Main US indexes red; Dow off most, down ~0.9%
Materials weakest S&P 500 sector; Energy sole gainer
Euro STOXX 600 index falls ~1%
Dollar gains; bitcoin up ~1.5%, US crude up ~3.5%; gold falls ~2%
US 10-year Treasury yield rises to ~4.45%
Welcome to the home for real-time coverage of markets brought to you by Reuters reporters. You can share your thoughts with us at markets.research@thomsonreuters.com
AI’S NEXT GOLD RUSH IS IN THE PLUMBING, NOT THE PROCESSORS
While the biggest names in AI - Amazon.com AMZN.O, Alphabet GOOGL.O, Meta META.O, Microsoft MSFT.O, and of course Nvidia NVDA.O - still dominate headlines and command trillions of dollars in market cap, the real market action is quietly shifting elsewhere. Investors are increasingly putting their money into smaller, lesser-known companies that are tackling the behind-the-scenes problems holding AI back.
In a note published late Friday, Philip Palumbo, founder, CEO, and chief investment officer at Palumbo Wealth Management, put it this way: the first major AI bottleneck was Nvidia’s chips, which are the “brains” powering artificial intelligence. But AI’s explosive growth has forced data centers to scale up far faster than ever before, and now the limits are physical. Power management, cooling systems, and network speed, or the “brawn” behind AI, have become the real constraints, and they’re fast becoming the focus for both the industry and investors.
The AI arms race has moved beyond simply buying the most advanced chips. It’s increasingly about building enough real-world infrastructure to run them, and that's where hyperscalers are starting to hit hard physical limits. According to Palumbo, that’s why companies solving these problems are emerging as Wall Street’s new AI favorites.
The biggest pressure points include networking and interconnects, power generation and cooling, and the supply chains that feed key components into data centers. And it hasn’t taken long for markets to catch on.
Those bottlenecks are already showing up in stock performance. Power infrastructure players like Vertiv VRT.N, power generation firms such as Bloom Energy BE.N, networking specialists including Astera Labs ALAB.O and Credo CRDO.O, and optical names like Lumentum LITE.O and Ciena CIEN.N have all outpaced the broader market.
Below is a recreation of Palumbo's chart that puts this shift into focus by comparing the year-to-date performance of these stocks to the Magnificent Seven ETF MAGS.K. So far in 2026, the MAGS is only up around 1%, while these stocks have scored gains ranging from more than 20% to more than 225%.
Palumbo's broader takeaway is striking: the global race to dominate AI is unfolding at a pace that hasn't been seen before. Innovations that once took decades are now happening in years, putting enormous strain on physical infrastructure. While software and engineering can scale quickly, things like transformers, factories, and raw materials simply can’t. And that’s why, even as the hyperscalers keep the spotlight, the next wave of AI investment is flowing toward the companies laying the pipes, and solving the physical bottlenecks that make AI possible.
(Terence Gabriel)
*****
EARLIER ON LIVE MARKETS:
FACTORY ORDERS BLAST PAST EXPECTATIONS, WITH WAR MACHINERY, AI SPENDING IN THE DRIVER'S SEAT
CLICK HERE
US STOCKS MIXED AS MIDDLE EAST TENSION RE-ENTERS THE PICTURE CLICK HERE
AI IS BOOSTING US INFLATION, GOLDMAN SAYS CLICK HERE
NASDAQ HITS NEW HIGHS AS MARKET BREADTH SENDS A WARNING CLICK HERE
MAGSvsAIbottleneckstocks05042026 https://fingfx.thomsonreuters.com/gfx/buzz/lgpdgemljvo/image-1777910091327.png
( )