Microchip’s beat is less about demand and more about a margin inflection the market may underprice

The print should change the debate from “when does Microchip’s top line recover?” to “how much margin recapture is already visible before the top line has recovered?” Street revenue expectations were not materially wrong: actual revenue of $1,140.4 million versus estimate $1,132.7 million produced only a +0.7% surprise. The real surprise was below the revenue line, with EPS of $0.35 versus estimate $0.33 and a +5.9% surprise. That separation matters because the stock’s controversy has been whether the company’s cycle trough in Q4 FY2025 revenue of $970.5 million and gross margin of 51.6% would require a clean demand rebound to repair profitability. This quarter says no: revenue is still below the Q2 FY2025 level of $1,163.8 million, with revenue YoY still -2.0%, yet gross margin has already moved to 55.9% and EPS has turned positive at $0.08 on the quarterly history basis. What was priced in was a bottoming revenue story after Q1 FY2026 revenue grew +10.8% QoQ; what surprised was that earnings could move faster than revenue even while the company still carried meaningful factory and inventory headwinds.

That distinction becomes clearer when the September quarter is placed against the last eight quarters rather than against one consensus line. Revenue peaked in the provided history at $2,288.6 million in Q1 FY2024, then fell through $1,765.7 million in Q3 FY2024, $1,325.8 million in Q4 FY2024, $1,026.0 million in Q3 FY2025, and $970.5 million in Q4 FY2025. The rebound to $1,075.5 million in Q1 FY2026 and $1,140.4 million in Q2 FY2026 is real, but still modest relative to the prior peak. The market’s mistake would be to anchor on the still-low sales base and miss the operating leverage already reappearing in the model. Gross margin moved from 51.6% in Q4 FY2025 to 53.6% in Q1 FY2026 and 55.9% in Q2 FY2026, while the December-quarter history line shows $1,186.0 million revenue and 59.6% gross margin. The company’s own call basis also framed the September quarter as $1.14 billion of net sales, with J. Bjornholt saying it was “up 6% sequentially and $10.4 million above the midpoint of our September quarter guidance provided on August 7.” That wording matters because it separates a narrow top-line beat from the more investable setup: management did not need a large demand surprise to begin restoring margins.

The capacity story explains the margin guide, because the biggest visible drags are quantified and therefore can shrink without requiring heroic end-market assumptions. On the call, J. Bjornholt put non-GAAP gross margin at 56.7%, including capacity underutilization charges of $51 million and new inventory reserve charges of $71.8 million. Steve Sanghi then converted the combined drag into the P&L language investors care about: “Divide that by the net sales of $1.1404 billion and you get a non-GAAP gross margin impact of 10.8 percentage points.” That quote earns its place because it is not generic color; it is management explicitly identifying a recoverable margin pool. The December guide makes the same point from another angle: management expects net sales of $1.129 billion, plus or minus $20 million, down 1% sequentially at the midpoint, but expects non-GAAP gross margin between 57.2% and 59.2% of sales. If revenue is guided lower and margin is guided higher, the improvement is not just mix or volume. It is the reversal of prior self-inflicted or cycle-driven costs, which is why the EPS guide of $0.34 to $0.40 matters more than the revenue guide.

That earnings-first recovery is also visible in the cost structure, but the company has not yet earned full credit because leverage remains constrained by opex and leverage on the balance sheet. J. Bjornholt disclosed total operating expenses of $549 million, including acquisition intangible amortization of $108.1 million, special charges of $6.3 million primarily tied to the closure of Fab 2, share-based compensation of $53.3 million, and $12.3 million of other expenses. Operating expense guidance for December is 32.3% to 32.7% of sales, with non-GAAP operating profit between 24.5% and 26.9% of sales. That is not a clean operating leverage story yet, because the guided opex ratio remains high relative to the revenue base, but it is enough to move the profit dollars: Steve Sanghi said non-GAAP operating profit is projected to increase by over $13 million sequentially at the midpoint of guidance. The balance sheet keeps the equity story from becoming a simple cyclical snapback: cash flow from operating activities was $88.1 million, adjusted free cash flow was $38.3 million, total debt decreased by $82 million, net debt increased by $247.7 million, trailing 12-month adjusted EBITDA was $1.103 billion, and net debt to adjusted EBITDA was 4.69 at quarter-end. That mix argues for multiple expansion only if the December margin guide is delivered, because debt reduction is not yet carrying the thesis.

The revenue composition supports the margin thesis more than the headline sales beat does, because Microchip’s MCU line showed the sharper sequential recovery while analog remained slower. Steve Sanghi said, “Specifically, our MCU business grew 9.7% sequentially with strong contribution from 32-bit MCU, while our analog business increased 1.7% sequentially.” The numbers matter for second-order implications inside the semiconductor stack: a 9.7% sequential MCU recovery points to a more meaningful inventory normalization in embedded control than in analog, while the 1.7% analog increase says broad-based analog demand is not yet accelerating at the same pace. That is relevant for TXN and NXPI comparisons. TXN’s latest reported revenue YoY was +18.6% with 58.0% gross margin, while NXPI reported revenue YoY of +12.2% with 56.2% gross margin. Microchip’s Q2 FY2026 revenue YoY was still -2.0% and gross margin was 55.9%, so it is not yet matching the peer demand recovery, but its December-quarter gross margin guide of 57.2% to 59.2% puts it in the same gross-margin neighborhood as TXN’s 58.0% and above NXPI’s 56.2% if delivered. That is the comparative point: Microchip is not winning the revenue-growth screen yet, but it is regaining the margin screen earlier than the YoY revenue line suggests.

