I've made exactly one investment decision in the past two years that felt genuinely contrarian — not contrarian for its own sake, not "I'm buying something nobody wants" just to be different — but contrarian in the way that actually matters: where you can see what the market is missing, and you can wait for it to get priced in.
STMicroelectronics was that decision for me.
I bought at $33. The stock had already fallen more than 60% from its peak. Sentiment was brutal. Revenue was down, margins were getting squeezed, and the automotive slowdown was everywhere in the headlines. I'll be honest with you — I hesitated for weeks before pulling the trigger. But when I finally bought, I was at peace with it in a way I rarely am.
Here's my full story.
Looking for the Ugly Duckling in Semiconductors
I was looking for a stock in the semiconductor space — specifically in the automotive and industrial intelligence theme — but I didn't want the obvious names. I didn't want to chase Nvidia at 40x revenue or own a pure EV play praying on adoption curves. I wanted something that made the infrastructure of the next computing cycle: the chips inside cars, inside factories, inside the sensors and microcontrollers that make machines think.
What I found was ST.
My investment angle was straightforward: STMicroelectronics is one of the only semiconductor companies in the world with serious depth in automotive-grade silicon carbide (SiC), industrial microcontrollers, and — increasingly — AI data-center power and connectivity chips. A company like that, down 60% from its highs, trading at a fraction of its normalized earnings power, started to look like exactly the kind of setup I wait for.
I bought a starter position at $33 and told myself I'd add more if it went lower.
What ST Actually Does (And Why It Matters More Than You'd Think)
STMicroelectronics (NYSE: STM) is a $28 billion Franco-Italian semiconductor manufacturer that most people have heard of but few can place. It's not a pure AI chip story. It's not a memory company. It's what I'd call a workhorse semiconductor business — the kind that doesn't get headlines but quietly powers an enormous slice of the world's electronics.
As per my study, ST operates across four main categories: automotive semiconductors (think SiC power modules for EVs and ADAS chips), industrial automation and energy management, personal electronics (the motion sensors in your iPhone are ST's), and — the newest and most exciting piece — cloud and AI data-center infrastructure chips.
The reason this space matters right now is that two massive tailwinds are running simultaneously. First, the electrification of transportation: every EV needs silicon carbide power modules that can handle high-voltage switching efficiently — and ST is among the top three SiC suppliers globally. Second, the AI infrastructure buildout: hyperscalers like AWS are desperately seeking power-efficient chips for data centers that generate enormous heat and consume staggering amounts of electricity. ST's mixed-signal and power management ICs are exactly what they need.
It's not one tailwind. It's two — and ST sits at the intersection of both.
The Moment That Convinced Me
I was deep in the Q1 2026 earnings transcript when CEO Jean-Marc Chery said something I had to read twice:
"ST is now strategically positioned to capture upside from new AI-driven programs, leveraging specialized technologies to enable the evolving AI infrastructure. We are confirming our data center revenue expectation to be nicely above $500 million for 2026 and well above $1 billion for 2027."
Let me put that in context. When I bought at $33, ST's entire data-center revenue was a rounding error on its income statement. The company was primarily known as an automotive and industrial chip maker. And here was the CEO — on a live earnings call, with real guidance, not vague aspiration — saying they expected to more than double their data-center revenue in a single year, and then double it again the year after.
That doesn't happen by accident. There was a reason behind it that most investors had missed entirely: the AWS deal.
My Investment Thesis — Five Pillars
1. The AWS Deal Is a Multi-Billion Structural Relationship
On February 9, 2026, ST announced a multi-year, multi-billion USD strategic collaboration with Amazon Web Services. This wasn't a small component purchase agreement. AWS designated ST as a strategic supplier of advanced semiconductor technologies across high-bandwidth connectivity, mixed-signal processing, and power management for its hyperscale data centers.
CEO Jean-Marc Chery said at the announcement:
"This strategic engagement establishes ST as an important supplier to AWS and validates the strength of our innovation, proprietary technology portfolio, and proven manufacturing-at-scale capabilities. Our advanced semiconductor solutions will directly power AWS's next-generation infrastructure."
Think about what that means for a company trading at $33 with a market cap below $30 billion. AWS is betting its data-center efficiency roadmap — worth hundreds of billions of dollars in infrastructure — partly on ST's chips. That's not a commodity contract. That's a design-win relationship with one of the most powerful infrastructure builders in the world.
