The AI-Accelerated Digital-Only Economy

Published: September 9, 2026

Graduate School of Information Science and Technology, The University of Tokyo
Professor Hiroshi Esaki

1. From Physical-Only to Digital-Only

　Having gone through the Agricultural Revolution and the Industrial Revolution, humanity has now entered the stage of the Digital (Information) Revolution. A steam engine is a system that uses the intense heat generated by the combustion of fossil fuels to vaporize water (phase transition) and then converts the kinetic energy of the steam into rotational energy via a turbine. However, once technology was invented to ”reciprocally” convert the rotational energy generated by steam engines into electrical energy (i.e., power generation and motors), that electrical energy could be supplied via copper wires—without the need for belts or gearboxes—to the motors of machine tools within factories. As a result, the machine tools in the factory were “freed from belts and gearboxes,” enabling them to start, stop, and even”Movement”It has become easier to do this. This seems to be the true Second Industrial Revolution.
　In recent years, humanity appears to be undergoing a major revolution in energy infrastructure—one comparable to the discovery of electrical energy and the development of methods for its utilization—through the invention of technologies for the transmission and processing of digital bits. Synchronous, centralized client-server systems based on electrical transmission and processing technologies are evolving into asynchronous, distributed peer-to-peer systems that do not require long-distance distribution networks capable of carrying large amounts of electrical energy. It is widely recognized that the costs of building and managing digital networks are more than two orders of magnitude lower than those of power grids, and this realization is sparking a major infrastructure revolution.
　In the pre-20th-century world, which was entirely physical, the power of bits enabled the optimization and streamlining of physical systems through their emulation in digital space. After this process, and following the design phase, cyberspace took center stage, and physical systems became, in a sense, something “printed out” from cyberspace into the physical world.“Unique and Niche”We can view the world as a system that is constantly changing and evolving. In other words, it is evolving into a system that is self-contained within cyberspace or the digital realm (or a system in cyberspace that interfaces with the physical world only to the bare minimum). For example, newspapers—a classic example—began as systems confined solely to physical paper media, but through digitization, they have moved beyond physical paper and evolved into a business model that is nearly ”entirely digital.” Physical space is subject to ”finite” constraints—the laws of physics—from which it cannot escape. However, since cyberspace has no “upper limits” or “restrictions,” it can be viewed as a diverse space for activity where space can be expanded almost infinitely and where free rules can be applied. The Age of Discovery and the subsequent era of colonial economies demonstrated that on Earth,Endless Lands and Oceans...existed, and as a result, business investments were made based on the idea that the economy could expand indefinitely. However, since the Earth is a sphere,The fact that there are upper limits to land and ocean areasAs this became clear, a scramble for physical resources among the major powers began, and humanity ended up experiencing two world wars.
　The digital industry, whose capabilities and density have improved exponentially in accordance with Moore’s Law, has created a digital/cyber space capable of infinite expansion (i.e., one with no upper limit). Almost,“A Digital Economy with No Upper Limit”and many economies“Fully Digital”It appears to be heading in that direction. As a factor that limits the growth of the digital economy,(1) Global Warming and (2) Electric PowerWouldn't 2026—the first year of the second quarter of the 21st century—be the year when this began to be recognized?
However,
(1) Digital infrastructure, which consists of electronic devices, relies on the availability of electricity, and it is necessary to ensure “energy and economic security for cyber infrastructure.”
(2) Discussions on “data governance”—which concerns the management and use of data—are gaining momentum among governments and companies around the world.
　This is why the “Public-Private Consultative Council on Watt-Bit Collaboration,” which has been the subject of full-scale government discussions since March 2024, was launched. Digital and cyber infrastructure underpins and enables the digital economy, which is set to expand in full force in the second quarter of the 21st century; among these, data centers (along with energy and information and communications infrastructure) constitute the most critical infrastructure foundation.^⑴The following is a list of the most common problems with the

Figure 1. Comparison of Infrastructure Costs for Logistics, Electricity, and Information Flow

