Japan Bets on Light to Win the Next Tech War

The lesson that haunts Japan’s tech strategy

Japan lost the mobile phone war, and the memory of that loss still shapes how the country approaches technological competition. Its so-called garake — feature phones that dominated the domestic market and held significant global presence in the late 1990s and early 2000s — became irrelevant almost overnight when smartphones arrived. Sony, Sharp, Fujitsu, and Panasonic watched their market shares evaporate not because they lacked engineering talent or manufacturing excellence, but because they misread the architectural direction of the industry. They optimized existing paradigms instead of betting on the next one.

The pattern has repeated across sectors. Japan ceded ground in flat-panel displays to South Korea and Taiwan, lost the memory chip business to Korean producers in the 1990s, and watched its automotive dominance face existential pressure from electric vehicles. Each time, the narrative was the same: brilliant incrementalism, catastrophic strategic blindness. The current government, led by Prime Minister Fumio Kishida and continuing under his successors, is haunted by this history. It is building an entire industrial strategy around avoiding that mistake again — by identifying where the architectural shift is happening and placing bets before the tide turns.

The bet on light

Japan has identified optical communication devices as one of seventeen priority fields in its growth strategy, a category selected not by accident but by design. These seventeen sectors — ranging from quantum technology and AI to biomanufacturing and next-generation batteries — were chosen through a rigorous assessment of where Japan holds existing competitive advantages and where global demand is about to surge. Optical communication checks both boxes emphatically.

The target is blunt: double the country’s global market share to 10 percent by 2030. The total addressable market is projected to reach 53 trillion yen — roughly $350 billion — by that date, driven primarily by the bandwidth requirements of AI data centers, 6G research, and the expanding submarine cable networks that connect global compute infrastructure. The focus areas are specific and interconnected. An all-optical network would eliminate the electrical conversion bottlenecks that currently limit data center performance. Submarine cable leadership would secure Japan’s position in the physical layer of global connectivity. Next-generation wireless would embed optical principles into the radio access networks of tomorrow.

These are not consumer gadgets. They are the bones of the information infrastructure that AI and everything built on top of it will run through. Self-defense forces information systems are already being planned around IOWN — the Innovative Optical and Wireless Network initiative — which signals that this is not only an economic play. It is an economic security bet, one that treats information infrastructure with the same seriousness that Cold War planners once treated nuclear weapons. The timing matters: AI workloads are creating bandwidth demand that electrical interconnects simply cannot satisfy at scale. Power consumption in data centers is already approaching critical thresholds, and the electrical-to-optical conversion process that underpins current networks is both energy-intensive and fundamentally limiting performance.

The architecture question

Here is where the strategy gets interesting, and where Japan could repeat its garake error if it picks the wrong technical path.

IOWN was originally conceived around APN — the All-Photonic Network — a vision in which data stays in the optical domain from end to end. No electrical conversion. Minimal latency. Massive throughput. NTT, Japan’s former monopoly turned private giant, was the natural champion of this approach and has genuine expertise that no competitor can replicate overnight. The company inherited decades of research from its telecommunications monopoly days, when it built the fiber optic backbone that connected Japan and much of the world. NTT Labs holds foundational patents in photonic integrated circuits, wavelength division multiplexing, and optical switching that remain essential to any all-optical architecture.

But recent NTT presentations suggest a pivot that could redefine the entire strategy. The company appears to be shifting IOWN’s emphasis from pure APN toward a hybrid optical-electric architecture — one that retains electrical processing stages alongside optical transport. That sounds incremental. It is not.

Hybrid optical-electric solutions are exactly what NVIDIA, Broadcom, and TSMC are productizing right now, and they represent a different value proposition entirely. NVIDIA’s optical interconnect strategies, for instance, are designed to reduce power consumption within existing data center architectures — a real and urgent need, but one that leaves the fundamental electrical processing paradigm intact. Broadcom and TSMC are pursuing similar paths with silicon photonics that augment rather than replace electronic computation. These hybrid approaches optimize the existing paradigm. APN rewrites it. The difference is not a matter of degree but of kind: APN promises fundamental latency reduction, a completely different performance envelope, and a supply chain that radiates from Japanese strengths rather than American ones.

If Japan follows the hybrid route, it enters a market where American and Taiwanese firms already have first-mover advantage, manufacturing scale, and deep customer relationships with every major hyperscaler. If Japan doubles down on APN, it competes on its own terms — where it actually leads. The choice is not merely technical. It is the definition of what kind of technological leader Japan intends to be.

Who wins, who loses

The APN path rewards Japan in ways that the hybrid path does not. Its firms have decades of fiber optics expertise compounding across generations of research and development. Fujikura dominates the global market for optical fiber splicing equipment. Sumitomo Electric produces the specialty fibers that underpin long-haul submarine cables. NTT Laboratories holds patents in photonic integration that predate most of the current AI boom. JAXA’s experience with optical laser communications for satellite networks adds another dimension to Japan’s capabilities. An all-optical network is the kind of infrastructure where these assets compound — each generation of optical component builds on the last, and Japan’s accumulated knowledge creates barriers to entry that are difficult for newcomers to overcome.

