Chinese VC's Decade-Long Photonics Bet Pays Off as AI Strains Data Centers

Mi Lei sits in his Shanghai office beneath a nameplate that reads "Wallfacer"—a reference to Liu Cixin's science fiction trilogy, where elite strategists serve as humanity's last defense against an incomprehensible threat. The irony is not lost on him. For more than a decade, he has been betting venture capital on a technology most of the world was not yet ready to take seriously. Now, as artificial intelligence reshapes the global computing landscape, that patient thesis is being vindicated in real time.

Mi founded CAS Star, a venture firm spun out of China's state-run Academy of Sciences, with a singular conviction: that light would eventually become the infrastructure of choice for moving data at scale. He holds a doctorate in optics from the Academy's Institute in Xian and has spent his career watching the physics of electricity bump up against hard limits. When you try to push massive amounts of data through copper wires, you hit a wall. The signal degrades. Heat builds. Power consumption climbs. The bottleneck strangles the system.

For years, this was a thesis in search of a moment. The moment has arrived. As AI models have grown to require tens of thousands of specialized chips working in concert, the demand for data transmission between those chips has exploded beyond what traditional copper interconnects can handle. Optical links—systems that transmit information using photons instead of electrons—suddenly became not a curiosity but a necessity. Light travels at the speed of light. It generates far less heat. It consumes far less power. In the calculus of modern data centers, these advantages are no longer marginal. They are existential.

CAS Star's portfolio reflects this long conviction. Of the firm's roughly 600 companies, more than 200 operate across the photonics sector: sensing, communications, computing, storage, display. The breadth suggests not a scattered bet but a systematic thesis about where computing infrastructure would need to go. One early investment, Yuanjie Technology, a laser-chipmaker based in Shaanxi province that CAS Star backed around 2019, has seen its stock price surge more than elevenfold in the past year alone. The driver is straightforward: revenue from data-center light-source products has exploded as companies racing to build and train large language models have discovered they cannot do so without optical infrastructure.

Mi describes the venture firm's role with the clarity of someone who has been waiting for this moment. "New technologies can drive industrial upgrading," he said. "Our role is to push scientific research towards commercialisation." It is a statement that sounds simple until you consider what it required: the patience to fund a sector that the market did not yet understand, the technical knowledge to recognize which companies had real solutions, and the institutional backing to sustain that bet through years when returns seemed distant. CAS Star had all three. Now the world is catching up to what Mi saw coming.

The implications ripple outward. As AI continues to scale globally, the companies that control photonics infrastructure—the light-based systems that allow data centers to function at the speeds and scales required by modern AI—will occupy a position of genuine strategic importance. China has positioned itself aggressively in this space. CAS Star's portfolio is not an accident. It is the result of a decade of deliberate, patient capital allocation in a sector that is now becoming central to how the world computes. The validation, for Mi, is not just financial. It is vindication of a vision about where technology was heading, and proof that sometimes the longest bets are the ones that matter most.