Two neurosurgeons just raised $25 million betting brain cells can (someday) outcompute silicon

The Biological Computing Co. raised $25 million in a seed round led by Primary with participation from Builders VC, E1 Ventures, Proximity, Refactor Capital, Tusk Ventures, and Wonder Ventures to commercialize neuron-based biological computing. Founders Alex Ksendzovsky and Jon Pomeraniec demonstrated that living neurons can be stimulated with encoded market and image data to produce usable computational outputs and plan near-term products such as an algorithm-discovery platform and software adapters; they position the technology as a far more energy-efficient alternative to conventional silicon for certain AI tasks, with five- to ten-year ambitions to integrate biological networks into real-time compute circuits.

Analysis

Market structure: Biological compute at seed-stage shifts near-term winners to lab-equipment, reagents, and automation suppliers (Thermo Fisher TMO, Danaher DHR, Illumina ILMN) and to niche startups building multi-electrode arrays and cell-culture scale-up; incumbent GPU winners (NVDA) retain pricing power for 1–3 years as mainstream AI demand continues. Adoption curve is long — expect commercial substitution pressure on silicon only in a 5–15 year window; initially demand for specialty consumables will spike, tightening reagent and microfluidics supply chains and lifting small-cap tooling vendors.

Risk assessment: Key tail risks are regulatory/ethical clampdowns (moratoria on certain neuron uses), biosafety incidents, and reproducibility failures — each could halt adoption for 12+ months; operational scaling risk makes probability of widespread data-center deployment within 5 years <25%. Hidden dependencies include IP bottlenecks (neuronal encoding patents) and hyperscaler partnership access; primary catalysts are published benchmarks showing latency <100ms and accuracy parity with silicon or a hyperscaler pilot within 12–24 months.

Trade implications: Near-term (0–12 months) overweight lab-tech suppliers and select synthetic-bio tooling names (TMO, DHR, ILMN) by 2–4% AUM; keep NVDA exposure but hedge longer-term disruption via 12–24 month OTM put spreads (NVDA). Rotate out of data-center REITs/utility-heavy capex plays (e.g., reduce EQIX exposure by 1–2%) because long-term biological compute roadmaps imply downside to perpetual data-center growth assumptions.

Contrarian: The market will likely short NVDA on “bio compute threat” headlines — that is premature. Historical parallels (GPU and ASIC cycles) show computation paradigm shifts take decades; mispricings exist in small-cap lab automation equities and private tooling rounds where revenues are growing but multiples compressing. Main unintended consequence: heightened regulatory scrutiny could create winners among well-regulated, vertically integrated instrument providers and losers among bench-scale startups lacking compliance infrastructure.

Key Decisions for Investors

• Establish a 2–3% long position in Thermo Fisher (TMO) or Danaher (DHR) within 2–6 weeks to capture increased capex for cell-culture, electrode arrays, and automation; target +15–25% upside on 12–24 month revenue re-rating, trim if revenue growth stalls below 8% YoY.
• Maintain a 1–2% core long in NVDA for near-term AI demand but buy a protective 12–18 month put spread (buy 15–20% OTM puts, sell 30–40% OTM puts) sized to 0.5–1% of portfolio to hedge a >25% downside event tied to faster-than-expected bio compute adoption.
• Reduce data-center REIT exposure (e.g., EQIX) by 1–2% of portfolio over the next 3 months; redeploy into lab/biotech tooling names—exit reduction if EQIX total-return outperforms NVDA by >10% over 6 months (signals persistent demand).
• Allocate 0.5–1% into public small-cap lab-automation or microfluidics names (or an equivalent private deal) with active due diligence, and set a binary exit trigger: sell if company fails to announce a commercial hyperscaler partnership or peer-reviewed reproducible benchmark within 12 months.
• Monitor regulatory milestones and peer-reviewed demos closely over the next 6–12 months; if a reproducible real-time neuronal compute demo appears with latency <100ms and benchmark parity, increase biotech tooling exposure by another 2–3% and reduce NVDA hedge proportionally.