Analysis
PicoJool, a maker of optical connectivity components for AI infrastructure, raised a Series A led by Socratic Partners, with participation from Hudson River Trading, according to the company's announcement:
- Series A -- $27.5M, led by Socratic Partners
- Prior seed round -- $12M, led by Playground Global
- Total funding to date -- $39.5M
PicoJool builds Vertical Cavity Surface Emitting Laser (VCSEL) and microVCSEL products -- the optical components that move data at high speed between processors inside AI data centers, the same category Ligent Technologies monetizes at a far larger scale through its recent Hong Kong IPO, covered elsewhere in this issue. Where Ligent is an established, revenue-generating manufacturer going public, PicoJool is an early-stage company still scaling its first commercial products, following the June unveiling of its 200G VCSEL and microVCSEL family for scale-up AI data centers.
The company will use the new capital to expand its teams and manufacturing facilities in both the United States and Taiwan -- a dual-geography build-out that spreads supply-chain risk across two jurisdictions rather than concentrating optical-component manufacturing in one country, relevant given how much attention chip and component supply chains have drawn amid ongoing US-China trade tension this year.
Hudson River Trading's participation is a detail worth noting on its own: a quantitative trading firm investing directly in a physical AI-hardware component maker, rather than only deploying capital into software, data or trading-adjacent infrastructure, suggests firms with deep computing expertise are looking further down the AI infrastructure stack for investment opportunities than their traditional software focus would suggest.
At $27.5 million, this round is a fraction of the size of the mega-deals covered elsewhere in this issue, but it's a useful data point on its own: early-stage venture capital is still funding the unglamorous physical layers underneath the AI buildout -- optical interconnects, in PicoJool's case -- not only the foundation-model labs and application-layer startups that dominate headlines.