It is hard to imagine exactly how we will know who won the AI arms race. We would need a global consensus about what AI that supersedes human-level intelligence (commonly referred to as artificial general intelligence) looks like, a benchmark that is much harder to pin down than a picture of Neil Armstrong on the moon.
Whatever the case, industry and government officials at the SelectUSA Investment Summit agreed: the US wants, needs, to be first.
Go deeper with GlobalData
A crucial part of winning the AI race in the US involves creating a high-tech manufacturing industry that will build the thousands of chips and data centres needed to power this technology. The creation of such an industry is aligned with one of the Trump administration’s main goals: bringing manufacturing jobs back to the US. However, delegates at the SelectUSA Investment Summit made it clear that this was not an easy task and that the lack of a trained labour force is one of the main obstacles they face. How will the US address its skilled labour shortfall in order to power its AI ambitions?
The will to rebuild industry
Since 1997, “the US lost around five million manufacturing jobs”, US Labor Secretary Lori Chavez-DeRemer told an audience at the summit. “Nearly 100,000 factories closed, and we experienced one of the largest drops in manufacturing employment in history.”
However, that was all about to change, Chavez said.
“From manufacturing to mining and technology to transportation, our industries are coming back to life because President Trump has sent a clear message: America is open for business.”
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataMuch of President Trump’s tariff regime, despite the actual consequences it has had so far, has been justified through this goal. Manufacturing jobs left the US when the world embraced globalisation, and the president argues that high tariffs will force companies to build locally, and therefore bring those jobs back to the US. For this, he has received support from unlikely allies such as United Auto Workers President Shawn Fain and the Brotherhood of Teamsters (one of the US’ biggest unions).
At SelectUSA, state governors from both parties and industry leaders also made it clear that they supported increasing manufacturing with AI integration at its core. Siemens CEO Barbara Humpton said the technological advances in manufacturing reflected an “AI industrial revolution”.
“There are still old notions of what manufacturing looks like. That it is hot and hard and dirty work, but it is cutting-edge technology,” Michigan Governor Gretchen Whitmer said about her state, traditionally the home of US auto manufacturing.
There are various objectives at play. There is the political goal of bringing back manufacturing and the industry goal of making these processes more efficient through AI integration. Underpinning both of these is the need to develop a high-tech manufacturing sector that can sustain it all. Developing sensitive technology independently from China seems to be a national security issue everyone can get behind.
Challenges of creating a workforce
Since 2020, the US has attracted more than $540bn in semiconductor supply chain investments. The CHIPS and Science Act, passed in 2022, has been a major factor for domestic and foreign companies giving grants, subsidies and other incentives to develop the sector. The CHIPS Act is the first major piece of industrial policy the US has passed since the 1950s. It is a notable effort in a country that is not accustomed to long-term planning. However, it means that the US is catching up with other countries that have spent decades developing a workforce with the necessary skills.
Taiwanese companies such as TSMC have been a major source of investment since the passing of the CHIPS Act; the company says its total investment in the US will reach $165bn (T$4.93trn). Even with this policy, however, the US may not necessarily have been an economically logical place for it to build semiconductors. The security guarantee that the US provides for Taiwan also played a role in attracting companies that might have otherwise produced more locally.
“I am not sure that purely on the economics, the US would be a very good bet, even with incentives for building leading-edge fab custody, mostly because of the workforce. We just don’t have a tradition of this highly consequential workforce that you really need for the most advanced chip manufacturing,” W. Patrick Wilson, vice-president of government relations at Taiwanese chip company MediaTek, told Investment Monitor. “We always understood that another component of it was deepening the US-Taiwan partnership.”
Whatever the confluence of factors that has led to vast investment in the semiconductor industry, it has meant that foreign companies are now taking on the challenge of building the necessary workforce.
The Semiconductor Industry Association predicts a deficit of 67,000 workers within the country’s semiconductor industry by 2030. According to GlobalData Strategic Intelligence principal analyst Isabel Al-Dhahir, this gap “is likely to affect not just domestic chip production but also adjacent industries such as data centre construction, potentially resulting in delays and escalating costs”.
At the SelectUSA summit, the need for the US to upskill workers was widely acknowledged by both economic development organisations and businesses with major investments in the US.
One official who is part of the Tech Hubs Programme, highlighted that in her city, they “put industry in the room first” and asked them “what are your needs for your workforce today? What are they tomorrow?” These conversations happen before they contact technical colleges to work together to develop curricula.
Representatives from US and Taiwanese semiconductor companies also expressed frustrations about the skilled labour shortage, particularly as the rise of automation changes the qualifications workers need.
“The area that we really focus on for workforce development is that mid tier. Ours will be the most advanced silicon wafer facility in the world, which means that we are implementing a lot of automation. So, we will have fewer operators. Operators typically have a high school-dependent education, but we will need more technicians with that community college, two-year degree,” said Brent Omhdal, executive vice-president for government affairs at Taiwanese chip company GlobalWafers.
While companies are betting that educational initiatives led by private-public partnerships will create this workforce in the long run, their time horizons do not always conform to the speed the industry wants to move at. TSMC, which announced in March that it would invest an extra $100bn to build five more fabs in the US, has already dealt with the consequences of this asymmetry.
As of January 2025, half of the 2,200 workers at TSMC’s Arizona plant had been brought in from Taiwan. The company realised they could not stay within budget and train a local workforce with the necessary skills on time. This caused pushback from unions, who were promised the plant would create jobs for local workers. The example underscores the difficulties of supercharging the development of an industry without a readily available workforce to match it.
Tech jobs, automation and the AI race
Barry Broome, Greater Sacramento Economic Council CEO, says he witnessed “the destruction of the economic base of the Midwest”, while working in Ohio and Michigan earlier in his career. The problem was not free trade agreements like NAFTA, he argues, it was that “there wasn’t an effort to increase US competitiveness”. When Investment Monitor asked whether the simultaneous push to reshore jobs while integrating AI and automation will bring back as many jobs as some people expect, Broome said: “It won’t.”
“I think it will bring back a lot of jobs. I mean, the fact of the matter is, we need a nationwide movement to upskill people, because you are not going to make a living wage unless you are technically skilled,” Broome noted.
Increased automation, while it might contribute to the growth of the industry and lower costs, would also undermine the goal of job creation.
“Trump’s goals are reshoring and increasing employment in the manufacturing industry,” Beatriz Valle, senior technology analyst at GlobalData, tells Investment Monitor. “Trump is cultivating ties with many Big Tech CEOs whose main goal is to grow automation and robotics, which, in some cases, may replace types of manual work, and there is a clear contradiction between these two goals.”
Wider economic questions also underpin the development of this workforce. Will companies invest more in automation to counteract higher operating costs in the US? What does that mean, in the long term, for the number of jobs that will be created? And, more importantly, will the US be able to pursue all these goals with enough speed to win the global AI race?
