Microsoft Builds a $2.5 Billion, 6,000-Person Operation to Support Enterprise AI Adoption

On July 2, Microsoft announced the establishment of a new organization dedicated to helping enterprises adopt AI. The investment is expected to total $2.5 billion (roughly ¥390 billion), with a large-scale setup of around 6,000 employees.

The new division will be staffed with specialists in engineering, corporate training, and industry-specific consulting, positioning Microsoft to support customers on strategy as well as technology. Amazon, OpenAI, and Anthropic all operate similar adoption-support units — a sign that competition among major AI players is shifting from model development to the stage of real-world deployment.

Anthropic Reportedly in Talks with Samsung on Custom AI Chip Development

According to reports from Bloomberg and The Information, Anthropic is in discussions to partner with Samsung Electronics and make it the manufacturing partner for its own AI chips. The plan is said to be at an early stage, with details such as the processor's specifications, performance, and how it would be implemented in servers still undecided.

With Google fielding its own TPUs and Amazon its Trainium chips, Anthropic's exploration of a custom chip path can be seen as a strategically important move from the standpoint of curbing inference costs and achieving supply independence.

Low-Cost Chinese AI Models Close In on Anthropic and OpenAI

Reuters reported on July 2 that new Chinese-made AI models are showing performance that approaches the frontier models of Anthropic and OpenAI despite their low prices. Details such as specific model names and benchmark results remain limited, but the report indicates that the Chinese trend of "cheap yet powerful" models is continuing.

The report also highlights how Chinese players are narrowing the gap through model efficiency, even as U.S. export controls on American-made chips persist.

OpenAI Considered Offering 5% of AI Profits to the Trump Administration

According to The Verge, OpenAI presented the Trump administration with a proposal to provide the government with 5% of the profits from the AI boom. The arrangement appears to be positioned as a return for government backing of AI infrastructure.

While the specifics of the mechanism and whether any agreement was reached remain unclear, the picture of a private AI company sharing its profits with the state is drawing attention as an unprecedented new framework.

Japan's Supreme Court: AI Cannot Be a Patent Inventor

According to reports from the Yomiuri Shimbun and other outlets, Japan's Supreme Court has finalized a ruling that AI alone cannot be recognized as the inventor on patent applications. The decision stands as an important judgment in the worldwide debate over AI and intellectual property rights.

Similar lawsuits have been filed in the United States and Europe as well, and Japan's ruling is likely to serve as a reference point in the international discussion.

The Economist: America Should Not Lock Up Frontier AI

In a July 2 editorial, the British newspaper The Economist sounded a warning against the United States over-regulating and restricting frontier AI models. The paper emphasized the need to balance the freedom to innovate with safety, noting that regulation going too far could erode competitiveness.

Read alongside OpenAI's profit-sharing proposal and the rise of Chinese models, it becomes clear that the debate over the direction of AI governance now stands at a critical crossroads.

---