US Expert: China's All-Optical Photonic Chip Route Is Highly Promising

At the 2026 Phoenix Bay Area Finance Forum held in Guangzhou on September 16, Joe Weinman, president of the US Institute for Future Industries and a pioneer in AI economics, said in his opening speech that China's deep investments in AI's underlying technology architecture are underestimated, and that all-optical photonic chips represent a highly promising route.

Weinman noted that China is conducting extensive basic research in multiple AI infrastructure fields with transformative potential, citing Tsinghua University's Taiji-1 and Taiji-2 all-optical photonic chips as one of the most representative achievements. In his view, this technology route directly addresses a core pain point of current AI computing systems.

Mainstream AI computing today relies heavily on silicon-based electronic chips, which generate significant heat during operation. High power consumption and bulky cooling systems have become hard constraints on the expansion of large-scale computing centers. Photonic chips, by contrast, offer disruptive advantages: they run without generating heat, substantially reducing power consumption and cooling requirements. Optical computations can also be performed in parallel, delivering a fundamental efficiency improvement over electronic circuits.

Beyond photonic chips, Weinman highlighted China's progress in quantum computing. Continued breakthroughs in quantum error correction and quantum photonic interconnection are helping quantum computers scale efficiently and move toward commercial deployment, already creating real business value in several industrial scenarios. Research in even more advanced directions, such as room-temperature quantum computing, is also gradually reducing dependence on ultra-low-temperature dilution refrigerators and scarce helium-3 resources.

Weinman argued that the global AI industry's current heavy reliance on GPU-centric technology paths may not withstand long-term scrutiny. China's sustained investment in underlying architectures such as photonic chips and quantum computing is exploring more diverse technological possibilities for the AI industry and building long-term momentum for the next round of global AI breakthroughs.