• At its AI Tour Tokyo event recently, Microsoft brought together its global leaders and local partners to highlight its strategy and vision in the AI space.
• Japan’s enterprise sector is quickly adopting AI solutions, with 85% of Nikkei 225 companies using Microsoft’s Copilot.
• Microsoft’s strategy for creating an agentic web rests on an improved interface, Copilot AI stack and on-device AI.
• Quantum computing will increasingly enter the digital transformation narrative.

Microsoft held its AI Tour Tokyo event on March 27, bringing together its global leaders and local partners to highlight its strategy and vision in the AI space. Japan has recently been hosting more international technology events, such as the recent NVIDIA Japan Summit, as it is clear the country is becoming more of a focus in emerging tech. The Microsoft event was inaugurated by its CEO and Chairman Satya Nadella who outlined the company’s vision and strategy to create a truly agentic web.

The Trillion-dollar Question: More Efficient AI

Satya started off his keynote by addressing the overarching question looming over the AI industry currently – very high energy costs posted by companies like OpenAI, an appropriate point to discuss given recent announcements by the likes of DeepSeek. He also mentioned that the AI sector is experiencing a phenomenon known as Jevons paradox where advances in technology for a product with elastic demand led to increased use of that product as opposed to less utilization due to higher efficiency. For example, when James Watt introduced a more efficient coal-fired steam engine in 1769, demand for coal soared instead of decreasing because more industries adopted the new engine, outweighing the per-unit efficiency gains at the aggregate level. This is now happening with AI and energy consumption, and Satya mentioned that more efficient software will be critical going forward.

Source: Microsoft

Three Pillars of an Agentic Web

In order to create a truly agentic web, Satya argued, three things need to happen – (1) The industry needs a better user interface which must be multimodal, (2) More intelligent software is needed for reasoning and planning and (3) More memory needs to be available to handle this increased processing. Microsoft is trying to enable this across the tech stack in three ways:

Copilot UI for AI: Microsoft showcased its voice-enabled Copilot feature for consumers and enterprises, which, for example, provides search results by speaking to a chatbot as opposed to using a browser. Microsoft also demonstrated its recently announced Researcher and Analyst AI agents, including integration of the technology into solutions such as Teams, PowerPoint and Excel.

Copilot AI Stack: This gives enterprises the opportunity to create bespoke AI applications using their own data. Microsoft also announced that it would be upgrading its data centers in both the eastern and western regions of Japan with AI capabilities to enable applications to run smoothly in the country. GitHub is also being granted AI capabilities under its Padawan program and the number of developers in Japan has surpassed 3.55 million, growing at over 20% per year.

Copilot Devices for AI: Hardware will also be critical to make the transition to agentic AI complete. Therefore, Microsoft showcased its current range of products which perform many generative AI functions on the device.

Japan’s Progress and a Taste of the Future

Satya mentioned that 85% of Nikkei 225 companies use Microsoft Copilot, many of which have created specialized teams to use the technology. Japan Airlines was mentioned as a case study where it is using an SLM-driven on-device application to enable its flight attendants to create flight reports when they lack connectivity. The average time to create these reports has fallen from over one hour to 20 minutes.

It is clear that Japanese enterprises are adopting AI at a much faster pace compared to technologies like cloud computing and advanced cybersecurity and this bodes well for digital transformation efforts in the country. It is also clear from this, and other events that foreign vendors are driving this change.

The keynote concluded with Satya mentioning that security, privacy and safety are three key pillars of Microsoft’s AI strategy. As a preview of what to expect from the company going forward, he mentioned quantum computing, which is only just starting to enter the IT narrative. Satya demonstrated Microsoft’s Majorana 1 “pre-quantum” solution which he believes will have game-changing applications in sectors like drug discovery and materials science and will be the next step for the industry after the AI boom.