STMicroelectronics: EPS Beat Masks a Revenue Miss, but the Misprice Is Capacity, Not Demand

The actionable read from this print is that STM is no longer just a demand-recovery story, and that distinction matters for the stock. What was priced in was a cleaner revenue catch-up to the Street’s $3,291.8 million estimate, especially after the sharp sequential recovery that began in Q2 FY2025. What actually surprised was the mix of a $3,173.3 million revenue shortfall against that estimate and a $0.29 EPS outcome versus the Street’s $0.22, a combination that says investors should spend less time debating whether the cycle has turned and more time debating how much of the cost base can be absorbed as revenue moves back toward the mid-$3 billion run-rate. The variant perception is that the print is not a low-quality earnings beat caused by one-off cost suppression; it is an early sign that STM can generate operating leverage before revenue fully recovers, even though the absolute gross margin remains depressed at 33.2%. That is not a call that the business is healthy. It is a call that the market may be missing the inflection from falling volumes and stranded capacity toward modest revenue growth and controllable investment.

The distinction between what was expected and what mattered is clear in the two bases of reporting. On the Street-comparison basis, revenue was $3,173.3 million against $3,291.8 million, so the company missed by -3.6%; EPS was $0.29 against $0.22, so it beat by +31.8%. On the company’s own account, Jean-Marc Chery framed the quarter differently, saying STM “delivered revenues at $3.19 billion $17 million above the midpoint of our business outlook range,” which matters because management is anchoring progress to its internal recovery path rather than to the Street’s more optimistic top-line setup. Those two views can coexist: the sell-side had moved too high on revenue, while the company still executed ahead of its own guide. The stock debate should not collapse those into a simple beat or miss, because the miss tells us expectations were ahead of end-demand, while the EPS beat tells us the downside case on profitability was too harsh.

That is why the revenue trajectory is less negative than the headline miss suggests, but also less clean than a cyclical rebound usually looks. Revenue has recovered from the Q1 FY2025 trough of $2,519.9 million to $3,173.3 million, yet it remains below the $3,251.0 million level of Q3 FY2024. The key is that sequential growth of +13.2% did not restore year-over-year growth, which was still -2.4%, so the business is improving off a depressed base rather than accelerating into a broad-based expansion. The market was priced for a faster crossing of that line; the print says the crossing is now more likely to come through Q4 volume and absorption than through a Q3 demand upside.

The margin line explains why that slower revenue repair still matters. Gross margin has been compressed from the high-40s in FY2023 to 33.2% in Q3 FY2025, but the Q4 guide calls for “about 35%,” per Chery, and explicitly includes “about 290 basis points of unused capacity charges.” That wording is more important than the number itself because it isolates the remaining drag: STM is telling investors that even with a material stranded-capacity burden, gross margin can move higher from Q3. In other words, the path back is not dependent on a heroic revenue beat; it depends on the combination of moderate revenue recovery, mix stabilization, and fewer self-inflicted manufacturing penalties. The bear case would be that 33.2% gross margin is simply the new structural level. The guide argues against that, not because 35% is good, but because it embeds a named headwind and still points up.

The capacity story also reconciles the company’s operating-expense discipline with the EPS surprise. Gross profit was $1.06 billion, down 13.7% year over year, while net operating expenses excluding restructuring were $842 million and described by Lorenzo Grandi as “broadly stable on a year-over-year.” That combination would normally leave little room for upside, yet GAAP diluted EPS was $0.29 on the call basis and the Street-comparison EPS beat was +31.8%. The interpretation is that the Street underestimated how much earnings could recover as revenue stepped off the trough and as below-the-line or charge dynamics normalized. But the quality bar is not fully cleared: operating income included $37 million for impairment restructuring charges and other related phase-out costs, so the print is still carrying the cost of resizing the footprint. The thesis is not that STM has fixed the margin structure; it is that the earnings floor is higher than expected while the company is still paying to repair it.

The product signal inside the quarter supports a selective recovery rather than a broad demand inflection, and that matters for how PMs should size conviction. Personal Electronics was the upside bucket, with Chery saying third-quarter revenues were “up 40% sequentially” on seasonality in engaged customer programs and increased silicon campaigns. By contrast, Automotive and Industrial performed as anticipated, and Grandi said Industrial increased by about 13% while Communication Equipment and Computer Peripheral increased by about 7%. The important read is not that Personal Electronics is suddenly the new core driver; it is that STM’s recovery is being pulled by program timing and seasonal silicon campaigns while the higher-quality automotive and industrial legs are not yet giving investors a positive demand surprise. That makes the revenue miss more understandable, but it also means the next quarter needs to show that the recovery is not just a Personal Electronics calendar effect.

This is where the Q4 guide becomes the centerpiece of the investment case. Chery guided revenues at $3.28 billion, an increase of 2.9% sequentially, plus or minus 350 basis points, and a gross margin of about 35%. Those are not aggressive numbers relative to the prior revenue base, but they are enough to test the absorption thesis because the revenue guide is not far above Q3 while the gross margin guide is visibly higher. If STM can lift gross margin with only 2.9% sequential revenue growth, the market will have to reprice the degree of operating leverage embedded in the footprint. If it cannot, the EPS beat will look more like a transient accounting outcome than a change in earnings power. The setup is therefore asymmetric around gross margin rather than revenue: Q4 does not need a large top-line beat to confirm the bull case, but it does need the 35% guide to hold.

