Photo: Arijit Raychowdhury, EE prof. at Georgia Tech (left), and Bolaji Ojo (center) and Junko Yoshida ( right)
What’s at stake?
Does a premier U.S. engineering school like Georgia Tech need a TSMC chip fab in Arizona?
Absolutely. Not for strengthening the physical supply chain but for creating an “intellectual supply chain,” says Arijit Raychowdhury, a Georgia Tech EE professor.
Decades into the practice of going fab-lite or fabless, the U.S. semiconductor industry has lost its mojo. No longer can the industry effectively connect advanced chip architecture and production capabilities with academia, labs, and IP in its own country. At stake — more than national pride — is the future of the engineering workforce.
Many key U.S. semiconductor players support TSMC’s fab buildout in Arizona, citing concerns over the future of Taiwan-China geopolitics and their need to fulfill local customers’ demands better and faster.
Left unexplored, however, is what these U.S. chip vendors have lost over the past two decades by ditching their own IC production.
Wall Street rewarded many U.S. chip vendors for their early decision to go fab-lite or fabless, but those companies now face the inconvenient truth that they left the keys to the future of the chip industry in the hands of Taiwanese. Today, U.S. chip companies can’t prototype and test new breeds of semiconductors without going through Taiwan. More troubling, the next wave of U.S. semiconductor engineering recruits with knowledge of semiconductor integration, advanced packaging, and manufacturability has become a dwindling pool.
Our recent conversation with Arijit Raychowdhury, an EE professor at Georgia Tech, reveals the gaps in the U.S. semiconductor industry and suggests why academia and future semiconductor engineers in the United States need a manufacturing site in their home country.
You can listen to our interview with Raychowdhury here, in which he raised these key points:
- TSMC is already proactive in engaging with U.S. academia. Its engineers spend time with professors and students and allow them to experiment with TSMC’s programs.
- Having a TSMC fab in the United States is important for two reasons: It expands opportunities for productive exchanges between academia and TSMC, and, in the process, it nurtures an expanded domestic workforce in semiconductor engineering and manufacture. “This, in my mind, is the most important thing, because I believe the purpose of academia is to create the future workforce,” said Raychowdhury. “It engages and employs more people in this country in semiconductors.”
- TSMC fabs in the United States create an “intellectual supply chain,” he added. Students go to work in Arizona, come back to the universities, and create a future basis for working together. “Today, we have 10 p.m. meetings with TSMC’s engineers in Taiwan, but it is always much easier if we can do it locally,” said Raychowdhury.
- Engineers in this country today lack domestic capabilities for prototyping, he said. “Assume you need new kinds of packaging technology. Without prototyping and testing capabilities, we don’t know if this is going to be a commercial success. We need to protype a chip to demonstrate new ideas. Or when students want to found startups, they should be able to take advantage of the locally available technical capabilities.”
Local fabs can advance everything from pure research and conceptualization to prototype development and, finally, manufacturing. Maintaining this entire intellectual supply chain locally ensures that the U.S. semiconductor industry has a future.
— Junko Yoshida is the editor in chief of The Ojo-Yoshida Report. She can be reached at email@example.com