The European chipmaker reported Q2 2026 revenue guidance of $3.45 billion, up 25% year-over-year, with AI data center demand now driving a target that doubled to $1 billion in under six months.
Most investors hunting AI chip exposure went straight to Nvidia, then AMD, then maybe a handful of memory names. STMicroelectronics barely registered. That was a mistake. The Geneva-based semiconductor company reported its Q2 2026 results on July 23, guiding to revenues of approximately $3.45 billion, a 24.9% year-over-year jump, with a gross margin of around 34.8%. The number driving that growth isn't automotive or industrial, the two segments that built ST's reputation. It's data centers, and the ramp is moving faster than the company's own forecasts.
In June, ST doubled its full-year 2026 data center revenue target from "nicely above" $500 million to approximately $1 billion. That revision came less than a quarter after the original number was set. Management now says that if the current order environment holds, data center revenues could double again to roughly $2 billion by 2027. Book-to-bill is well above 1.0 across all end markets and all regions, and distribution inventory, which was a significant drag through much of 2025, has normalized.
The catalyst behind the numbers is a multi-year, multi-billion-dollar commercial agreement with Amazon Web Services announced in February 2026. That partnership positions ST as a strategic supplier across high-bandwidth connectivity, mixed-signal processing, advanced microcontrollers for infrastructure management, and analog and power ICs for hyperscale energy efficiency. The agreement also includes share warrants linked to purchase volumes, which structurally ties ST's revenue trajectory to AWS's capital spending on AI compute. That's not a typical supplier relationship. It's something closer to a co-investor arrangement.
What ST makes is worth understanding here. This isn't a GPU company. STMicroelectronics specializes in silicon carbide power semiconductors, microcontrollers, and analog chips: the unglamorous components that control power conversion, manage thermal loads, and connect systems in a data center rack. Every hyperscale facility running AI inference or training at scale needs them in volume. Nvidia gets the headlines; ST gets the order book.
The AWS deal gives ST something that most chip suppliers spend years chasing: long-dated demand visibility. When a hyperscaler commits multi-year, it makes capital allocation decisions easier and capacity planning more aggressive. That confidence is almost certainly what allowed ST to revise its data center guidance upward so quickly after setting it. The company isn't guessing at the ramp. It has contracted demand underpinning the forecast.
Why the market hasn't fully priced this in yet
Here's the thing: Q1 2026 results beat estimates, and the stock barely moved. That pattern, a beat followed by a muted reaction, usually signals one of two things. Either investors are skeptical that the new growth driver is durable, or the narrative hasn't spread far enough yet to attract the buyers who would reprice the stock. In ST's case, it looks like both.
STMicroelectronics has spent the better part of two years dealing with an ugly automotive semiconductor correction. Inventory gluts, order cancellations, and a prolonged reset in the electric vehicle supply chain hammered the stock from its 2024 highs. The company saw net income fall 89% at one point during the downturn, as Bloomberg noted in its coverage of the data center guidance revision. That context trained the market to view ST as a cyclical chipmaker, not an AI infrastructure play. The pivot to data centers hasn't yet changed that framing in the minds of enough investors.
But the framing is wrong, or at least increasingly incomplete. As Bloomberg's analysis in June made clear, the AWS engagement is not a one-quarter bump. It's a structural realignment of what ST sells and to whom. Power management and mixed-signal chips for AI data centers aren't a commodity addition to ST's catalog. They require the same precision manufacturing and thermal expertise that ST has applied to automotive for decades. The company is moving its most defensible capabilities into a faster-growing market with a committed anchor customer.
The consensus estimate for Q2 adjusted earnings per share sits at 26 cents, a modest figure given the revenue scale, but margin recovery takes time after a deep downturn. What matters more at this stage is the revenue trajectory and the quality of the order book. On both counts, the Q2 report looks like confirmation, not a turning point still to come.
At its current valuation, ST is still being priced largely as a cyclical industrial and automotive chipmaker catching a lucky break. The AWS deal and the data center ramp suggest a different business is quietly taking shape beneath that label. The investors who figure that out before the next guidance revision will have been early. Everyone else will say they should have seen it coming.
Also read: Nokia beats Q2 2026 earnings estimates as AI data center demand rewrites the company's story • Intel and AMD are locking Chinese AI data centers into multi-year CPU deals as a 40% price surge signals a shortage nobody saw coming • OpenAI's own AI models broke out of a test sandbox and autonomously hacked Hugging Face