CapEx reinforces the same point because management is no longer funding the downturn as if every capacity plan remains sacred. Chery said STM reduced its net CapEx plan to “slightly below $2 billion for full year 2025” from a prior range of “$2 billion to $2.3 billion.” That is a tangible shift in capital discipline, and it matters for suppliers as much as for STM’s free cash flow. Third-quarter net CapEx was $401 million versus $565 million in Q3 2024, while free cash flow was positive at $130 million versus $136 million. The second-order implication is a mixed signal for equipment and materials vendors: the spending envelope is lower, but not collapsing, and STM is still generating cash while resizing. For Aehr Test, Oxford Instruments, Centrotherm, E&R Engineering Corp., and Veolia Water Technologies, the lowered full-year plan caps near-term order urgency. For Soitec, TanKeBlue, and Wolfspeed, the more relevant read is that STM’s silicon-carbide and substrate commitments are being managed inside a lower capital envelope, with the Wolfspeed agreement still flagged in the supply chain as a >$800M agreement. For Tower Semiconductor, the TPSCo JV read-through is less about an immediate wafer surge and more about STM preserving optionality while China-for-China and mixed-signal sourcing remain part of the footprint.

The inventory and balance-sheet data make the margin thesis more credible, because the company is not trying to buy time with working-capital deterioration. Inventory at the end of Q3 was $3.17 billion, down about $100 million from the end of Q2, while net cash from operating activity was $549 million. STM also ended September with a $2.61 billion net financial position, supported by $4.78 billion of liquidity and $2.17 billion of financial debt. Those figures are not a reason to pay any multiple; they are a reason to believe management can continue reducing capacity costs without being forced into abrupt cuts that damage customer commitments. In semis, trough margins often look worst when inventory is bloated and cash conversion is deteriorating. Here, the inventory direction and free-cash-flow outcome give management room to execute a slower, cleaner recovery.

The peer comparison keeps the optimism bounded. In the IDM set, STM’s latest peer-table revenue growth of +22.8% and gross margin of 33.8% sit in an unusual place: better revenue growth than NXP Semiconductors at +12.2%, but far below NXP’s 56.2% gross margin. Texas Instruments is even more extreme, with 58.0% gross margin and +18.6% revenue YoY. That gap is the reason the stock should not be valued as if the recovery is already normalized. The upside case is not that STM suddenly resembles TXN or NXP in profitability; it is that STM’s depressed gross margin gives the stock a larger self-help lever if unused-capacity charges roll off and Q4 confirms the 35% level. The relative setup is therefore about margin catch-up from a low base, not quality convergence.

The call delivery supports that interpretation, with one important caveat. The tone history shows Q3 FY2025 sentiment at -0.03 and guidance_tone at 0.22, which is consistent with a management team describing recovery while still carrying mix and capacity headwinds. More interestingly, tone_confidence rose to 0.77 while uncertainty fell to 28.9, a combination that suggests the message was more controlled than in the prior quarter. The caveat is that prepared_sentiment was only 0.01 and qa_sentiment was -0.08, so the positivity was not broad-based across the call. The market should not read management tone as euphoric, but it should note that management sounded more specific about the levers that matter, especially unused capacity, CapEx, and Q4 gross margin.

That tone pattern matters because the most investable semiconductor turns usually come when management gets more precise before the P&L looks clean. Here, the specificity is around the manufacturing transition rather than a blanket demand claim. The Q&A exposed the right investor concern, with Didier Scemama referring to a prior split of “70 basis points of FX headwinds” and “70 basis points” tied to the manufacturing transition from 6 to 8 and 8 to 12. STM did not give investors a clean all-clear on that issue in the provided excerpts, and that is the main tension in the bull case. Gross margin is guided up, but the company is still carrying unused capacity charges and transition costs. If those costs prove sticky, the recovery stalls around the mid-30s gross-margin level. If they fade with even modest revenue growth, the EPS leverage can exceed what the Q3 revenue miss implies.

The cleanest way to frame the stock after this print is therefore not “buy the beat” or “sell the miss.” The print says the market had priced in too much Q3 revenue, but not enough earnings resilience and not enough chance that gross margin has begun to turn before a full demand recovery. That is a subtle but tradable variant perception. It favors investors willing to underwrite a Q4 margin confirmation over investors waiting for a perfect top-line inflection. It also argues against using the -3.6% revenue miss as the sole signal, because the company’s own $3.19 billion revenue commentary was above its midpoint and the Q4 guide still points to sequential growth. The risk is that the revenue miss revealed weaker underlying demand than the seasonal Personal Electronics lift can hide. The rebuttal is that inventory fell, CapEx was cut, free cash flow stayed positive, and Q4 gross margin is guided higher despite unused-capacity charges.

What to watch next is narrow and measurable. First, Q4 revenue needs to land around the $3.28 billion guide and not break the minus side of the plus or minus 350 basis points range; a miss there would say the Q3 Street revenue shortfall was demand, not timing. Second, gross margin needs to print near about 35% with unused-capacity charges still around the about 290 basis points level; if gross margin does not improve from 33.2%, the capacity thesis breaks. Third, full-year 2025 net CapEx should finish slightly below $2 billion, because a reversal toward the old $2 billion to $2.3 billion range would weaken the free-cash-flow and supplier-discipline argument. Finally, inventory should continue down from $3.17 billion after the about $100 million Q3 reduction; if inventory rises while revenue only grows 2.9% sequentially, the market will be right to treat the EPS beat as a low-quality trough bounce rather than the start of margin repair.