Semiconductor manufacturing technology – has Asia managed to gain a significant advantage in the sector?

Seventy-five percent of semiconductors, or microchips are manufactured in Asia.

Intel Corporation, which has produced $63 billion in net income over the last three years, has been urging the US government to assist resuscitate chip production at home — through incentives, subsidies, and/or tax breaks, as Taiwan, Singapore, and Israel have done.

The efficacy of Intel’s lobbying was evident in late March, when the White House announced $50 billion in subsidies for US semiconductor companies as part of its $2 trillion infrastructure plan to address shortages and US vulnerability to foreign chip makers. According to the White House, the subsidies for the semiconductor industry enjoy bipartisan backing in Congress. After spending $84 billion on share buybacks since 2011, Intel pleaded for $50 billion in subsidies for the US chip industry.

25 years ago, the United States produced 37% of the world’s semiconductor manufacturing in the U.S. The United States’ semiconductor production has decreased to the point that it currently accounts for barely 12% of global output.

“… It’s a problem, because relying on one region, especially one as unpredictable as Asia is highly risky.”

“For chips with the tiniest transistors – there is no ‘made in the US.’ option. Intel currently doesn’t have the know-how to manufacture the most advanced chips that Apple and the others need.”

“We (Intel Corporation) had some product stumbles, some manufacturing and process stumbles. Perhaps the biggest stumble was in the early-2000s, when SteveJobs of Apple needed chips for a new idea: the iPhone. Intel wasn’t interested. And Apple went to Asia, eventually finding TSMC: the Taiwan Semiconductor Manufacturing Company – today, the world’s most advanced chip-manufacturer, producing chips that are 30% faster and more powerful than Intel’s”

– Pat Gelsinger during CBS 60 minutes interview May 2, 2021.

Can we openly say that the technological axis of the world is located presently in Asia?

In July 2021, Pat Gelsinger presented IDM 2.0 Intel’s roadmap with its plans to ramp its capacity and catch up with leaps in processing technologies by 2025. Intel has 15 fabs all over the world, assembly and test sites in Vietnam, Malysia, Costa Rica, China and US making around 8000 products (2 bln units per year) for its 2000 customers. Intel has also fab expansion underway in Ireland and is in talks on projects in Italy and Germany. In March Intel announced its plans of spending USD 20 BLN to build two new chip plants in Arizona. It takes a long time to build the chemical delivery and electrical systems to make it possible for a fab to run. At the moment, Intel is not capable of producing the 5 Nm chips (advanced data processing, AI processing). Making 5 Nm chips is possible with using the extreme ultraviolet lithography machine where the tin droplets fall into the vacuum chamber , where the droplets are illuminated by high-energy laser pulses – 50,000 times per second. The tin atoms are ionized and an intense plasma is formed. The collector mirror collects the EUV radiation that is emitted by the plasma in all directions. The light is collected in a beam and transmitted to the lithography system, where it illuminates the silicon wafer. So the first to produce 5Nm chips in US will be TSMC with its plans to build USD 12 BLN Fab in Arizona. Intel is highly volumed in 10 Nm chips that were supposed to be 7 years ago and the productions processes have been constantly delayed (14Nm, 10Nm, 7Nm). Intel is switching from designing and manufacturing only its own design chips to the role of a subcontractor acting on behalf of companies such as Apple or Amazon, Qualcomm or US government.

Will it be enough?