Management Strategy for an AI-Driven Society — Fujitsu CTO Takagi on the Reality of "Human-Centered AI x Corporate Transformation" [SusHi Tech Tokyo 2026]

Hello, this is Hamamoto from TIMEWELL.

"In Stanford University's AI vibrancy ranking, Japan dropped to ninth place in 2023." When Fujitsu CTO Kazuhiro Takagi projected that single slide on the screen at SusHi Tech Tokyo 2026, the room froze.

Takagi's talk that day was an extraordinarily candid "declaration of crisis" delivered by the CTO of one of Japan's leading technology companies. It was not a presentation that simply stoked anxiety. He spent an hour mapping out, in fine detail, how Fujitsu intends to convert four decades of accumulated AI research into the implementation of an "AI-driven society." Among the executive talks I have attended this year, this one had the sharpest edge.

In this article, I summarize the talk's logic from a management perspective and share concrete pathways for legacy enterprises to transform into AI natives. I hope it can serve as material for anyone in new business development or transformation roles to audit their own company's AI strategy.

I should add a quick note on why a CTO talk like this matters in 2026. For most of the past decade, the dominant AI narrative inside Japanese companies has been "let's pilot something with ChatGPT and see what happens." That mindset is reaching the end of its useful life. The conversation needs to move to platform choice, organizational structure, and capital allocation. Takagi's talk reframed AI from a tool to be experimented with into infrastructure to be designed at the same level of seriousness as ERP, security, or supply chain. That reframing alone is worth the trip to SusHi Tech.

---

SusHi Tech Has Become the Main Battleground for AI Discussion

SusHi Tech Tokyo 2026, held at Tokyo Big Sight from April 27 to 29, 2026, is one of Asia's largest global innovation conferences. This year's focus themes spanned four areas, including "AI and Robotics," "Resilience," and "Entertainment." Among them, the density of AI-related sessions was overwhelming, and the Fujitsu CTO's solo talk drew one of the largest crowds.

Takagi opened by recalling his attendance at the "AI Impact Summit 2026" held in India in February, where he joined a roundtable with national leaders including Prime Minister Modi. Witnessing India's intense national-level commitment and the energy of its young top talent, he developed a deep sense of crisis about Japan's "current position in the AI domain." Starting with this personal experience gave every subsequent message the urgency of an executive speaking as an individual.

"Japan's AI Lag" in Numbers

The numbers shown during the talk were so clear that they were painful to confront head-on.

• Japan ranked 9th out of 36 countries in the Stanford AI Vibrancy ranking (2023)
• Japan's working-age population is projected to fall 30%, from about 75 million in 2020 to roughly 53 million by 2050
• Heavy rain events have roughly doubled compared with the 1980s
• South Korea produces 12 million tons of artificial fertilizer, about twice Japan's output

Across both R&D (research output and patents) and economic indicators (startup investment and M&A), Japan is conspicuously behind. The fact that the CTO of a top company stated openly, in a public forum, that this "lag" exists as a structural issue is significant. For a long time it was taboo in Japanese management culture for major companies to admit they were "losing." That atmosphere is finally beginning to shift.

The shift is partly generational. Executives who came of age during the bubble era often defended Japan's ranking by pointing to manufacturing prowess or the quality of basic research. The current generation of CTOs and CFOs has watched too many global indices to use that defense credibly. They have also seen what happens when a country denies its position for too long — recovery takes decades, not years. Naming the problem clearly, in public, is the first step toward mobilizing capital and talent against it.

Physical AI — When AI Steps Out into the Physical World

Among the AI domains Takagi presented, the concept that had the biggest impact was "Physical AI." Until now, AI has stayed in the digital world, supporting human decision-making. Physical AI is different. It completes the cycle of sensor data acquisition → judgment → execution in the physical world. Robots, autonomous mobility, and self-driving vehicles are typical examples.

Why is Physical AI decisively important for Japan? Because it is a response to labor shortages. As humans become scarce in manufacturing, logistics, infrastructure maintenance, and dangerous work, Physical AI becomes indispensable. Takagi also introduced healthcare cases. AI agents handle reception duties and assist with consultations, freeing medical staff from routine work and letting them focus on patient care.

The Monaka CPU — A Hardware Strategy Aimed at 2027

Fujitsu's next-generation CPU "Monaka," scheduled to start being deployed in 2027, is built around a high-efficiency, low-power-consumption design. Implementing Physical AI requires "low power consumption" because it is used in space-constrained settings such as small drones and home robots, while also demanding "high computing performance" to process large volumes of sensor data in real time. Balancing this trade-off is the core challenge of hardware design.

Furthermore, the planned installation of a 1,000-qubit-scale quantum computer this year is symbolic. The simultaneous evolution of software and hardware will form the foundation of an AI-driven society. Fujitsu, having already brought a 64-qubit superconducting quantum computer online jointly with RIKEN in 2023, is now reaching a scale that is orders of magnitude larger only a few years later.