2. Digital Economic Security

　The key elements of the Second Industrial Revolution were “fossil fuels,” “iron and concrete (i.e., protons and neutrons),” and “electricity (i.e., high-voltage electricity).” In contrast, the key elements of the Third Industrial Revolution are electrons (i.e., low-voltage electricity) and photons (i.e., electromagnetic waves). This signifies that systems, which were previously premised on and driven by the physical world, have transformed and evolved into systems premised on and driven by the digital world. At the same time, we must recognize that it has become essential to establish and build the foundation for data governance—that is, “Data Sovereignty”—which ensures the accuracy and legitimacy of data. Blockchain can be viewed as an example of a digital system that implements a democratic decision-making process to guarantee and ensure the legitimacy of data. Voters—who are multi-stakeholders—make decisions through voting while mutually monitoring one another.
　Another important transformation and evolution is this: (1) Since a nation is a “unit that shapes social and economic activities in physical space” and must operate under the constraint of the Earth—a finite three-dimensional physical space—a struggle for physical resources (particularly fossil fuels) inevitably arises. On the other hand, ➁ “activities in digital space” have virtually no limitations (upper bounds) on their scope; the scope of these activities can be expanded without restriction, and as a result, the competition for physical resources is significantly alleviated. However, the physical infrastructure that constitutes digital space and the energy required to operate it exist within physical space and are strongly influenced by “nations.” That said, this pertains to the environment on the Earth’s surface; in higher dimensions and outer space, the concept of a “nation” does not currently exist. On the Earth’s surface, “public”The Sea”"such as 'public”Sky”It is a space where the potential exists for a struggle for hegemony and vested interests to erupt among nations (or corporations, etc.). To avoid or resolve conflicts over this new space, we must pool the wisdom derived from humanity’s past experiences and implement solutions; otherwise, we may once again find ourselves facing wars between nations (or between corporations).

3. Development of Digital Infrastructure

The rapid increase in power consumption within digital infrastructure—specifically, in computing infrastructure, such as data centers, and high-speed wireless infrastructure—is becoming a major concern. Computing that utilizes vast amounts of digital data, such as artificial intelligence and deep learning, is generating a wide range of new value and is poised to trigger large-scale digital transformation (DX). Since these communications and computing infrastructures require massive amounts of electricity, shouldn’t future infrastructure development evolve these two types of infrastructure—which have traditionally operated independently—into an integrated infrastructure based on collaboration and coordination?
　The increase in electricity consumption by the data center and semiconductor industries is poised to bring about a major shift in Japan’s energy policy. Total electricity consumption, which had been on a downward trend, is now set to reverse course and begin rising due to these two industries, as outlined in the “7th Basic Energy Plan”^⑵Consequently, strategic collaboration between these two industries—along with other cutting-edge sectors—and the renewable energy sector has emerged as a key direction. It is important to note that the total electricity consumed by these two industries, which constitute the data center infrastructure, accounts for less than a few percent of the total; while a significant increase is inevitable in the future, it does not yet represent a large proportion. Nevertheless, these two industries are key core sectors driving digital innovation (DX), and they are expected to make significant contributions to the advancement of GX through data-driven productivity improvements and structural transformation across all industries; from this perspective as well, they are vital industries. As discussed in Andrew McAfee’s *Capitalism Is Dematerializing*, the concept of “generating and creating large outputs from limited resources”—“More from Less by Bits over Watts”—is central to this discussion.”^⑶However, this is driven by digital technology. With the advent of generative AI, the data center and semiconductor industries have seen their prominence increase dramatically. Through the massive reduction in energy consumption—achieved through the use of AI in both “AS IS” and “TO BE” scenarios—in industries that were previously nearly entirely physical due to digitalization, it may not be unreasonable to expect that the total energy consumption of digital infrastructure could drop to 20–25 percent, similar to that of the human body.
　From the perspective of latency requirements, it is unacceptable for AI servers to be located in distant data centers. In particular, while the training process—which requires a large amount of power—has no latency requirements and can therefore be run in large-scale data centers, the inference function requires low latency, leading to situations where operation in remote data centers is no longer acceptable. This represents a shift from centralized (client-server) to distributed (peer-to-peer) architectures in data centers. China’s national policy, known as “Computing in the West, Data in the East,” outlines the next phase of this architecture in line with this trend.

