Before they are legible.
Three clusters of physics-native founders on the hinge between a far better world and a far worse one. Plus the engine I would build to reach them first.
The best founders are hardest to see at the moment they are cheapest to back.
Consensus is expensive by the time it is legible. A ZBLAN draw on the ISS, a viral counter-surveillance orb, a stroke-recovery device out of Brazil: each looked like a stunt or a science project before it looked like a category. BlueYard's manifesto asks for the two-sided bet, the technology that moves us toward utopia or holds the line against oblivion. This brief names three places that bet is live, the people carrying it, and how I would build a machine that meets them before the round, instead of chasing them after.
Neuroscientist and biomedical engineer building non-invasive neuromodulation for motor recovery. Orby's device sits on the skin, sends patterned signals to the nervous system, and helps people with spinal-cord injury stand and take first steps, no surgery. Forbes Under 30 Brazil; now based in the Bay Area.
Building Capable, a rapid-iteration therapeutics company for human capability, starting with sleep and cognition. Separately, at MIT's Boyden lab, he co-authored the State of Brain Emulation Report 2025 and a worm-to-human scaling thesis: the reference roadmap for a field of fewer than 500 people worldwide.
Harvard physicist, ex-CERN researcher. Her device, Spectre I, scans for nearby microphones and emits signals that garble what recorders capture while humans still hear normal speech. The launch video passed four million views in 48 hours; pre-orders are open at a $1,199 refundable deposit.
I pressure-tested the premise that off-world founders skew senior. It fails. Space Forge began in 2018 as a pub conversation between two Thales Alenia colleagues; the registration cost £10 and the first office was a Bristol garage. Western had entered the industry only four years earlier. Varda was co-founded by a Founders Fund partner still in his twenties, paired with an ex-SpaceX engineer. The true pattern is a pairing: a young commercial founder sets the clock speed, while flight heritage concentrates in the CTO seat and on the cap table. Where a team is all heritage with no young engine, as at Flawless Photonics, the cadence follows the ISS schedule. That is the filter applied below.
after the first meeting
The world's first contract research organization for drug discovery in orbit. Its Orbital High-Throughput Screener runs more than 2,000 automated experiments per mission and holds a payload slot on Vast's Haven-1, the first crewed commercial station, with a maiden mission scheduled for 2027. Founded in Luxembourg in late 2024; roughly twenty people; paying clients signed in year one.
Semiconductor crystals grown on returnable satellites. ForgeStar-1 launched in June 2025 as the first UK-licensed in-space manufacturing satellite, and in December 2025 generated the first commercial plasma in orbit: the first free-flying commercial semiconductor tool ever operated in space. Bacon designed the payload that did it. Pridwen, the deployable heat shield, is the return piece ForgeStar-2 will exercise.
The engine is the thesis.
The founders above make a list. The reason I can name them, and the reason I will name the next twenty before their rounds, is a compounding machine I already run in prototype: media that pulls, capital-light capture that wins first look, and convening that deepens the relationship. Built to BlueYard's standard, and to BlueYard's advantage.
Public theses draw the right people toward me instead of me chasing them. The podcast runs in three formats; the Substack organizes into named Research Groups in the register of MIT Media Lab groups like Cyborg Psychology and Critical Matter.
Small, fast, non-dilutive instruments that put the Frontier State in a founder's story before a priced round exists. Modeled on the fastest programs in the world, and tuned wider than any of them.
Rooms turn a first contact into a relationship a screen cannot replicate. Branded to the standard of a16z, Khosla, and General Catalyst, and classier than DeepTech Week, which built real gravity in 18 months.
- Launch Frontier State Dinners: Chatham House, monthly, SF and Cambridge.
- Work the circuit as a scout: Deep Tech Week, Foresight and Cosmos salons.
- Restart the Substack and podcast on the new Research Group structure.
- Scale Dinners into a named, waitlisted series across SF, Cambridge, and a European node.
- Launch Frontier State Grants and the first Fellows cohort for first look.
- Open a Catalysts salon for pre-company lab and PhD talent.
- A flagship annual Assembly for the frontier-of-frontier.
- Global founder immersions no other GP runs.
- The loop compounds; talent arrives on its own.
Networks most partners cannot buy, built for a geography-free fund.
Ideas first. Money second, as access.
That is the argument in Paul Graham's Cities and Ambition, which my next podcast takes apart with a philosopher: power flows from ideas, and the city tells you which ideas it rewards. Curation is where ideas and founders meet. The rooms give me the ideas, the ideas bring the founders, and the engine becomes the thesis. I would like to build it here.
- blueyard.com: track record (Castelion pre-seed, $100M A, $350M B), four sections, $500K–$5M pre-seed to A model. Accessed Jul 2026.
- Space Forge + World Fund releases, May 14, 2025: $30M Series A led by NATO Innovation Fund; World Fund first backed 2021.
- spaceforge.com mission log, Jun 2025 to Mar 2026: ForgeStar-1 launch on Transporter-14, first UK CAA in-space manufacturing licence, Pridwen plan.
- Space Forge, Dec 2025: first commercial plasma in orbit; first free-flying commercial semiconductor tool. Payload credited to Bacon on the team page.
- Space Forge totals: Tracxn $40.9M; PitchBook $39.2M (2026). Site says roughly $41M.
- Millennial Masters interview, Feb 2025: pub origin, £10 registration, Bristol garage, 70-plus team.
- Varda: CNBC, Jul 10, 2025 ($187M C); mission log per Wikipedia, 2026 (W-1 Utah re-entry Feb 2024; W-4 lost 2026); Series D figure per Clay dossier, verify before quoting aloud.
- Flawless Photonics: SpaceNews Feb 2024; The Register Apr 2024 ($1,000/m orders); Electro Optics Oct 2024 (11.8 km); New Space Economy profile Feb 2026 (senior bench).
- Exobiosphere: Startup Luxembourg Apr 2025 (€2M seed); Forbes Luxembourg Jun 2026 (Haven-1 2027, next round, Cedars-Sinai extension); med-tech.world Apr 2026 ($1M Drapers prize).
- Founder backgrounds: spaceforge.com/team and about-us; theorg.com (Western at Thales from 2014).