2. Silicon Carbide (SiC) — The EV Infrastructure Play Nobody Is Talking About
ST is in a three-way race with Wolfspeed and Onsemi for dominance in SiC power semiconductors. These chips are non-negotiable for any EV manufacturer that wants to maximize range, charging speed, and thermal efficiency. They're also used in industrial motor drives and renewable energy inverters.
As per my study, the SiC semiconductor market is projected to grow from approximately $2 billion in 2023 to over $10 billion by 2030. ST's SiC revenues in FY2025 were already a meaningful contributor to its automotive segment. And unlike Wolfspeed — which is losing money and burning cash trying to scale its GaN-on-SiC fabs — ST already has vertically integrated SiC wafer supply, meaning it controls its own substrate economics.
That's a competitive advantage that took years to build and is genuinely hard to replicate.
3. The NXP MEMS Acquisition — Buying Exactly the Right Asset at the Right Time
In February 2026, ST completed the acquisition of NXP Semiconductors' MEMS sensor business. This is the unglamorous kind of acquisition that analysts undersell and that pays off for years.
MEMS sensors — accelerometers, gyroscopes, pressure sensors — go into automotive safety systems, smartphones, IoT devices, and increasingly into AI edge computing applications. ST was already a leader in MEMS. By acquiring NXP's MEMS business, it extended that leadership specifically into automotive safety applications where regulatory tailwinds (mandatory ADAS in new vehicles) virtually guarantee demand growth.
The acquired business contributed mid-forties million dollars to Q1 2026 revenue immediately on close — that's not a long integration cycle, it's a plug-and-play revenue addition.
4. Microcontrollers — The 30–50% Growth Engine the Market Is Missing
Here's the part that convinced me the most. Craig-Hallum upgraded ST to Buy specifically calling out that microcontrollers — general-purpose and automotive MCUs — could account for 30–35% of total revenues and potentially grow 30–50% year-over-year as AI edge applications proliferate.
Think about what that means. Every smart sensor, every edge AI device, every connected industrial machine needs a microcontroller. ST is one of the largest MCU makers in the world. If MCU volumes inflect on the back of AI edge adoption — and I believe they will — then this line item alone could drive a revenue rerating that most models haven't priced in.
5. The Restructuring Is a Feature, Not a Bug
When I bought at $33, ST was in the middle of a painful company-wide restructuring — closing or downsizing manufacturing facilities, resizing its cost base. The market hated it. Revenue was down, margins were compressed, headlines were grim.
But as per my study, management confirmed the annual cost savings target is in the high triple-digit million-dollar range exiting 2027. That's potentially $700–900 million in structural cost removal on a revenue base that's now recovering. When that drops to the bottom line, the EPS impact is enormous relative to where the stock was priced.
I was buying the trough. That's always uncomfortable. That's usually where the money is made.
What the Numbers Told Me
What concerned me:
FY2025 revenue of $11.80 billion was down 11.1% year-over-year — and FY2024 had already fallen 23.2% from FY2023's peak of $17.3 billion. Two years of consecutive decline is hard to stomach
GAAP operating income for FY2025 was just $175 million on nearly $12 billion in revenue — an operating margin of barely 1.5%
Gross margin had contracted from 39.3% in FY2024 to 33.9% in FY2025 — meaningful compression
Q1 2026 EPS of $0.04 GAAP missed analyst estimates of $0.17 — not a small miss
What gave me conviction:
Q1 2026 revenue of $3.10 billion came in above mid-point guidance, and Q2 2026 was guided at $3.45 billion — that's +24.9% year-over-year growth. The recovery was accelerating, not decelerating
Despite the EPS miss, non-GAAP operating income was $171 million — showing the underlying business was generating cash even through the restructuring
Gross margin trajectory was recovering: Q4 2025 at 35.2%, Q1 2026 at 33.8% (impacted by acquisition accounting), but Q2 guidance back to 34.8%
Cash position remained solid; balance sheet was not under stress
Analyst consensus of Moderate Buy with an average price target of $49.07 from 13 firms — against a stock I was buying at $33
The numbers told me: this is a cyclical trough, not a structural decline. Those are very different things. And I've learned over the years to make sure I know which one I'm looking at before I invest.
The Hidden Catalyst: ST's Role in AI Is Much Bigger Than Automotive
Here's what I think most people — including most professional investors — are missing about ST.
The narrative around ST has always been "automotive semiconductor company." EV plays. SiC. ADAS. That's the box the market has put it in.