Figure 2. 13th Session of the National People's Congress of China, 2021: “East Data, West Computing” (i.e., learning in the west, inference in the east)

4. Digital Economic Security

　We must recognize that virtually all industries now rely on digital systems and that their operations depend on the Internet. All industries and economies are predicated on the availability of “electricity” to power electronic devices and on the existence and use of the “Internet.” As a result, discussions on “digital economic security”—concerning the management and use of data—are gaining momentum. In these discussions, the concept of “self-reliance” is frequently accompanied by the option of “disconnection.” Security is increasingly being viewed as the means to enable “self-reliance and disconnection.” The Cold War structure between the U.S. and the Soviet Union, which symbolized “self-reliance and disconnection,” was characterized by “vodka and Coke”—^⑷It collapsed due to the expansion and advancement of interdependence in the real economy, as symbolized by [...]. By increasing this interdependence, they made conflict impossible. Furthermore, as a result of a shared recognition that joining forces to collaborate and cooperate was a wiser course than division and conflict, the process led to the fall of the Berlin Wall. Cooperation and collaboration require mutual trust and respect. In asymmetrical relationships, it is impossible to build and maintain cooperative ties. This also underscores the importance of respecting diversity and promoting inclusivity. To establish digital economic security, shouldn’t we steer the course toward “mutual cooperation and interdependence” rather than “self-reliance and isolation”?

5. How Generative AI Is Changing Investment Amounts and Dynamics

　The rapid increase in investment in generative AI (artificial intelligence) and the shifting dynamics seem to have created a situation where “investment capacity” (scale) and “integration capacity” (structure) determine the survival of a business. Building the infrastructure for the AI era requires two things: (1) investment on an unprecedented scale and (2) investment based on a new geopolitical landscape. It seems that companies will not be able to survive without redesigning their capital and business structures.
　With the advent of generative AI, the data center industry—which is driving the rapidly accelerating digital economy—requires massive investments that the companies traditionally comprising these industries are simply unable to meet. In industries such as pharmaceuticals, biochemistry, and semiconductors, the market structure has shifted such that only companies with the focus, financial strength, and risk-taking capacity to make massive investments can survive—a shift that has caused Japanese companies to rapidly lose their competitiveness in the global market. Is the data center industry on the verge of facing the same situation? The NTT Group was quick to accelerate the international expansion of its data center business—which supports NTT Data’s operations—alongside NTT Data’s overseas expansion. In doing so, the group established a new company in Singapore to conduct global business operations, as Singaporean REITs (real estate investment trusts) offered advantages over Japan’s investment structure, rules, and regulations. In Japan, too, investment in data centers is rapidly accelerating and expanding, with participation not from traditional telecommunications companies, system integrators, or service providers, but rather from fund management firms and real estate companies capable of making significantly larger investments (one to two orders of magnitude greater). Consequently, special provisions for J-REITs have been established in Japan regarding data centers, and thereby achieving deregulation of investment in data centers.

