Every technological revolution has been built upon a material that quietly reshaped civilisation.
Stone enabled permanence. Steel transformed cities into industrial powers. Silicon, perhaps more than any material before it, became the invisible foundation of the modern economy. It gave rise to personal computing, smartphones, cloud infrastructure and, ultimately, artificial intelligence. For more than half a century, silicon has not simply powered innovation—it has defined its pace.
But history has a habit of replacing the materials that once seemed irreplaceable.
Artificial intelligence is exposing the limits of an industry that has spent decades relying on increasingly smaller transistors to produce increasingly larger gains. The challenge facing computing today is no longer just about making chips faster. It is about sustaining an exponential demand for computation without an equally exponential increase in energy consumption, heat generation and manufacturing complexity.
That question has prompted researchers around the world to ask a more fundamental one: what if the next era of computing is built on something other than silicon?
One company convinced the answer lies in light is Taiwan’s LongServing Technology.
Founded by Dr. Ko-Cheng Fang, the company has spent recent years developing photonic computing technologies that replace electrical signals with photons—the particles of light—to move and process information. After unveiling its photonic quantum chip architecture, the company has now introduced what it calls X-Photon, a proprietary optical material designed to guide light through integrated circuits while overcoming one of photonics’ most persistent engineering obstacles.
Whether X-Photon becomes a commercial breakthrough remains to be seen. Yet it illustrates a broader shift taking place across the global semiconductor industry: the search for the computing architecture that succeeds silicon.
Dr. Ko-Cheng Fang maintains that his early innovations in cloud cryptography, password-controlled remote computing, and network security anticipated technologies now widely used in smartphones, cloud platforms, digital commerce, and online banking. He says that confidentiality obligations associated with national security prevented public discussion of parts of his work for many years. Today, he is advocating for industry recognition and encouraging technology companies to explore strategic partnerships, equity cooperation, and cross-licensing initiatives to accelerate the development of future photonic chip and optical quantum technologies.
Silicon’s Success Has Become Its Constraint
For decades, Moore’s Law functioned as both prediction and business model. Every new generation of semiconductor manufacturing produced smaller transistors, delivering more performance at lower cost and with greater energy efficiency.
That cycle is becoming increasingly difficult to maintain.
As fabrication approaches atomic dimensions, engineers face mounting physical limitations. Heat becomes harder to dissipate. Electrical interference becomes more pronounced. Manufacturing grows exponentially more complex and expensive.
These are not merely technical problems. They are economic ones.
Artificial intelligence now consumes computational resources at a scale few anticipated only a decade ago. Training frontier AI models requires enormous data centres whose energy demands increasingly resemble those of national infrastructure projects. As governments compete to build AI capacity, electricity has become almost as strategic as silicon itself.
The industry’s next breakthrough therefore may depend less on making electrons move more efficiently and more on finding an entirely different way to move information.
Why Light Matters
Photonic computing has long occupied an intriguing position within advanced computing research.
Instead of transmitting information through electrical current, photonic systems use light.
The theoretical advantages are considerable. Photons travel faster than electrons, generate significantly less heat and allow information to move with remarkable efficiency. For AI systems that perform trillions of mathematical operations every second, those characteristics are especially attractive.
The difficulty has never been understanding the physics.
It has been engineering practical hardware.
Unlike electricity, which can be directed through conductive pathways with relative ease, light naturally travels in straight lines. Persuading photons to navigate microscopic circuits, change direction and maintain signal integrity has remained one of integrated photonics’ greatest engineering challenges.
This is the problem LongServing Technology believes X-Photon helps address.
According to the company, the material enables photons to make controlled 90-degree directional changes within nanoscale optical circuits using an embedded reflective structure. Dr. Fang compares the mechanism to the familiar behaviour of a mirror, except the reflection occurs inside the material itself rather than across an external surface.
If that capability proves scalable, it would solve one of the practical problems preventing photonic circuits from becoming substantially more complex.
Building Computing at the Scale of Atoms
LongServing’s ambitions extend well beyond a single material.
The company says X-Photon operates at optical wavelengths averaging between two and three nanometres while supporting optical circuitry fabricated at the 10-nanometre scale. Those dimensions are significant because photonic computing has historically struggled with miniaturisation compared with electronic semiconductors.
Shrinking optical components sufficiently to compete with silicon has remained one of the field’s central engineering challenges.
LongServing argues that overcoming this hurdle could make densely integrated photonic processors and memory systems commercially viable rather than purely experimental.
The distinction matters.
Many technologies demonstrate impressive laboratory performance without ever becoming manufacturable products. The future of photonics will ultimately depend as much on fabrication economics as scientific achievement.
Beyond Faster Chips
The implications extend beyond processor performance.
If photonic architectures mature, they could reshape the economics of AI infrastructure itself.
Today’s AI boom has triggered an unprecedented global race to construct data centres capable of training increasingly sophisticated models. Those facilities require vast amounts of electricity, specialised cooling systems and continuous investment in increasingly expensive semiconductor hardware.
Photonic systems promise a different equation.
Because light generates considerably less heat than electrical current, future optical computing platforms could reduce both power consumption and cooling requirements while delivering higher computational throughput.
LongServing has outlined an ambitious long-term vision that includes two-nanometre multi-bit photonic quantum chips, optical memory technologies and large-scale Photonic Cloud Computing Centres designed specifically for AI workloads.
The company has also suggested future platforms could eventually achieve computational performance many times greater than conventional electronic systems while dramatically reducing energy consumption. These remain aspirational commercial targets rather than demonstrated capabilities, but they reflect the scale of ambition increasingly defining next-generation computing research.
Betting on the Next Computing Economy
Scientific ambition, however, requires financial backing.
Developing an entirely new computing architecture is as much an industrial undertaking as a technological one. Success depends not only on scientific discovery but also on manufacturing ecosystems, supply chains and patient capital.
LongServing recently announced a US$500 million strategic financing initiative based on a stated valuation of US$2.5 billion. According to the company, the investment will support photonic fabrication, cloud infrastructure and the continued commercial development of its technologies.
It has also introduced what it describes as a Strategic Equity Hedging Protocol aimed at creating long-term partnerships with global technology companies as the photonic ecosystem expands.
The announcement reflects a broader reality of the AI economy.
The next generation of computing is unlikely to emerge from research laboratories alone. It will require industrial alliances, infrastructure investment and sustained confidence that alternative architectures can eventually compete with one of the most mature manufacturing industries ever built.
A Different Kind of Computing Race
For decades, the semiconductor industry has measured progress by how effectively it could manipulate electrons.
The coming decade may ask a different question altogether.
As AI continues to redefine economies, geopolitics and industrial strategy, attention is shifting from software alone to the physical foundations that make advanced computation possible. The companies shaping tomorrow’s AI may not simply build better algorithms. They may invent entirely new ways for information to move.
LongServing Technology’s work sits within that larger transition.
Whether X-Photon ultimately becomes a cornerstone of future computing or one step along a longer scientific journey, it reflects an increasingly important reality: the race to define artificial intelligence is no longer confined to models and software. It is also a race to discover the material capable of carrying the next generation of computation.
Silicon transformed the digital age.
The next chapter may depend on whether light can do the same.
Contact Information
Dr. Ko-Cheng Fang
Founder, CEO & Chairman
LongServing Technology Co., Ltd.
Email: service@longserving.com.tw
Website: https://longserving.com.tw/en/
Instagram: @ko_cheng_fang
