The Overlooked Chips Powering the AI Boom
Artificial intelligence is stoking demand for power semiconductors. But Chinese competition, surprise shortages and lackluster investment are looming risks.
Profit and prosper with the best of Kiplinger's advice on investing, taxes, retirement, personal finance and much more. Delivered daily. Enter your email in the box and click Sign Me Up.
You are now subscribed
Your newsletter sign-up was successful
Want to add more newsletters?
To help you understand the trends surrounding business and technology and what we expect to happen in the future, our highly experienced Kiplinger Letter team will keep you abreast of the latest developments and forecasts. (Get a free issue of The Kiplinger Letter or subscribe.) You'll get all the latest news first by subscribing, but we will publish many (but not all) of the forecasts a few days afterward online. Here's the latest…
Power semiconductors are surfing the wave of soaring AI electricity usage. Rather than moving and controlling data, power chips move and control electricity in each step that goes into converting high-voltage power from a power plant to the current that AI chips use.
"Power underpins everything in modern technology," said Leonard Shtargot, a fellow at Analog Devices, in a recent panel that brought together top industry executives. The skyrocketing amount of electricity that AI data centers use “all has to go through power semis one way or another,” said Shtargot.
Global yearly revenue for power chips will jump to $100 billion by 2029, up from about $80 billion in 2026, according to Omdia. The yearly growth rate of 7% or so is faster than previous years, with the AI segment seeing even faster growth.
Power chips are also used in EVs, smartphones, laptops, robotics, motors, wireless gear, military radar and much more. Many of the chips last for decades and have been traditionally treated as commodities, which has meant the sector hasn’t seen as much change or investment as other advanced chips, such as central processing units or graphics processing units.
Advanced power chips use high-performance, but costlier, materials such as silicon carbide and gallium nitride. The industry is also researching newer materials such as gallium oxide and diamond.
The AI industry needs power chip innovation to support the AI buildout. "I think the entire ecosystem has to evolve to meet the demand for AI end users," said Dinesh Ramanath, senior vice president at Onsemi, during the panel. He expects significant innovations in the AI market, especially in power density, or how many watts can be squeezed in a certain amount of space on a chip.
Germany’s Infineon Technologies is the industry leader with about 45% market share. The rest of the market is fragmented, with dozens of companies. Major vendors include Onsemi, STMicroelectronics, Mitsubishi Electric, Texas Instruments and Analog Devices. A big focus for chipmakers is getting power to AI chips more efficiently. For example, 10% to 15% of the high-voltage power that goes to a GPU is lost as heat, which is "not acceptable," said Jeff Halbig, product marketing manager at STMicroelectronics, during the panel.
One pressing concern is China’s push to win more of the market, including by relying more on its domestic suppliers. China already nabs 40% of global sales and Beijing is "focused on building up an ecosystem of emerging power chipmakers," noted MorningStar analyst Brian Colello in a research report last year. Other worries include cost pressure stemming from intense competition, demand volatility and worker shortages.
One way to deal with China’s effort to dominate the market is by boosting U.S.-based production, but that’s easier said than done. Industry players want more federal backing, including higher spending on research. They are also calling for efforts to ensure that chip shortages don’t become a surprise AI bottleneck.
Paul Pickering, research director at Omdia, used his presentation as a reminder about chip chaos during COVID: Cars were stuck in parking lots because they couldn’t get basic $1 chips. He wonders if in, say, 2031, there could be a data center ready to switch on, except for a missing $1 power chip and suggests that this hypothetical scenario is something that policymakers and companies can mitigate by acting early.
This forecast first appeared in The Kiplinger Letter, which has been running since 1923 and is a collection of concise weekly forecasts on business and economic trends, as well as what to expect from Washington, to help you understand what’s coming up to make the most of your investments and your money. Subscribe to The Kiplinger Letter.
From just $107.88 $24.99 for Kiplinger Personal Finance
Become a smarter, better informed investor. Subscribe from just $107.88 $24.99, plus get up to 4 Special Issues
Sign up for Kiplinger’s Free Newsletters
Profit and prosper with the best of expert advice on investing, taxes, retirement, personal finance and more - straight to your e-mail.
Profit and prosper with the best of expert advice - straight to your e-mail.
Related Content
- What a Time to Run This T. Rowe Price Tech Fund
- 17 Ways to Cut Your Energy Bill
- AI Race for Memory Chips Drives High Prices for Tech
- Why AI Superiority is Measured in Gigawatts
Profit and prosper with the best of Kiplinger's advice on investing, taxes, retirement, personal finance and much more. Delivered daily. Enter your email in the box and click Sign Me Up.

John Miley is a Senior Associate Editor at The Kiplinger Letter. He mainly covers AI, technology, telecom and education, but will jump on other business topics as needed. In his role, he provides timely forecasts about emerging technologies, business trends and government regulations. He also edits stories for the weekly publication and has written and edited email newsletters.
He holds a BA from Bates College and a master’s degree in magazine journalism from Northwestern University, where he specialized in business reporting. An avid runner and a former decathlete, he has written about fitness and competed in triathlons.