New Data Centers Will Be Powered by Human Brain Cells

Cortical Labs is building two "biological data centers" with DayOne Data Centers: 120 CL1 units in Melbourne and up to 1,000 units in Singapore, deploying racks of living-neuron CL1 "wetware" compute. CEO Hon Weng Chong says each CL1 node uses less power than a handheld calculator—orders of magnitude below modern GPUs—suggesting materially lower energy and water footprints if scaled. The company has demonstrated gameplay and prototypes but has not shown CL1 can match the computational performance of current top-tier data-center chips, leaving significant execution and technical risk.

Analysis

This development is best viewed as a long-duration technology option rather than an immediate structural shock to silicon incumbents. If biological compute captures even a low-single-digit share of AI inferencing workloads over 3–7 years, it would convert a portion of GPU demand growth into a different capex bucket (biomanufacturing + lab automation) and compress long-term GPU pricing power; conversely, failure to scale reproducibly would relegate it to an R&D curiosity with negligible market impact.

Second-order winners are likely to be suppliers of cell-culture infrastructure, sterile automation, and high-throughput biomanufacturing (equipment, reagents, facility conversion), while traditional colocation economics could bifurcate — some operators will see lower power/water intensity per rack and others will need to invest in wet-lab compliance to capture new revenue. Local utilities and water-dependent cooling suppliers face asymmetric downside in growth forecasts where these systems scale, which could change municipal revenue models and data-center siting decisions within a 2–4 year planning horizon.

Key risks: reproducibility, regulatory/ethical constraints, and biological supply chain scale (cell sourcing, contamination control) — any of these can stall commercial rollout for multiple years. For investors, the prudent posture is to treat this as a convex theme: size exposure in supply-chain names with clear revenue pathways while using cheap, time-limited hedges against incumbent GPU upside being re-priced too quickly.

Key Decisions for Investors

• Buy Thermo Fisher (TMO) or Danaher (DHR) equal-weight exposure (initiating a 12–24 month overweight) to capture incremental revenue from lab automation and biomanufacturing conversions; target 15–30% upside if wetware adoption materializes, primary risk is biotech capex cyclicality—trim on a 25% run-up.
• Protect NVDA exposure with a 6–12 month put spread (buy ~15% OTM put, sell ~30% OTM put) sized to cover 10–20% portfolio exposure—cost financed by selling a small portion of upside (call spread). This limits downside from an accelerated market re-rating while capping the hedge cost if GPU dominance persists.
• Initiate a 12-month pair: long Digital Realty (DLR) / short NVDA (dollar-neutral). Rationale: DLR can retool real estate and capture conversion premium; short NVDA expresses the risk of demand reallocation. Target asymmetric pay-off: 20–30% if wetware narrative drives capex shift, with stop-loss if NVDA total-return outperforms by +15% in 3 months.
• Allocate a small (5% thematic sleeve) to lab-specialist small caps via call spreads (9–18 month) on names with >50% revenue in bioprocess equipment; high idiosyncratic risk but offer 3x+ upside if adoption accelerates. Keep exposure size small and monitor reproducibility/outsourcing headlines as primary catalysts.