You are not selling anything and not offering a term sheet. You are (1) showing genuine, specific interest in their work, (2) being a useful contact (intros, feedback, a sounding board), and (3) securing one soft outcome: “if you ever start a company or spin this out, talk to Daxos first.” That’s the whole ask. Right-of-first-look, not a deal. Never overpromise money or imply a decision Mark hasn’t made.
Who to reach — priority order
| Target | Where | Why them | The ask | Mode |
|---|---|---|---|---|
| Daniel Murfet Jesse Hoogland | Timaeus (Singular Learning Theory / developmental interpretability). Murfet: mathematician, Melbourne. Hoogland: ED. | The purest “cypherpunks today” scene — a real theory of what neural nets learn, near-zero VC. If AI interpretability becomes a deployment gate, the spinout is a category. | Relationship + right-of-first-look on any for-profit spinout (a training-time model-audit / deception-detection company). | Relationship |
| Gizem Gumuskaya | MIT / Tufts. Inventor of Anthrobots (human-cell “living machines”), Michael Levin diaspora. | Highest-convergence scene in the scan (bioelectric morphogenetics). The most natural young founder in the Levin orbit, near Jurvetson & Carmack money. | Relationship + first-look on her spinout. If she is already raising, flag to Mark for a small optionality check. | Relationship |
| CMU Hackerfab lead / Sam Zeloof orbit | Carnegie Mellon Hackerfab; the homebrew-silicon community. | A proven talent factory. The apex (Atomic Semi) is already funded, but the community that mints the founders is cheap to embed in. | Sponsor / show up; identify the next person about to spin out; get first-check rights. | Relationship |
| Nikhil Malvankar | Yale Microbial Sciences Institute. Cytochrome protein nanowires (living bioelectronics). | Owns the correct mechanism and the recombinant work closest to a manufacturable conductive protein. The one investable node in electromicrobiology. | Pre-spinout relationship + first-look; hold any check until there is a spinout with a manufacturable material. | Watch+track |
| Convergent Research + Astera ecosystem | The FRO (Focused Research Org) world — connectomics (E11 Bio, Forest Neurotech), brain data. | Least-VC-touched frontier. FROs are explicitly designed to spin out for-profits later; being inside the network is the early move. | Get on their radar; track which projects are heading toward a for-profit spinout. | Watch+track |
| MIT LENR group (Metzler / Hagelstein) | MIT + the ARPA-E LENR capability cohort (condensed-matter nuclear). | The “actively dismissed, quietly government-funded” reverse-signal — but the highest crank risk. A relationship costs nothing; a check needs a trigger. | Relationship + a pre-committed small check that fires ONLY on a clean, independently replicated excess-heat result. | Watch+trigger |
| Twente (van der Wiel) · CNRS (Grollier) · Tokyo (Nakajima) | The academic core of physical / in-materia / coupled-learning computing. | The startup edge (Extropic, Normal) is funded; the unpriced alpha is a fresh spinout from these labs. | Track for a formation-stage spinout; build the relationship before the round exists. | Watch+track |
Companies to have an investment conversation with (not researchers)
| Company | What | Why now | Mode |
|---|---|---|---|
| QZabre (Zurich) | NV-diamond quantum-sensing magnetometers — the “shovel” sold into every altermagnet / spintronics lab. Bootstrapped, revenue, $0 VC. | Real revenue and a real moat with no VC on the cap table. Foreign, so a negotiated minority or secondary, not a hot round. | Invest convo |
| Morphoceuticals (Medford MA) | Ion-channel regeneration drugs — the one bioelectric vehicle chasing mammalian translation. Chair: Roger Pomerantz (ex-Seres CEO). | Last raised a $3.5M seed (Sept 2024); a syndicate slice on the next round buys the mammalian-drug shot. Optionality, not conviction. | Invest convo |
| Noumenal Labs (US) | Active-inference robotics autonomy with calibrated uncertainty. ~6 people, founded 2024. | The one clean pure-play in active inference — but confirm what has actually closed (Harmonic shows $0), and treat it as a small lottery ticket. | Invest convo |
How to reach out
Two modes. With researchers, lead with the work and never with money — they get pitched constantly and dismiss anything that reads transactional. With companies, be straight that Daxos is a fund and you’re opening a conversation. Short, specific, human. No corporate voice.
Hi [name], I’ve been following [the specific paper/repo/result — name it exactly]. [One genuine, specific sentence about why it matters / a real question about it.] I work with Daxos Capital, a small fund that likes to know the people doing the most interesting technical work early — long before there’s ever a company. No agenda here beyond that. Would you be open to a short call sometime? Happy to be useful where I can (intros, feedback, a sounding board).
Goal: a call. Do NOT mention checks, valuations, or investing on first contact. If they raise it, say Daxos writes $100–500K first checks and would love the first look if they ever start something.
Hi [name], I’m with Daxos Capital ($100–500K first checks, US-focused deep tech). We think [one specific, correct sentence about what they do and why it’s differentiated — prove you understand it]. Are you raising, or would you be open to a conversation regardless? We move fast and do real diligence.
Goal: a call + their funding status. Be honest we’re a fund.
Track it
One simple sheet. Columns: Target · Scene · Mode (relationship / invest / watch-trigger) · Contact channel · Last touch date · Status · Next step · Trigger (for watch items) · Notes. Cadence: initial outreach, then a light, genuine touch every 4–8 weeks (share a relevant paper, a useful intro — not “just checking in”). The only metrics that matter: relationships opened, first-look commitments secured, and triggers watched. Deals will come later or not at all; that’s expected.
• Never promise money, a valuation, or a decision — only Mark decides that. • Never imply Daxos has committed to anything. • With researchers, curiosity first, always; if it feels like a pitch, you’ve lost them. • Get funding facts from Mark/the team (we use Harmonic), never quote a number you found online. • Log every contact in the sheet the day it happens. • When something looks genuinely hot (a spinout forming, a company raising), flag Mark immediately — don’t sit on it.