Analysts Ponder Impact of Intel’s $8.5 Billion CHIPS Act Subsidy Leave a comment

//php echo do_shortcode(‘[responsivevoice_button voice=”US English Male” buttontext=”Listen to Post”]’) ?>

Intel’s nearly $20 billion in loans and subsidies from the U.S. government will be the largest boost for a single chipmaker under the CHIPS Act. The sole U.S. memory chipmaker, Micron, will win a smaller, though similar, chunk of the CHIPS money, according to analysts who spoke to EE Times. They expect the world’s chip-tech leaders—Taiwan Semiconductor Manufacturing Co. (TSMC) and South Korea’s Samsung—to win smaller slices of the CHIPS pie.

“The $20 billion total of funding for Intel is a critical catalyst for the U.S. reaching [Commerce Department Secretary] Gina Raimondo’s target of a 20% share of global advanced chip production by 2030,” said Dan Hutcheson, a senior fellow at research group TechInsights. “I’m impressed with the breadth: not only across multiple states, but also that it includes packaging, which is a critical component as advanced chips move from single-die monolithic integration to multi-die polylithic integration. These are not your grandfather’s chips. They are chiplet based.”

The CHIPS Act provides $38 billion in subsidies to revive U.S. semiconductor manufacturing, especially production of advanced chips. To date, the Department of Commerce (DoC) has agreed to provide $35 million to military contractor BAE Systems to expand an existing chip facility in New Hampshire, as well as $162 million to Microchip Technology to help boost its U.S. production of microcontrollers. Both companies make chips for defense equipment.

More recently, the DoC agreed to provide GlobalFoundries (GF) $1.5 billion. The GF and Intel subsidies account for more than a quarter of the $38 billion in CHIPS money.

Micron will win the next big package from the U.S. government given its critical roles as a U.S. firm that’s also one of the world’s top-three memory chip providers, Hutcheson says.

“I expect the next major step will be onto Micron, as high-bandwidth memory (HBM) is another critical part of advanced chip manufacturing. Without it, the packaging part makes little sense. Without Micron, it’s doubtful we can hit the 20% target. After that, the DoC needs to make sure TSMC and Samsung are taken care of if for no other reason than to repay their confidence in America by investing here.”

The Asian companies are building new chipmaking facilities in the U.S., but they are keeping their most advanced R&D and manufacturing tech at home to maintain their geopolitical strength in the chip industry. At the same time, the U.S. wants to draw semiconductor capacity away from Asia given the geopolitical and supply chain risks exposed by the recent covid pandemic.

“America invented these chips,” U.S. President Joe Biden said in prepared remarks this week at Intel’s subsidy announcement in Arizona. “Even though we invented the most advanced chips, we make zero percent of them today. Nearly all manufacturing of leading-edge chips across the entire industry moved overseas to Asia years ago.”

$200 Billion in Global Subsidies

The U.S. CHIPS Act coincides with stimulus efforts in nations like China, Japan, South Korea, Taiwan and India to build more resilient local chip supply chains after semiconductor shortages crippled global car and electronics makers during the Covid pandemic. TechInsights estimates that the total amount of known stimulus funding globally is around $200 billion, or enough to build eight fully equipped gigafabs. That number will still be insufficient to achieve industry growth targets by 2030, according to TechInsights.

“While $200 billion sounds like a lot, it is only an eighth of the total investment needed to meet 2030 semiconductor production requirements,” Hutcheson said.

Even with the CHIPS stimulus, the U.S. is likely to fall short of its goal to make a fifth of the world’s most advanced chips by 2030, according to Paul Triolo, an associate partner at Washington D.C.-based Albright Stonebridge Group, who advises companies in the tech business.

“This funding alone will not be nearly enough,” Triolo said. “That 20% figure is likely only achievable if TSMC and Samsung receive comparable packages to Intel, and there is a second CHIPS Act after 2026.”

Secretary Raimondo has voiced support for a “CHIPS Two.”

“This funding is one step that could lead to Raimondo’s goal, but this alone will not,” Patrick Moorhead, industry analyst, told EE Times. “There will need to be a CHIPS Act 2.0 and 3.0.”

TSMC, Intel and Samsung have encountered major issues in the U.S. with construction costs, qualified contractors and workforce development that will be critical to reaching Raimondo’s 2030 target, according to Triolo.

“TSMC has already announced major delays at its Arizona facilities, and Intel just two weeks ago announced a two-year delay in reaching production at its Ohio facilities,” Triolo said.

He expects that Micron will win a package like what Intel received, leaving smaller amounts of CHIPS funding for other key parts of the supply chain that will be necessary to onshore an electronics ecosystem over the long term. There is a shortage of companies in substrates and materials that are critical suppliers to front-end chipmakers like TSMC ramping up new U.S. fabs in the 2025-2027 timeframe, Triolo noted.

“It is very hard to tell from the outside whether the CHIPS funding is being apportioned in a way that will eventually lead to a complete and sustainable semiconductor ecosystem in the U.S. by 2030,” Triolo said. “Advanced packaging is a good example. While Secretary Raimondo has talked about an end-to-end supply chain for advanced-node semiconductors by 2030, right now, the commercial justification for companies like TSMC to place advanced-packaging facilities in Arizona remains unclear as there is not sufficient volume to justify the high capital expenditure it would take to put a full up CoWoS or other advanced-package facility in Arizona, for example.”

Worker shortage

The U.S. also faces a shortage of qualified semiconductor workers that may persist for decades, according to John Dallesasse, associate dean for facilities and capital planning at the University of Illinois at Urbana-Champaign. The U.S. workforce of around 345,000 today will need to grow to 460,000 by 2030 despite an existing shortage of about 70,000 qualified people, according to Dallesasse.

Openings range from fab technicians to construction workers, he told EE Times.

“Obviously, people with an associate’s degree level to do basic work inside the fabs,” he said, warning that the U.S. industry also needs the expertise to make emerging technology like photonics and wide bandgap devices.

“We have to focus on what’s beyond silicon. How to augment silicon. That’s going to require people at the Ph.D level.”

He worries that U.S. university enrollments are declining, especially in STEM fields. The U.S. talent gap may push the nation toward opening more H-1B visas to overseas nationals, according to Dallesasse.

“You hear industry executives saying we need to make sure that we’re not shutting down access to talented workers from around the globe. We also need to make sure that we’re creating opportunity for other workers in the U.S.”

The U.S. has lost an alarming amount of its electronics industry in the past 40 years, said Dallesasse, who began his career as an electronics engineer during that time.

“I view [the CHIPS Act] as an exercise in ensuring that we don’t lose everything,” he added.