The hybrid path rewards NVIDIA, Broadcom, and TSMC with a different kind of compounding. These companies are already shipping silicon photonics products and advanced packaging solutions at scale. Their customers are the hyperscalers — Amazon, Google, Microsoft, Meta — building AI clusters today, not tomorrow. Japan would be a follower in this world, assembling components designed elsewhere and sold through channels it does not control. The value capture would be minimal.

The domestic political calculus is not simple. Hybrid approaches may deliver faster near-term results and appeal to incumbents with existing electrical interconnect businesses who would see their positions weakened by full APN adoption. APN requires more patience, more willingness to disrupt established architectures, and more confidence that the long-term payoff justifies the short-term disruption. That is precisely the same tension that destroyed Japan’s mobile industry — the tendency to optimize the present rather than bet on the next thing. Industry groups with vested interests in the electrical paradigm will lobby hard for hybrid solutions. The question is whether the government’s strategic vision can override those pressures.

The supply chain angle

What makes this strategically significant beyond Japan’s borders is the supply chain dimension, which sits at the intersection of three critical areas: semiconductors, telecommunications infrastructure, and data center economics. A Japanese-led APN ecosystem would create sustained demand for specialized photonic chips, precision interconnects, and advanced packaging — areas where Japan has deep capability but limited manufacturing scale. That gap represents an opportunity to rebuild domestic manufacturing in segments that Western companies have largely offshored over the past three decades. Japan already produces the raw materials — specialty glass, lithium niobate substrates, indium phosphide wafers — that photonic chip manufacturing requires. The missing piece is the Fab-scale production capacity.

This dynamic creates second-order effects that extend well beyond Japan’s economy. A Japanese-led APN supply chain would give countries looking to diversify away from US-dominated tech stacks an alternative supplier base. It would reduce the concentration risk that currently makes the global information infrastructure vulnerable to geopolitical disruption. For ASEAN nations, whose submarine cable connections pass through Japanese territory, a Japanese APN ecosystem would offer a third option between American and Chinese infrastructure providers — a scenario that aligns with the strategic preferences of Japan’s closest partners in the Indo-Pacific.

Submarine cable strategy reinforces this entirely. Japan’s cable network connects it to major data centers across the Pacific and through Southeast Asia, giving it leverage at the physical layer of global connectivity. Controlling the optical terminals and repeaters on those cables is leverage in any scenario where data sovereignty matters — which is increasingly often. As nations impose data localization requirements and restrict cross-border data flows, the physical routes that data takes become matters of national security. Japan’s position as a cable hub is not accidental. It is strategic.

What happens next

The window is real but narrow. The 53 trillion yen market projection assumes continued AI-driven bandwidth growth that shows no signs of plateauing. But if that growth slows — if hybrid solutions prove good enough for most use cases and the market settles into an equilibrium that does not require full all-optical architectures — the APN thesis weakens considerably. The government’s timeline of 2030 gives roughly four years to establish a differentiated position, and those four years will determine whether Japan becomes a shaper of the next infrastructure era or a supplier of components designed elsewhere.

NTT’s direction will be the signal everyone watches. The company’s internal debates are being observed closely by industry groups, foreign governments, and competing corporations. If NTT returns to APN as the centerpiece of IOWN, the industrial base will follow — suppliers will invest, universities will adjust curricula, and the government’s subsidies will flow toward the all-optical architecture. If hybrid wins internally, expect procurement decisions to tilt toward established American and Taiwanese suppliers, and Japan’s global market share to stagnate rather than double. The company’s next major IOWN demonstration project will be the clearest indicator available.

The garake lesson was supposed to teach Japan to bet on architecture, not just components. It was supposed to teach the country to recognize when the underlying paradigm is shifting and to commit resources before the shift becomes irreversible. Optical communication is the test. The technology exists — APN is not science fiction, and NTT and its partners have prototypes that demonstrate feasibility. The market is growing at a rate that rewards early movers. The question is whether Japanese industry and government can commit to the harder, longer play — or whether they will chase the easier win and find themselves dependent again.

How Japan answers that question will matter for the rest of the world’s infrastructure supply chains. A successful APN bet would create an alternative technology stack centered on Japanese manufacturing, weakening the concentration of infrastructure control that currently rests with American and Taiwanese firms. A failed bet — a retreat into hybrid solutions — would reinforce that concentration and leave Japan as a secondary player in the infrastructure that underpins the digital economy. The garake taught Japan that architectural bets define national prosperity. The optical pivot is whether the country has learned the lesson.