6. Cybersecurity

　Providing users with a sense of ”peace of mind” is referred to as “trust.” It can be understood as a situation in which there is sufficient credibility to engage in business activities or maintain relationships. Trust is relative; it is not an absolute standard. It must take into account the risk of unforeseen incidents (accidents or events) while considering an acceptable level of risk. The fundamental responsibility must lie with the end user (i.e., self-reliance). However, to achieve trust in the digital space as a whole, we must establish a structure comprising, first, “mutual aid” through collaboration and cooperation among relevant stakeholders, and finally, “public assistance” provided by the government and other entities. “First self-reliance, then mutual aid, and finally public assistance.” The necessity of maintaining and upholding this order is demonstrated by humanity’s past history (particularly the state-imposed information control leading up to World War II).
　Laws vary from country to country, and in digital networks that facilitate the global exchange of digital information across borders, security measures and system optimization must be implemented across countries with different regulations.
　In December 2017, the Ministry of Economy, Trade and Industry launched its “Industrial Cybersecurity” initiative and outlined the CPSF (Cyber/Physical Security Framework) as a roadmap for the design, construction, and operation of a cybersecurity infrastructure that enables data integration within supply-and-demand networks formed by diverse, autonomous data spaces. This has come to be recognized as a critical issue for individual countries and the world as a whole by 2026.
　Furthermore, the National Cyber Coordination Office, established in July 2025, will issue “Unified Standards for Cybersecurity Measures for Critical Infrastructure,” which will outline the following key guidelines:
　(1) Emphasizing and Enhancing Resilience
　(2) Countermeasures that prioritize areas of focus based on security risk assessments
　(3) While this assumes a zero-trust model,
　　　(A) Cannot be protected by a gateway firewall (self-protection is essential)
　　　(b) Cannot be protected by a proprietary OS or protocol (AI will attack it)
　　　(c) However, segmentation is important as a defense against the spread of damage.

　To address these environmental and situational challenges, cybersecurity measures are being promoted across four levels.There are four levels: ➀ Product, ➁ Production Sites (factories, etc.), ➂ Organization (companies, etc.), and ➃ Supply Chain Network.Thus far, we have discussed vulnerabilities in cybersecurity measures related to the product described in (1). However, considering the sustainability and continuity of social and industrial activities, cybersecurity measures at the three levels mentioned above must also be implemented. This refers to a model in which numerous virtual supply chain spaces—known as “Data Lakes” or “Data Spaces”—coexist within cyberspace. We must ensure the secure and trustworthy circulation and integration of digital data both within Data Spaces and across them.

7. Conclusion

　It is said that we must achieve carbon neutrality by 2050. Solving environmental pollution problems is in the public interest. However, at first, the business community is reluctant to take action, arguing that measures to combat environmental pollution are a cost that will drive down productivity, output, and profit margins. Nevertheless, we must create a win-win situation in which the resolution of environmental pollution issues is ultimately achieved through the reduction of waste and the invention and adoption of new technologies, so that the pursuit and enhancement of private interests leads to the resolution of environmental problems—which serve the public good.
　The so-called “GAFAM” (Google, Apple, Facebook , Amazon, and Microsoft) have been able to raise massive amounts of capital from the stock market. Furthermore, by establishing a corporate governance structure that allows for management and investment decisions unconstrained by shareholder opinions, they have built a system capable of owning and operating an end-to-end infrastructure—from energy infrastructure to computing infrastructure and, ultimately, application infrastructure. The industrial structure of digital infrastructure has changed. It seems that Japanese companies, by prioritizing shareholder interests and profit-driven restructuring alongside short-term business stability in their corporate governance, have ultimately lost both the motivation and the opportunities for investment and innovation. I hope that, using AI as a reason or catalyst, Japan’s industrial sector will take on the challenge of revitalization—one that differs from the “Lost 30 Years.”

References

Author Profile

Hiroshi Esaki
Professor, Graduate School of Information Science and Technology, The University of Tokyo

1987: Completed the master’s program in the Department of Electrical Engineering, Faculty of Engineering, Kyushu University. Joined Toshiba Corporation.
For two years starting in 1990, Bellcore, New Jersey, U.S.; for two years starting in 1994
Visiting Researcher at Columbia University in New York City, U.S.
In 1998, he became an associate professor at the Large-Scale Computer Center, University of Tokyo; in 2001, he became an associate professor at the Graduate School of Information Science and Technology, University of Tokyo.
He has held his current position (Professor, Graduate School of Information Science and Technology, The University of Tokyo) since 2005.
Chair of the WIDE Project; Chair of MPLS-JAPAN; Chairman of JPNIC (Japan Network Information Center);
Representative of the University of Tokyo Green ICT Project; Vice President and Chair of the Steering Committee of the Japan Data Center Association;
Chairman of the IPTV Forum, Chairman of the JNSA (Japan Network Security Association),
Chairman of the CSAJ (Cloud Security Association of Japan). Ph.D. in Engineering (University of Tokyo).