Investment Thesis
Betting on the physical layer before the market is obvious.
Coexin was founded on a single insight: the hardware layer of computation is being rebuilt from the ground up — and the best time to invest is when the physics is proven but the commercial path is opaque to generalist investors.
Why we focus on deep silicon and quantum
The dominant compute paradigm — general-purpose CPUs scaling on Moore's Law — is reaching physical limits. What comes next is not one technology but several competing physical substrates: reconfigurable dataflow architecture, wafer-scale integration, neutral-atom qubit registers, superconducting processors, photonic interconnects. Each requires a different kind of diligence. You cannot evaluate a neutral-atom qubit startup from a spreadsheet. You need to understand the fidelity numbers, the gate times, the error correction overhead. That's not a generalist skill. It's why Coexin exists.
Our investment criteria
We lead or co-lead Series A rounds in deep-tech hardware companies where the underlying physics or materials science is novel. We look for: (1) a founding team that understands the physics deeply enough to predict which engineering problems will turn out to be hard versus easy; (2) a defensible substrate advantage — not just a product roadmap, but a reason why this company's approach is harder to replicate than it appears; (3) a market where the value capture is at the device or system level, not easily commoditized to software. We are not the right partner for software-only AI plays or consumer hardware.
What we mean by conviction capital
Every Coexin check has been made at a stage when the company's core claims were unproven in the market — before Lightmatter had a shipping product, before PsiQuantum had announced its billion-dollar scale-up, before Cerebras had placed its first chip in a data center. We don't require market proof. We require scientific plausibility and a team that can close the gap between proof-of-concept and manufacturable product. That's a judgment call that depends on our partners' direct technical experience, not on competitive-landscape analysis.
Stage and check size
We invest primarily at Series A — typically $17M–$38M check size, leading or co-leading. In rare cases where a pre-Series-A syndicate needs a technically sophisticated institutional anchor, we will participate earlier. Our two funds reflect our evolution: Coexin Fund I ($50M, closed May 2017, fully deployed) proved the thesis at smaller scale; Coexin Fund II ($130M, closed August 2022, currently deploying) allows us to write larger checks and reserve more capital for follow-on in existing portfolio companies.