---
title: "STMicro revenue outlook shows AI chip demand spreading"
author: "Avery Lin"
datePublished: 2026-07-23T05:56:00.000Z
canonical: "https://scramnews.com/post/00tilpc0zib0h/stmicro-ai-boom-revenue-growth-outlook-2026"
---

[STMicroelectronics](https://investors.st.com/static-files/bdacd978-231f-4e8c-bfde-f5cc6c53250d) forecast about $3.70 billion of third-quarter revenue on Thursday and said fourth-quarter sales should top $4 billion. The point was narrow but useful: artificial-intelligence data-centre demand is starting to reach a European chip supplier, not only the usual hyperscaler-adjacent winners.

Investors still have to decide how much quality sits behind that growth. STMicro reported $3.49 billion of second-quarter revenue and a 34.8 per cent gross margin, then lifted its 2026 data-centre ambition to [above $1 billion](https://newsroom.st.com/media-center/press-item.html/c3396.html). Profitability is thin, cyclical end-markets remain uneven and the new growth engine is smaller than the businesses it is trying to offset.

That gap explains the cautious read. [Bloomberg’s reporting](https://www.bloomberg.com/news/articles/2026-07-23/stmicro-forecasts-revenue-growth-as-ai-boom-boosts-chipmaker) and [Finimize’s analysis](https://finimize.com/content/stmicroelectronics-sees-demand-return-but-its-q3-guide-fell-short) both centred on the softer quality of the recovery, not just the headline beat: the current-quarter guide was solid enough to keep the story alive, but not strong enough to settle whether AI is broadening STMicro’s base or merely masking weak auto and industrial demand.

Chief executive Jean-Marc Chery cast the quarter as a beat driven by a business mix already in motion in the [earnings release](https://investors.st.com/static-files/bdacd978-231f-4e8c-bfde-f5cc6c53250d).

> “Q2 net revenues came above the mid-point of our business outlook range, driven by higher revenues in CECP and Automotive.”
>
> — Jean-Marc Chery, STMicroelectronics

## Where the AI growth is showing

The AI case starts in the plumbing around compute clusters. In the [company’s data-centre update](https://newsroom.st.com/media-center/press-item.html/c3396.html), STMicro described a revenue stream tied to 800V direct-current power, silicon photonics and optical interconnect, the less glamorous parts of the stack that become critical when operators cram more compute into each campus.

![Server racks in a data centre, illustrating where power and optical components sit beneath AI computing clusters.](https://images.pexels.com/photos/4508751/pexels-photo-4508751.jpeg?auto=compress&cs=tinysrgb&dpr=2&h=650&w=940)

This is where the builder-optimist camp matters. [CNBC reported in May](https://www.cnbc.com/2026/05/29/nvidia-photonics-investment-ai.html) that Nvidia is spending heavily on photonics to move data between GPUs, memory and networking gear, while separate coverage from [CNBC on UMC](https://www.cnbc.com/2026/07/14/umc-starts-mass-production-in-singapore-citi-sees-improving-outlook.html) and [Tom’s Hardware on Tower Semiconductor](https://www.tomshardware.com/tech-industry/semiconductors/tower-semiconductor-revives-shuttered-panasonic-era-fab-in-3-billion-japan-photonics-expansion) showed suppliers racing to add photonics capacity. STMicro is pitching the bottlenecks around the GPU, where power conversion and optical traffic can become content winners if AI clusters keep scaling.

June’s record makes this harder to dismiss as a one-quarter talking point. As [Bloomberg reported on June 2](https://www.bloomberg.com/news/articles/2026-06-02/stmicro-raises-estimates-for-data-center-revenue-to-1-billion), STMicro had already nearly doubled its 2026 data-centre revenue view to $1 billion. Thursday’s update pushed the 2027 ambition to well above $2 billion. Management keeps returning to the same trio of programmes, 800V power, photonics and optical interconnect, because those are the lines now winning committed customer budgets.

Concentration risk remains. STMicro’s own description of the opportunity points to engaged programmes rather than a fully diversified order book, including an [AWS-linked AI stack](https://newsroom.st.com/media-center/press-item.html/c3396.html) and low-Earth-orbit communications. That is better than a concept slide, but it also means the market is underwriting a handful of design wins rather than a broad secular rerating. Additive growth is real; dominant growth is still a future tense.

Market context helps the demand case and complicates the valuation case. [CNBC’s survey of European AI winners](https://www.cnbc.com/2026/05/21/ai-frenzy-european-stocks-rally-nokia-aixtron.html) showed investors rewarding companies tied visibly to infrastructure buildout, but more cyclical industrial chip names still had to prove they were more than temporary beneficiaries. STMicro’s update places it in the overlap: not a pure AI winner, but no longer just an auto-and-industrial recovery bet either.

