The AI Power Balance Could Shift Dramatically Next Year

Speaking on a science and know-how program aired on December 26, 2025, Matsuo stated the “surprise” situation for 2026 is a sudden rearrangement of the AI pecking order, together with the chance that the semiconductor sector’s present dominance may very well be disrupted, that progress towards synthetic normal intelligence (AGI) may alter company competitiveness, and that so-called “physical AI,” robots powered by superior fashions, may start altering on a regular basis life quicker than many anticipate.

In the race between OpenAI and Google, Matsuo pointed to rising consideration on Google’s Gemini 3, which he stated is extensively seen as a high-performing mannequin and is reported to be scoring strongly on benchmarks, whereas additionally noting that OpenAI is unlikely to cede floor simply. He cited public indicators of heightened urgency within the sector and stated the competition will seemingly hinge on how either side leverages expertise, funding and product velocity, with Google’s scale on one aspect and OpenAI’s aggressive iteration on the opposite.

Matsuo additionally highlighted the momentum behind lower-cost fashions and open methods, particularly from China, saying the unfold of open-source or open-weight approaches is making a scenario the place fashions developed with huge funding will be approached by techniques which can be cheap, and even free to make use of. He prompt the stability between closed and open ecosystems could depend upon the tempo of technical progress, with closed fashions tending to guide throughout fast advances, and open approaches catching up when progress slows.

On AGI, Matsuo stated predictions are diverging, with some distinguished voices arguing it may arrive inside a number of years, whereas others have begun to query whether or not the tempo of enchancment is cooling. Even so, he stated the size of worldwide competitors makes it believable {that a} breakthrough may seem out of the blue, and that the trail could require strategies past the present transformer structure, even when “workarounds” or sudden improvements emerge.

For Japan, Matsuo argued for pursuing each adoption and home growth, framing the controversy by way of the idea of “sovereign AI,” the concept a rustic ought to retain the aptitude to construct core AI techniques at residence. Relying too closely on different nations, he stated, carries dangers starting from service disruption to produce constraints in semiconductors. At the identical time, he cautioned that making an attempt to match the world’s prime fashions head-on would demand funding on the size of U.S. and Chinese giants, and stated Japan ought to weigh cost-effectiveness whereas pushing industrial innovation by way of real-world purposes.

The dialogue broadened into semiconductors, the place Matsuo stated strikes to problem Nvidia are accelerating, together with efforts primarily based on Google’s in-house TPU chips, alternate options from AMD, new startups, and China’s push to construct home chips similar to GPUs. The program additionally launched a Japanese challenger: a semiconductor startup launched on the Japan Advanced Institute of Science and Technology, led by Kenji Fujiwara, a former Sony engineer concerned in chips for the PlayStation 2 and PlayStation 3. The startup is creating a brand new computing structure known as CGLA, which it says can run AI with dramatically decrease energy consumption, probably utilizing as much as 90% much less electrical energy than Nvidia’s GPUs, by lowering the quantity of information shuttled backwards and forwards throughout computation.

Fujiwara argued that whereas Google’s TPU is extremely power-efficient, it’s strongest in a narrower set of computations, and stated the chance lies in combining excessive versatility with low energy use. The program framed the trouble as a part of a wider Japanese push to regain competitiveness in superior chips, together with the government-backed foundry venture Rapidus, which was described as receiving help totaling 2.9 trillion yen. Matsuo stated the strategic worth of home manufacturing capability is important, however added that long-term success is dependent upon constructing prospects and linking fabrication to the broader AI worth chain, from purposes to information facilities.

A separate section aired on December 23, 2025, checked out how U.S. financial coverage and “AI bubble” worries may form markets in 2026, with commentary from Miyuki Obata of SMBC Nikko Securities America. The program stated the S&P 500 had risen about 15% year-to-date by late December regardless of volatility, and centered on uncertainty across the Federal Reserve management transition anticipated by May 2026, with Kevin Hassett and Kevin Warsh mentioned as main contenders and Christopher Waller additionally talked about in reporting. Obata stated market pricing pointed to round two charge cuts, whereas the median of Federal Open Market Committee projections implied fewer, and warned that overly aggressive cuts underneath political strain may reignite inflation considerations or elevate questions concerning the Fed’s independence.

On the query of an AI-driven bubble, Obata stated there are indicators of froth, however argued the chance of a full-scale collapse is restricted, partly as a result of traders have begun scrutinizing enterprise plans and profitability extra carefully, resulting in higher stock-by-stock choice inside the AI sector. She additionally stated AI demand stays sturdy and that coverage help may proceed to underpin funding.

In Tokyo, a market outlook section aired on January 5, 2026, described a powerful begin to the yr, with the Nikkei common rising 1,493 factors to 51,832 on the primary buying and selling day, and avenue traders naming AI, semiconductors and financials as areas of curiosity whereas acknowledging geopolitical dangers. Commentators on this system debated how far the rally can prolong, with some suggesting a variety that features potential volatility and a mid-year correction, and others holding a 60,000 stage in view later in 2026 or past, whereas warning that the clearest draw back situation can be a sharper-than-expected unwind of enthusiasm round AI.

Source: テレ東BIZ