But the AWS deal and the data-center guidance blew that box open. ST's chips do something that is extraordinarily valuable in AI data centers: they manage power with extreme precision at the margins of thermal efficiency where silicon starts to fail. As AI workloads get more intense, the heat generated per rack goes up. The power consumed per server goes up. And the demand for chips that can switch, regulate, and manage that power efficiently — at hyperscale — goes vertical.
ST doesn't just have one product for this. It has a portfolio: high-bandwidth connectivity ICs, mixed-signal processors, analog power management, advanced microcontrollers for intelligent infrastructure management. The AWS deal confirmed that this portfolio is genuinely differentiated — not a "we also have some power chips" footnote, but a multi-billion USD multi-year strategic relationship.
Baird put a price target of $90 on ST and explicitly cited AI infrastructure as a key revenue catalyst by 2027–2028. When I bought at $33, that upside math was striking.
Why I'm Still Holding
The stock has moved from $33 toward the mid-$40s since I bought. I haven't trimmed. I don't intend to.
Here's my honest reason: the thesis hasn't played out yet. The data-center revenue ramp to $500M+ in 2026 and $1B+ in 2027 is still in front of us. The MCU growth acceleration is still in front of us. The full benefit of the restructuring cost savings is still in front of us. When I feel like a thesis is already priced in, I get antsy. Right now, I feel like the market has started to notice ST — but hasn't yet fully appreciated what it's looking at.
The Q2 2026 earnings call in late July will be the next important checkpoint. If automotive revenues are up low double digits and industrial revenues are up mid-20% as guided — and the data-center business is tracking toward that $500M+ full-year target — I expect the stock to move materially higher.
I'll be watching that print very closely.
What I'm Watching Next
Green flags that would increase my conviction:
Q2 2026 revenue at or above the $3.45 billion guidance midpoint
Data-center revenue explicitly disclosed and tracking above $125M for the quarter (implying $500M+ run rate)
Gross margin recovery above 35% — signaling the restructuring is flowing through
Any additional hyperscaler contract announcement beyond AWS
SiC design-win announcements from major EV OEMs in Europe or North America
MCU bookings inflecting — management commentary on order strength here would be significant
Red flags that would make me reconsider:
A second consecutive EPS miss combined with guidance cut — would suggest the recovery is weaker than the revenue line implies
Automotive revenue declining further despite the low-double-digit sequential recovery guided for Q2
Any deterioration in the AWS relationship or loss of a major engaged customer program
Gross margin failure to recover — staying below 34% through H2 2026 would suggest structural, not cyclical, margin impairment
Tariff escalation specifically targeting European semiconductor imports to the US — ST manufactures heavily in Europe and this is a real, if underappreciated, geopolitical risk
The Bottom Line
Let me summarize why I bought STMicroelectronics at $33, in plain language:
A world-class chip maker at a trough valuation. ST makes the silicon for EVs, factories, AI data centers, and billions of connected devices. At $33, the market was pricing it like the decline was permanent. As per my study, it was cyclical.
Two massive tailwinds, one stock. SiC for automotive electrification and power chips for AI infrastructure are both genuine multi-year growth stories. Most investors are chasing them separately in expensive pure-plays. ST gives you both at a discount.
The AWS deal changed the narrative. A multi-year, multi-billion USD strategic engagement with one of the world's largest infrastructure builders is not something you see every day. It confirmed that ST's data-center capabilities are real — and that the market had badly misjudged them.
The restructuring is a tailwind in disguise. High triple-digit millions in annual cost savings exiting 2027, dropping onto a recovering revenue base, should drive meaningful EPS expansion that most current models are underestimating.
Analyst targets of $49–$90 vs. my entry at $33. That's the kind of asymmetry I look for when I invest.
I won't pretend this is a sure thing. Semiconductors are a cyclical business. The tariff environment is unpredictable. The auto market has been soft for longer than almost anyone expected. These risks are real and I've sized my position accordingly.
But when I look at where ST is today — with an AWS partnership, a recovering revenue trajectory, a data-center business that management is now explicitly guiding to exceed $1 billion by 2027, and a stock that's still well below its 52-week high — I feel like I'm holding something the market has undervalued. And I'm comfortable waiting for that to change.
The author holds a long position in STMicroelectronics N.V. (NYSE: STM) as of the date of publication. All financial data sourced from ST's official SEC filings, Q1 2026 earnings press release, and publicly available analyst research. This article is written for informational purposes only and does not constitute financial or investment advice. Always do your own due diligence before making investment decisions.