STMicroelectronics Raises Its Data Center Revenue Target and Turns AI Demand Into a Power-Silicon Story

STMicroelectronics’ June 2 reset is worth publishing because it sharpens an AI-infrastructure point many markets stories still miss. The useful signal is not simply that another chip company raised expectations. The stronger signal is that ST, a broad-line semiconductor company with deep exposure to power, analog, industrial, and embedded systems, now expects about $1 billion of data-center revenue in 2026 versus a prior expectation of nicely above $500 million.

That matters because the AI semiconductor trade is still too often framed as if the whole story sits inside accelerators. Dense compute racks do not work on GPUs alone. They need power conversion, sensing, control, protection, and a wider set of supporting electronics that help turn expensive silicon into operating capacity. When ST lifts its data-center ambition this sharply, the read-through is that AI capex is broadening deeper into that supporting bill of materials.

ST’s updated target matters because it shows AI spend moving beyond accelerators and deeper into the power-and-control silicon that makes dense compute operable.

The timing strengthens the signal. ST said the increase is based on continued strong AI-infrastructure-led demand and recent progress on capacity ramp-up. That means the company is not presenting the opportunity as distant optionality. It is saying current engagements and manufacturing progress are already strong enough to justify a materially higher near-term revenue target.

For investors, this is useful because it expands the watchlist beyond the obvious names. If AI demand is flowing into power and control semiconductors, then basket construction has to widen as well. “AI chips” is becoming too blunt a category. The better question is which suppliers are becoming indispensable as rack densities rise and data centers need more sophisticated electrical, thermal, and system-management hardware around the core compute.

For operators, the same point matters from the other direction. The more power-dense the rack becomes, the more strategic the supporting silicon around efficiency and control becomes. Those parts are less glamorous than flagship accelerators, but they can matter just as much to deployment schedules if they tighten or scale slower than expected.

This clears the duplicate block because the Grid Report already has coverage on advanced packaging, cooling, and GPU financing, but not this specific supporting-silicon angle. The ST signal is not another NVIDIA read-through or another pure-play AI multiple story. It is a clue that power electronics and adjacent semiconductor layers are moving deeper into the AI capex stream.

The Grid Report view is that ST’s raised target is one of the cleaner recent signs that the AI buildout is reaching further into the electrical guts of the data center. That makes it useful both as a search target and as an investor clue because it answers a better question than “which AI stock is hot?” It shows where the capex map is getting wider.

STMicroelectronics press release, “STMicroelectronics raises its revenue ambition for Data Centers amidst continued strong demand for AI infrastructure,” published June 2, 2026: https://investors.st.com/news-releases/news-release-details/stmicroelectronics-raises-its-revenue-ambition-data-centers