Triatomic Q1 2026 - Private II
The Q1 2026 investor update from Triatomic Capital outlines the firm’s view of the current AI landscape, macro‑economic implications, and a deep‑dive into an emerging investment theme—optical interconnects for next‑gen datacenter compute. The letter notes that AI capability scaling is now constrained by multiple supply‑chain bottlenecks (compute, memory, power, permitting, talent, and even niche inputs like helium and chip‑bonding films). These constraints are creating valuation dislocations across public and private markets. Triatomic frames AI as this decade’s analogue to mid‑20th‑century offshoring, arguing that while AI will accelerate productivity, the labor impact will mirror past manufacturing shifts—gradual, uneven, and ultimately leading to re‑skilling and new demand for high‑touch services. The firm presents data on AI adoption curves, compares them to smartphones, internet, and offshoring, and predicts a slower transition from pilot to full‑scale deployment. A second major focus is the inflection point in optical networking. Triatomic details how per‑rack optical bill‑of‑materials (BoM) costs are projected to rise 20× from the H100 generation (2023) to Rubin Ultra (2028), driven by co‑packaged optics (CPO) and scale‑up domain requirements. The letter explains the technical advantages of optics over copper, charts the historical wave of optical displacement, and argues that AI workloads have pushed the industry past the electrical I/O bandwidth limit, making optics essential for clusters >100 accelerators. The firm highlights supply‑chain concentration in China and the opportunity for Western start‑ups to on‑shore. Fund‑specific updates: Triatomic Capital Private II LP is ~80% funded toward a $350 M target, with two closes and two investments made—core investment in Radical Numerics (engineered biology/automated economy) and an opportunistic investment pending. Deal flow numbers (195 opportunities, 65 meetings, 34 diligences) are provided. Contact information for final fundraising is included. The letter ends with standard legal disclosures.
Topics
AI Supply‑Chain Bottlenecks and Market Dislocations
Analysis of compute, memory, power, permitting, talent, and niche material constraints throttling AI scaling and causing valuation gaps.
AI as the Decade’s Globalization Analogue
Comparison of AI‑driven productivity gains to mid‑20th‑century offshoring, with emphasis on labor displacement, re‑skilling, and demand shifts toward high‑touch services.
AI Adoption Curve vs. Historical Technology Diffusion
Logistic S‑curve data contrasting generative AI adoption with smartphones, internet, and offshoring, highlighting the pilot‑to‑full‑scale gap.
Optical Interconnect Inflection in Next‑Gen Compute
Detailed thesis on co‑packaged optics, per‑rack optical BoM growth (20× from H100 to Rubin Ultra), and the technical drivers (bandwidth, power, latency).
Supply‑Chain Concentration and On‑shoring Opportunities
Analysis of Chinese dominance in optical components and the strategic opening for Western start‑ups to supply domestic datacenter markets.
Fund II Capitalization, Deal Flow, and Portfolio Highlights
Status of Triatomic Capital Private II LP fundraising (≈80% of $350 M target), recent investments (Radical Numerics), and pipeline metrics.