Social Digital Twin — Simulating Policies Before Deployment

What personally excited me the most was the concept of a "Social Digital Twin." It is the idea of using real-world data to digitally recreate cities, regions, and society as a whole, and to explore solutions to complex problems in that environment. The core point is that policies can be simulated and validated before they are deployed.

This has the potential to fundamentally transform decision-making at municipalities and governments. "If we introduce car sharing into this region, what will traffic congestion look like in 10 years?" "What impact will this regulatory easing have on the regional economy?" Such questions can now be answered in advance on a digital twin. Even if a simulation fails, the real world is unaffected. Few changes in policy formulation are more profound.

Human Motion Analytics — Applications in Healthcare

Another striking technology is the ability to convert camera footage of human movement into 3D data at high speed. It is already used in scoring decisions for global competitions where complex, fair judgment is required. In healthcare, it is being applied to solutions that analyze the gait of elderly people and recommend exercises that improve physical function.

For a super-aged Japan, this is an extremely important technology. It compensates for the labor shortage in elderly care while delivering individually optimized exercise prescriptions. Among the healthcare startups TIMEWELL supports, several are well positioned to scale by riding this layer. The combination of foundational providers like Fujitsu and startups that go deep into specific use cases will, I sense, build the next healthcare market.

The reason this layered model works in healthcare specifically is that the regulatory burden of foundational technology — clinical validation, device certification, data privacy — is too heavy for early-stage startups to carry alone. By outsourcing that part of the stack to a platform partner, startups can focus their limited engineering resources on the parts of the experience that patients and clinicians actually feel: the design of the gait-analysis interface, the language of the exercise instructions, the integration into a specific clinic's daily workflow. This division of labor is exactly the pattern that turned cloud computing into a startup-friendly category over the past fifteen years, and it is now arriving in healthcare AI.

The "Hybrid Strategy" for Legacy Enterprises Becoming AI-Native

I want to dig a bit deeper into how Fujitsu itself is transforming into an AI-native company. This is exactly the challenge many legacy enterprises face.

At the heart of Fujitsu's AI strategy is "Fujitsu Kozuchi." It is a platform that unifies the company's AI services internally, while also selectively integrating external partner AIs (foundation models and specialized models). It is fair to call this a hybrid strategy that is neither pure self-reliance nor full outsourcing.

For companies carrying legacy IT assets, the difficulties of AI-native transformation come down to three points: (1) existing systems cannot be stopped, (2) there is no budget to replace everything at once, and (3) there are not enough internal talent. Fujitsu's answer is phased replacement combined with per-use-case optimization. Defend the core systems while migrating front-office work, internal knowledge, and customer touchpoints to AI-native operations in that order — this sequence is the realistic answer.

This is a directly applicable approach for executives at mid-sized and large companies. Switch from "company-wide simultaneous DX" to "redesign the work that delivers the highest value first using AI." This is the only way to keep moving while still carrying legacy systems.

In our consulting work at TIMEWELL, we often see companies stall because they try to choose "the one AI platform" before they have done any real implementation work. That sequence is backwards. The right sequence is: pick two or three concrete workflows where AI can deliver visible improvement within 90 days, build those out using whatever combination of tools is most pragmatic, and then — armed with real production data and real organizational learning — make the platform decision. Fujitsu Kozuchi exists not because Fujitsu picked a platform first, but because they accumulated enough use cases internally to know what a platform actually had to do.

Is Japan's AI Disadvantage Really a "Loss"?

Right after Takagi's talk, while walking through the venue, I kept coming back to one question. "How should we accept the reality of being ninth in the AI vibrancy ranking?"

In data, Japan is clearly behind. We trail the U.S. and China significantly in startup investment, and we are late on foundation model development. But it is also true that Japan has its own strengths.

• First, world-class manufacturing and hardware capabilities. These connect directly to Physical AI implementation
• Second, aging and labor shortage as "forcing functions." There is a counter-logic that makes Japan's urgency for AI adoption the highest in the world
• Third, a stable social infrastructure. It is an environment where AI implementation can be tested in society at large scale

These three factors do not exist in combination in the West or in China. The areas Takagi emphasized — Social Digital Twin, Physical AI, Human Motion Analytics — are all narrowed down to domains where Japan's strengths can be fully leveraged. This is not a coincidence; it can be read as an extremely deliberate strategic choice.

An "AI-First" Organizational Mindset

The question Takagi posed at the end was the heaviest of all.

"Are we truly ready to accept a world in which AI becomes the new normal?"

This is not merely about adopting technology. Starting from an AI-first mindset across every business — without this fundamental shift in organizational culture, it is impossible to extract the full benefits of AI or discover new sources of value.

What TIMEWELL faces every day is precisely this organizational-culture problem. At many Japanese companies, the order to "make use of AI" has been issued. But on the ground, AI is only used as a minor efficiency tool layered on existing work. Companies that have actually gone as far as redesigning their business processes "AI-first" are still in the minority. From my time supporting internal entrepreneurs at Panasonic, this "gap between the order and the implementation" has been a shared challenge for Japanese companies. Developing the middle managers who translate that order into action on the ground is the key to moving Japan's AI implementation forward.

