In 2019, the US Department of Energy selected Intel as the processor supplier for the $ 500 million Aurora supercomputer, which is supposed to provide a breakthrough in artificial intelligence for climate modeling, medical research, and computing nuclear interactions. Aurora has the potential to be the first computer in the world to cross the one exaflops threshold of computing power. However, given the delays, China is likely to bypass the United States, as the Celestial Empire is currently developing three supercomputers in parallel, capable of developing similar computing power.

The delay in launching Aurora sheds light on a more pressing issue for the United States. The country’s semiconductor industry is noticeably lame. In its July quarterly report, Intel said it was up to 12 months behind internal plans to implement a 7nm process technology. This was a significant setback for Intel, which subsequently lost nearly $ 50 billion in market capitalization. The company is one of the few in the United States that still manufactures chips in-house. Other leading manufacturers such as Apple, Nvidia, or AMD are outsourcing chip manufacturing to Samsung or TSMC, which have the most advanced manufacturing facilities available today. While Intel is struggling to get 7nm chips into production, TSMC has already announced that it is preparing to begin full-scale production of 3nm chips. Thus,

The semiconductor industry plays a key role in the US military’s plans to use artificial intelligence. As of early 2020, the country was only marginally ahead of China in artificial intelligence, according to a study commissioned by the US Army by RAND. This is only because Chinese companies continue to rely heavily on American developments, as well as on Korean and Taiwanese manufacturing facilities. The report also highlights the key role of AI in US military strategy. It argues that maintaining an advantage over China’s military systems and weapons using artificial intelligence directly depends on the ability of the United States to maintain an advantage over China in the field of AI at the national level.

This dynamic, along with Aurora’s rather palpable setback, means the US is likely to significantly increase subsidies to tech companies for AI research and development. Just this week, the Trump administration announced a $ 1 billion subsidy program that will focus on AI and quantum computing research by Intel and Microsoft.

These kinds of programs can be predicted to become more frequent as the US government closely monitors investment activity in the Pacific region. Recall that in early July, a state-backed Chinese investment fund funded local chip maker SMIC worth $ 6.6 billion to create a viable alternative to Samsung and TSMC in China. And this looks like a problem for the United States.