We’ve made a $100K OTC investment into Macrocosmos’ Subnet 25 – Mainframe.

Mainframe’s mission is simple: help scientists understand how molecules behave – faster and cheaper than ever before. Instead of waiting weeks to run expensive experiments in traditional wet labs, researchers can now simulate those experiments using AI in hours. That means a pharmaceutical company can model hundreds of drug candidates before committing a single dollar to real-world testing.

This isn’t a small upgrade – it’s a shortcut through billions in R&D spend.

The top 10 global pharma firms spend over $100B a year trying to find the next big treatment. Even a tiny improvement in speed or accuracy has massive commercial value. Mainframe offers both.

Mainframe works like a decentralised folding supercomputer – one that’s always on, always learning, and getting smarter with every protein it folds. It’s already completed over 400,000 simulations and is being used by partners like Rowan Scientific to generate high-quality training data for next-gen AI drug discovery tools.

You don’t need to be a PhD to understand the upside:

• Faster experiments mean quicker drug pipelines

• Better models mean fewer failed trials
• Lower costs mean bigger margins

Mainframe lets pharma companies simulate what used to require labs full of scientists and months of work – and makes that capability available through a simple API. It’s a no-brainer for teams building in biotech, materials science, or AI-assisted healthcare.

Siam Kidd, DSV CIO:

“Mainframe is decentralised science done right. No hype, just hard compute. The simulations they’re running would take traditional labs weeks. Steffen and the Macrocosmos crew are turning physics into product – and we’re backing that with full conviction.”

At DSV, we don’t just bet on tokens – we back tools that push science forward. If you’re building something that makes discovery faster, cheaper, or more open – let’s talk.