AI boom can withstand a development slowdown
Add Axios as your preferred source to
see more of our stories on Google.

Photo illustration: Sarah Grillo/Axios. Photo: Aaron Schwartz/UPI/Bloomberg via Getty Images
President Trump is pushing back hard against calls to slow AI development, casting it as a powerful economic growth engine.
- But sustaining that economic boom in the near term may not require ever-more-powerful AI models.
Why it matters: The economic payoff in the months ahead may depend less on the next generation of models than on continued infrastructure investment and broader adoption of the AI tools that already exist.
- AI now underpins a huge share of business investment and stock market gains, while hopes for faster productivity growth rest heavily on the technology.
Driving the news: Trump is railing against calls to slow AI development after AI frontier-lab leaders proposed slowing development of their most advanced models over mounting safety concerns.
- "AI, and Data Centers, will be the Greatest Economic Development Engine in History," Trump posted on Truth Social, predicting an impact "bigger than Oil, Gold, Diamonds, or even the Internet."
- Trump's allies argue that if the labs want to slow down, they can do so on their own, without the blessing of Washington.
The big picture: AI's economic impact is taking shape through two distinct, potentially staggered channels.
The first is the infrastructure buildout already meaningfully contributing to economic growth.
- There's a "strong pipeline of infrastructure projects in place" as companies race to secure the energy and computing capacity needed to meet demand for existing AI tools, ING chief international economist James Knightley tells Axios.
- AI-related spending has added an average 0.4 percentage point to annualized GDP growth since 2025, according to Morgan Stanley chief U.S. economist Michael Gapen.
- Bank of America analysts said that AI networks are being fully utilized and rental rates are rising for even older-generation chips as signs that demand remains robust. In a note yesterday, they called the economic stakes "too large for any sustained meaningful deceleration."
Yes, and: Salesforce president Patrick Stokes said a slowdown could actually be good news for builders, who "have a lot of catching up to do," Axios' Ina Fried reports.
The second is the hoped-for productivity boost, with AI potentially helping businesses produce more at lower cost — a payoff much less developed than the infrastructure boom.
- A slowdown "need not trigger an equivalent slowdown in broader tech-related economic activity," Mohamed El-Erian, Allianz chief economic adviser, wrote on X yesterday.
- He noted that many commercial uses don't require frontier models as businesses still figure out how to deploy capabilities that already exist.
- "There's a lot of focus on the hardware capex cycle," Goldman Sachs economist Joseph Briggs told Neil this morning at an event hosted by the Business Roundtable. "What I think is less appreciated is how much is being invested ... in softer forms of investment at the company level," including strategy and data needed to deploy AI effectively, Briggs said.
What to watch: A slowdown at the AI frontier could still hit the economy indirectly if it rattles financial markets. (AI chip stocks fell nearly 6% yesterday, compared with a much smaller decline in the broader market.)
- "An equity market correction might make [high-income households] more cautious, risking a pullback in spending," Knightley says, noting that the AI-fueled stock market boom has helped drive consumer spending among affluent households.
- "With so many of these projects now debt-financed, a drop in valuations could tighten financial conditions" and ultimately curtail the infrastructure buildout over the medium term, Knightley adds.
