Major technology and infrastructure stocks have experienced significant declines as investors rotate capital away from Magnificent Seven companies and into semiconductor and energy firms. Microsoft, Meta, and Amazon have fallen over 30% from recent highs, while companies like Micron have benefited from accelerated data center spending. Simultaneously, over 75 planned data center projects totaling $130 billion were canceled in the first quarter of 2026, citing electrical grid connection delays and community opposition.
The rotation reflects a maturing AI infrastructure cycle. As initial enthusiasm for AI deployment broadens beyond software and into hardware and energy requirements, investors are reassessing both the timeline and feasibility of large-scale data center expansion. Grid infrastructure constraints and permitting challenges are now recognized as material risks to the pace of AI infrastructure buildout.
Semiconductor and energy companies are capturing durable margin growth from Big Tech's data center capex, signaling healthy underlying demand and a healthy reallocation toward profitable hardware suppliers.
Mass cancellation of data center projects and steep declines across AI infrastructure stocks suggest the pace of AI deployment is slowing faster than expected, and grid constraints may prove more limiting than previously anticipated.
Think of it like this: everyone was rushing to buy AI company stocks because they thought the AI boom would be huge. But now that people are actually trying to build the data centers to power AI, they're hitting roadblocks like power grid shortages and local complaints. So investors are selling the original AI stocks and buying the companies that make the computer chips and power plants instead, because those might actually be the real winners.
Grid connection timelines improve and canceled projects resume, restoring investor confidence in data center plays and potentially halting the rotation away from Mag7 stocks.
scenario – not a predictionElectrical grid upgrades take longer than expected, forcing Big Tech to prioritize existing facilities and slow new builds, keeping data center-related capital allocation depressed.
scenario – not a predictionCompanies deploy AI infrastructure regionally where grid capacity exists, leading to mixed results: some data center suppliers thrive while others struggle, creating sector fragmentation.
scenario – not a prediction