Opportunities for Startups — Become Fujitsu's Partner

When a major company like Fujitsu sets out the banner of an "AI-driven society," it actually creates significant opportunities for startups. Most of the application layers and industry-specific layers running on platforms such as Monaka CPU, the quantum computer, the Social Digital Twin, and Physical AI will be provided by startups.

When Japanese startups partner closely with infrastructure companies like Fujitsu, both wheels — domestic scaling and overseas expansion — start to turn. At the SusHi Tech venue, I saw many startup representatives gathered at the Fujitsu booth. This is the new way Japan's ecosystem is moving. Rather than a major company hoarding partners, the open posture of welcoming partners was clearly raising the energy in the room.

For founders weighing whether to pursue this kind of partnership, my practical advice is to enter the conversation with a clearly scoped pilot in mind, not an open-ended request for collaboration. Large companies have learned to be wary of vague proposals that consume months of legal and procurement bandwidth without delivering measurable outcomes. A startup that arrives with a defined use case, a clear success metric, and a willingness to instrument the pilot from day one is far more likely to graduate from the booth conversation into an actual contract.

Fujitsu's Philosophy of "Human-Centered AI"

What CTO Takagi emphasized again and again was the philosophy of "Human-Centered AI." It is the perspective that AI is not a replacement for humans but an augmentation of human capabilities.

This is not just rhetoric. At the implementation level, Fujitsu rigorously enforces the design principle of "AI proposes the judgment, humans make the final decision." In medical diagnosis, financial review, and manufacturing quality control alike, AI is a tool that accelerates decision-making, while final responsibility lies with humans.

This philosophy also has strong affinity with future AI regulation. The EU AI Act mandates "human oversight obligations for high-risk AI," and Fujitsu's design principles already preempt this regulation. The "human-centeredness" of Japanese AI is going to be a powerful differentiator in why global markets choose it.

The three pillars of implementation that Takagi cited were: (1) transparency in decision-making, (2) employee education to understand AI's limits, and (3) the development of an AI governance framework. This is not just for unique large enterprises — it is a framework that any organization can drop directly into place. Before introducing AI, it is worth checking the current state of these three pillars first.

How Do You Upskill 400,000 Employees for the AI Era?

Fujitsu is a giant employer of around 400,000 people worldwide. The question CTO Takagi raised was, how do you upskill this workforce for the AI era?

Fujitsu operates an internal university called "Fujitsu University," which offers AI, data science, and cloud courses to all employees. Several million hours of learning per year reportedly accumulate across the organization. This is not mere training — it is a rewriting of organizational DNA.

The key for large Japanese companies to survive the AI era lies in this kind of upskilling investment. TIMEWELL's support for internal entrepreneurs is part of this same context. The work of raising the AI adaptability of the entire organization is, I believe, a contribution to the entire Japanese economy that goes beyond any single company's profit.

There is also a more uncomfortable truth here: not every employee will adapt at the same rate, and the gap between fast adopters and slow adopters is likely to widen rather than narrow. Companies need to plan honestly for what to do with that gap. Pretending it does not exist, or treating it purely as an HR risk, will lead to either cosmetic training programs or quiet resentment. The companies that will thrive in this transition are those that treat AI fluency as a real skill — one that requires time, support, and accountability — rather than as a cultural slogan that everyone is assumed to have absorbed.

Conclusion — From "Using AI" to "Designing Society with AI"

The conclusion I took away from Takagi's talk was clear.

Victory in the AI era will not be decided by organizations that stop at "using AI in their work" — it will be decided by organizations that move at the level of "redesigning society itself through AI."

The number "ninth in AI vibrancy" is admittedly a low starting point. Yet I sense that with the two wheels of Social Digital Twin and Physical AI, Japan still has a real chance to forge its own path. For that to happen, large companies, startups, government, and universities all need the resolve to redesign their organizations "AI-first."

Takagi's talk was an hour that interrogated that resolve. I myself, even from the small platform of TIMEWELL, renewed my commitment to redesigning the infrastructure for challenge "AI-first." For Japan to once again take a leading role in the AI era, what we need is not technology so much as resolve. This talk made that point unmistakably clear.

---

Related Articles

• SusHi Tech Tokyo 2026 Keynote Report — Governor Koike and Prime Minister Takaichi on the "Vision of a Startup Nation Born in Japan"
• Who Drives Discontinuous Growth? Four Turning Points for Japanese Companies to Take Risk at the Next Level
• Complete Guide to Enterprise AI — Conditions for Large Companies to Master AI

A Note from TIMEWELL

We also offer individual consultations through TIMEWELL's AI consulting service WARP. You can start with a 30-minute online consultation.

---

References

[^1]: YouTube. An AI-Driven Society Shaped by Technology. https://www.youtube.com/watch?v=5tSCXM0h3y0 [^2]: Fujitsu official site. https://www.fujitsu.com/jp/ [^3]: Stanford HAI. Global AI Vibrancy Ranking. https://aiindex.stanford.edu/vibrancy/