problems here are largely product design and go-to-market, not protocol.
Agentic commerce. A growing share of internet activity is being generated by AI agents rather than humans, and that share will continue to grow. Agents cannot open conventional bank accounts, which makes internet-native money a natural fit for agent-to-agent value transfer. We’re interested in founders building beyond the basic “give an agent a wallet” pattern — shopping agents that work across merchants, agent-compatible commerce experiences, infrastructure for agents to discover and pay one another for tasks, and the harness layer around models more broadly.
Fintech and money. The financial stack built on legacy rails — credit, lending, insurance, trading, payments — is being rebuilt on faster, cheaper, 24/7 rails that don’t require permission to cross borders. Bitcoin’s properties make it particularly well-suited as collateral and as a settlement layer, and credit systems on Bitcoin are emerging in a credible way for the first time. We’re interested in founders building any part of this stack, particularly products aimed at a younger generation that is more native to internet-based finance.
A practical note on AI: when we sat down to design the curriculum at the end of last year, we scrapped and rewrote the product section. The pace of change in agentic coding tools has meaningfully changed what a small team can ship in a short time, and it has changed what seed investors expect to see at the point of fundraising. Founders we work with are expected to make heavy use of these tools, and we do the same in how we run Antidote.
If you’re working on something in or adjacent to these areas, we’d like to hear from you. You can apply at antidotebtc.com — we’ll either set up a video call or, if you’re in London or able to travel, invite you in to talk through your idea in person.
Listen to the full episode wherever your get your podcasts for the longer conversation, including more detail on the curriculum, what we’ve learned from Cohort 1, and where each of these theses came from.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.