Nvidia goes big and U.S. chip projects stutter

Hello everyone, this is Akito from Singapore.

About two decades ago, I had a sneak peek at the PlayStation 3 before it was released at the former Sony Computer Entertainment headquarters in Aoyama, Tokyo. One of the standout features of the gaming console was its graphical performance. Although I'm not a gamer, I remember being astonished by how realistic a race car game being played on a large-screen TV looked.

The PlayStation 3 was a strategic product for Sony as the company aimed to increase its presence in the semiconductor industry. This console was equipped with a processor called "Cell'' that was jointly developed by the Sony Group, Toshiba and IBM. At that time, Sony aimed to incorporate Cells into a wide range of products, such as flat-screen TVs and home servers. The grand vision was for these chips to be used in all home digital equipment, just as Intel's chips were (and still are) ubiquitous in personal computers.

The Cell strategy ultimately failed as Sony couldn't recoup the massive manufacturing investment.

But the PlayStation 3 was equipped with another chip that would become the centerpiece of today's semiconductor industry -- a graphics processor that used Nvidia's technology to generate realistic graphics. Some 20 years later, Nvidia has ridden this technology to the top of the semiconductor industry.

Nvidia has unveiled its latest graphics processing unit (GPU) for artificial intelligence computing as the Silicon Valley company looks to cement its grip on the global market for AI chips, writes Nikkei Asia's Yifan Yu.

During its annual GPU Technology Conference, the company announced the new Blackwell GPU, which it says provides better AI training performance and energy efficiency than its popular