Europe’s own chip base makes that distinction sharper. As [The Register noted in June](https://www.theregister.com/systems/2026/06/11/dutch-chip-startup-claims-all-european-fab-flow-with-help-from-a-very-american-friend/5254023), much of the region’s semiconductor demand still comes from automotive and industrial applications. STMicro is trying to add a higher-growth AI layer on top of mature end-markets, not replace them outright. If that mix shift works, the company gets optionality. If it stalls, investors are left with the old cycle carrying a richer narrative multiple.

## Why margins still matter

The margin line is why the income statement still matters more than the slogan. Q2 revenue rose 26 per cent year on year, yet operating margin was only 5.4 per cent and the quarter still absorbed restructuring and unused-capacity charges, according to [STMicro’s results](https://investors.st.com/static-files/bdacd978-231f-4e8c-bfde-f5cc6c53250d). A company can grow out of a cycle and still fail to rerate if each dollar of recovery arrives with too much baggage.

![Fiber-optic cables connected to network equipment, a visual stand-in for the optical interconnects chip suppliers are racing to supply.](https://images.pexels.com/photos/2420212/pexels-photo-2420212.jpeg?auto=compress&cs=tinysrgb&dpr=2&h=650&w=940)

For skeptics, the issue is straightforward. If the AI line is already as important as management says, why did the current-quarter guide still land below what the bullish camp wanted? [Finimize argued](https://finimize.com/content/stmicroelectronics-sees-demand-return-but-its-q3-guide-fell-short) that the revenue recovery was real but the outlook fell short, and [Bloomberg’s account](https://www.bloomberg.com/news/articles/2026-07-23/stmicro-forecasts-revenue-growth-as-ai-boom-boosts-chipmaker) likewise left investors focused on whether the company’s top-line acceleration is broad enough to overcome margin drag. The answer may be that AI is helping the mix before it is helping the whole model. Encouraging, but not a full-cycle turn.

Management wants the market to look one quarter further out. In the [results statement](https://investors.st.com/static-files/bdacd978-231f-4e8c-bfde-f5cc6c53250d), Chery tied the expected fourth-quarter acceleration directly to customer programmes already in hand:

> “We anticipate a revenue growth acceleration in Q4, mainly driven by our engaged customer programs in AI datacenters and LEO satellite communication.”
>
> — Jean-Marc Chery, STMicroelectronics

Sequencing is the partial answer. Revenue can inflect before margins do when a manufacturer is still carrying underused capacity from the previous downturn. For STMicro, the bullish read depends less on whether AI demand exists, it plainly does, and more on whether that demand arrives fast enough to absorb the fixed-cost drag still attached to the rest of the portfolio. That is a narrower bet than the phrase AI boom suggests.

Chery made the same case more bluntly in the [data-centre release](https://newsroom.st.com/media-center/press-item.html/c3396.html), saying strong AI data-centre demand was enough for STMicro to lift its ambition again. The line matters because it signals management sees the upside in committed programmes, not just in a better macro tape. It also sharpens the standard against which the next two quarters will be judged.

## What has to happen next

Valuation discipline is the analyst debate now. [Morningstar’s Brian Colello](https://www.morningstar.com/stocks/stmicro-company-lifts-revenue-forecasts-amid-ai-boom) asked, in effect, whether a business moving from above $1 billion of data-centre revenue this year to well above $2 billion in 2027 deserves a rerating if gross-margin and medium-term profitability goals still look stretchier than the revenue story. That is the cleanest version of the bull-bear split. The bull case says STMicro has found a second engine. The bear case says the engine is still attached to a soft legacy chassis.

Second-half execution matters more than the quarter itself. Management is guiding to more than $4 billion of fourth-quarter revenue, a clear step up from the roughly $3.70 billion expected for Q3. If that bridge closes, investors can start to argue the AI programmes are scaling through the income statement rather than just through backlog language. If it does not, the story reverts to an ordinary cyclical rebound with an unusually fashionable customer set.

Across semiconductors, the tape offers both support and caution. [Micron’s June results](https://www.cnbc.com/2026/06/26/micron-stock-earnings-tech-selloff-ai.html) showed hyperscalers are still spending aggressively on AI memory, while [Onsemi’s move for Synaptics](https://siliconangle.com/2026/06/25/analog-chipmaker-onsemi-buys-synaptics-7b-stock-deal-push-physical-ai/) underlined how analog and power chipmakers are trying to reposition around physical AI rather than pure handset or auto exposure. STMicro fits that second camp more than the first. Its win condition is not to become another GPU proxy, but to prove that the plumbing around AI clusters can be a durable profit pool in its own right.

Thursday’s report matters beyond one quarter because STMicro has put real numbers around the idea that AI demand is spreading into Europe’s power and connectivity suppliers. What it has not yet shown is that those wins can carry the whole company, not just cushion the weak spots. Until margins start following revenue and the AI line looks less programme-led, the stock remains a test of breadth, not a settled verdict on the chip cycle.