The supplier read-through is narrow but concrete, because the data pack names only one supplier relationship: TSMC as wafer fabrication supplier for 40nm at JASM Kumamoto. Microchip’s September inventory went down by $73.8 million sequentially, per Steve Sanghi, and management guided December sales to $1.129 billion, plus or minus $20 million, down 1% sequentially at the midpoint. That combination is not a signal of an immediate wafer-order surge for TSMC tied to Microchip; it is a signal that Microchip is still consuming or rationalizing inventory while keeping revenue roughly stable around the $1.1 billion to $1.2 billion range. For TSMC, the magnitude available from this print is therefore not a demand uplift but a normalization marker: a customer using 40nm at JASM Kumamoto is reducing inventory by $73.8 million sequentially while guiding only a 1% sequential revenue decline. There are no named Microchip customers in the data pack, so any customer-specific read-through would be false precision; the defensible second-order point is that embedded MCU customers appear to be normalizing faster than analog customers, as shown by 9.7% sequential MCU growth versus 1.7% sequential analog growth.

That same discipline is needed when interpreting the call tone, because management sounded more constructive on guidance even as model-derived optimism declined. The tone history shows Q2 FY2026 sentiment at -0.15, guidance_tone at 0.02, prepared_sentiment at -0.09, qa_sentiment at -0.14, ai_optimism at 0.40, uncertainty at 37.3, and qa_evasiveness at 10.9. The next call in the series, Q3 FY2026, moved sentiment to 0.08, guidance_tone to 0.15, prepared_sentiment to 0.01, and qa_sentiment to 0.09, while ai_optimism fell to 0.23. The call-over-call delta captures the contradiction: sentiment +0.23, guidance_tone +0.13, prepared_sentiment +0.10, and qa_sentiment +0.23 improved, but tone_confidence fell -0.05, ai_optimism fell -0.17, and qa_evasiveness rose +24.6. The right interpretation is not that management became uniformly bullish. It is that management became more willing to guide margin improvement and profit improvement, while the Q&A still carried enough evasiveness to keep investors focused on proof points rather than narrative.

The conflicting tone metrics are not a reason to dismiss the print, because the hard guide is more specific than the sentiment. Management’s December-quarter revenue guide is $1.129 billion, plus or minus $20 million, and its non-GAAP gross margin guide is 57.2% to 59.2%. J. Bjornholt added, “But we are guiding a pretty nice sequential increase in the non-GAAP gross margins to like 58.2% at the midpoint of guidance.” The wording matters because the commitment is not to revenue acceleration; the commitment is to margin expansion at the midpoint even with a down 1% sequential sales guide. That is the variant perception in one sentence. Investors waiting for a clean revenue breakout may miss the first leg of earnings recovery, while investors buying only the revenue beat are overpaying for the wrong surprise. The EPS beat of +5.9%, the September non-GAAP gross margin of 56.7% despite $51 million of capacity underutilization charges and $71.8 million of new inventory reserve charges, and the December gross margin guide of 57.2% to 59.2% are the actionable data.

The bear case is still anchored in the same numbers, so it should be stated plainly. Revenue actuals beat by only +0.7%, December revenue is guided down 1% sequentially at the midpoint, and Q2 FY2026 revenue YoY remained -2.0%. GAAP earnings were also thin on the call basis, with J. Bjornholt reporting GAAP net income attributable to common shareholders of $13.9 million or $0.03 per share, positively impacted by an IRS audit settlement dating back to fiscal year 2007. Free cash flow is not yet enough to make leverage disappear quickly, with adjusted free cash flow of $38.3 million, net debt increasing by $247.7 million, and net debt to adjusted EBITDA at 4.69. Those numbers prevent a full-cycle re-rating unless the margin path is sustained. But they do not break the thesis, because the thesis is not that demand has already recovered to prior-cycle levels; revenue is still far below $2,254.3 million from Q2 FY2024 and below $1,163.8 million from Q2 FY2025. The thesis is that margin improvement can precede revenue normalization, and this print provides more support for that view than the headline revenue beat suggests.

The next confirmation point is therefore unusually clean. For the December quarter, the thesis requires revenue to land inside or above the $1.129 billion, plus or minus $20 million, guide and, more importantly, non-GAAP gross margin to land within the 57.2% to 59.2% range, with the midpoint marker at 58.2%. EPS should stay within the $0.34 to $0.40 guide, and non-GAAP operating profit should increase by over $13 million sequentially at the midpoint. The thesis strengthens if Microchip shows another inventory reduction after the $73.8 million sequential decline while keeping MCU growth above analog growth, using the September comparison of 9.7% sequential MCU growth and 1.7% sequential analog growth as the baseline. It breaks if gross margin fails to move toward 57.2% to 59.2% despite revenue near $1.129 billion, because that would mean the $51 million underutilization charge and $71.8 million inventory reserve drag are less recoverable than management implied. The date to underwrite is the next quarterly report after the 2025-11-06 call, when December-quarter revenue, gross margin, EPS, inventory, and operating profit will show whether Microchip is truly entering an earnings-led recovery or merely pausing after a shallow top-line bounce.