Key Takeaways
- • CAS Star, a venture capital firm spun from China's state-run Academy of Sciences, has validated a decade-long photonics investment thesis as AI infrastructure demand accelerates, with over 200 of its ~600 portfolio companies operating in photonics-related sectors (sensing, communications, computing, storage, display).
- • Photonics—glass chips that transmit data at the speed of light—has emerged as critical AI infrastructure as data centers scale to unprecedented sizes; tens of thousands of specialized computing chips now require high-speed interconnectivity beyond traditional copper cable capacity.
- • Major institutional investors including Nvidia have deployed $4 billion in photonics bets, investing $2 billion each in companies Lumentum Holdings and Coherent, signaling convergence around photonics as the next AI infrastructure investment frontier.
- • CAS Star founder Mi Lei, a doctorate-holder in optics from CAS's Xi'an Institute of Optics and Precision Mechanics, positioned the firm as a bridge between scientific research commercialization and industrial upgrading, leveraging deep technical expertise in a strategically important sector.
- • Photonics investment has accelerated globally since March 2026, with Intel Capital investing $91 million in North Ocean Photonics and multiple photonics-focused ETFs emerging, reflecting Wall Street's recognition of photonics as the next major AI infrastructure opportunity.
Summary
CAS Star, a venture capital firm established from China's state-run Chinese Academy of Sciences, is capitalizing on a decade-long strategic thesis in photonics that has suddenly become central to AI infrastructure development. Founder Mi Lei, who holds a doctorate in optics from CAS's Xi'an Institute of Optics and Precision Mechanics, has consistently funded photonics companies since the firm's inception. Today, the portfolio spans over 200 companies across the photonics ecosystem—covering sensing, communications, computing, storage, and display applications—representing a significant portion of CAS Star's roughly 600 total portfolio companies.
The timing of photonics' ascendance reflects fundamental infrastructure constraints facing the AI industry. As artificial intelligence models scale to unprecedented sizes, training operations require tens of thousands of specialized computing chips to constantly exchange massive volumes of data. Traditional copper cables and wire infrastructure cannot sustain the bandwidth and speed requirements at this scale. Photonics—glass chips that transmit data at the speed of light—offers a solution that aligns with the physical limits of the universe itself, making it an increasingly indispensable component of next-generation data center architecture.
CAS Star's early positioning in photonics has proven prescient, with major institutional capital now validating the sector's strategic importance. Nvidia has deployed $4 billion in photonics investments, committing $2 billion each to companies Lumentum Holdings and Coherent, reflecting the technology giant's confidence in photonics as essential AI infrastructure. Since March 2026, photonics investment has accelerated broadly, with Intel Capital allocating $91 million to North Ocean Photonics and multiple photonics-focused exchange-traded funds launching to capture institutional investor demand.
Mi Lei frames CAS Star's role as catalyzing the transition from scientific research to commercial application—a strategic imperative he describes as "push[ing] scientific research towards commercialisation" in service of industrial upgrading. This positioning leverages CAS's deep research heritage while operating as a commercial investment vehicle, creating a distinctive advantage in identifying and nurturing early-stage photonics technologies before broader market recognition. The convergence of AI infrastructure demands with a decade of accumulated portfolio construction positions CAS Star as a significant beneficiary of photonics' emergence as the next critical AI infrastructure investment frontier.
The timing of photonics' ascendance reflects fundamental infrastructure constraints facing the AI industry. As artificial intelligence models scale to unprecedented sizes, training operations require tens of thousands of specialized computing chips to constantly exchange massive volumes of data. Traditional copper cables and wire infrastructure cannot sustain the bandwidth and speed requirements at this scale. Photonics—glass chips that transmit data at the speed of light—offers a solution that aligns with the physical limits of the universe itself, making it an increasingly indispensable component of next-generation data center architecture.
CAS Star's early positioning in photonics has proven prescient, with major institutional capital now validating the sector's strategic importance. Nvidia has deployed $4 billion in photonics investments, committing $2 billion each to companies Lumentum Holdings and Coherent, reflecting the technology giant's confidence in photonics as essential AI infrastructure. Since March 2026, photonics investment has accelerated broadly, with Intel Capital allocating $91 million to North Ocean Photonics and multiple photonics-focused exchange-traded funds launching to capture institutional investor demand.
Mi Lei frames CAS Star's role as catalyzing the transition from scientific research to commercial application—a strategic imperative he describes as "push[ing] scientific research towards commercialisation" in service of industrial upgrading. This positioning leverages CAS's deep research heritage while operating as a commercial investment vehicle, creating a distinctive advantage in identifying and nurturing early-stage photonics technologies before broader market recognition. The convergence of AI infrastructure demands with a decade of accumulated portfolio construction positions CAS Star as a significant beneficiary of photonics' emergence as the next critical AI infrastructure investment frontier.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.