TXN’s beat is not the story; the market is underpricing the cost of owning the next analog upcycle

The actionable read on this print is that Texas Instruments is not giving investors a clean analog recovery trade yet. What was priced in was a modest top-line recovery, with Street revenue at $4,645.4 million and EPS at $1.49. What actually surprised was the mix of outcomes: revenue beat by +2.1% at $4,742.0 million, while EPS missed by -0.7% at $1.48. That combination matters because the stock’s debate has been shifting from trough demand to earnings power, and this quarter says the top line is healing faster than the income statement. The market may be treating the revenue beat as confirmation that Texas Instruments is exiting the industrial and auto correction; the variant perception is that the company is exiting it with a heavier fixed-cost base, enough restructuring noise to reduce EPS by $0.10, and a gross margin still stuck below the pre-downturn regime.

The first-order surprise was not demand direction, because management did not frame the quarter as an upside breakout. Haviv Ilan’s wording was deliberately restrained: “Revenue came in about as expected at $4.7 billion, an increase of 7% sequentially and an increase of 14% year over year.” That sentence earns attention because “about as expected” reduces the informational content of the revenue beat, even though the reported basis beat the Street by +2.1%. The market wanted evidence that analog units and channel inventory were normalizing; it got enough evidence to keep the recovery thesis alive, but not enough earnings conversion to raise confidence in near-term EPS. The EPS miss at $1.48 versus $1.49 is numerically small, but it is qualitatively important because Rafael R. Lizardi said, “Earnings per share included a $0.10 reduction not in our original guidance.” A business with a revenue beat and an EPS miss is telling investors that operating leverage is still being intercepted by costs that were not in the original guide.

That earnings conversion issue shows up cleanly in the financial trajectory: revenue has recovered from the trough and is now back above the 2023 run-rate, but gross margin has not recovered with the same slope. Q3 revenue of $4,742.0 million is above the prior-year $4,151.0 million, yet gross margin at 57.4% remains below the 62.1% level from Q3 FY2023. This is the crux of the print: demand has improved enough to beat the Street, but the manufacturing model has not yet restored the margin structure investors associate with Texas Instruments. The revenue line says the analog cycle is no longer contracting; the gross margin line says the company is still carrying the cost of capacity decisions made for a larger future business.

The capacity story explains why the revenue beat did not translate into an EPS beat, because the cost base is now the swing factor in the model. Lizardi’s answer to a gross margin question was notable for refusing to soften the depreciation headwind: “You know, we let the EPS guide speak for itself, but you have lower revenue, you fall that through, you have increases in depreciation, for the year is $1.8 billion to $2 billion.” The phrase “let the EPS guide speak for itself” matters because it declines to isolate a reassuring bridge; it tells PMs to take the EPS range at face value. The follow-on comment that depreciation for next year has been framed at $2.3 billion to $2.7 billion, “but to be on the lower end of that range,” is still not a margin reset, because the lower end is still higher than the current-year range. That is the mispricing risk: investors may be extrapolating revenue acceleration while underweighting the lag before fixed-cost absorption repairs EPS.

The Q4 guide reinforces that this is not a straight-line recovery, even though the longer data series has begun to look better. The company guided Q4 revenue to $4.22 billion to $4.58 billion and EPS to $1.13 to $1.39, which sits below Q3’s $4.7 billion company-reported revenue and $1.48 company-reported EPS. This is not a Street comparison sentence, and it should not be read against the beat metrics; it is management’s own basis for near-term planning. The important point is that the guide embeds a seasonal or demand pause while the depreciation burden continues, so the next quarter is more likely to test investor patience than confirm a clean inflection. The Q4 history line later shows revenue at $4,423.0 million and EPS at $1.27, which is consistent with the guide midpoint debate rather than a post-Q3 acceleration narrative. If investors were paying for an analog restocking cycle to overwhelm capacity cost, the next quarter’s numbers argue that timing remains contested.

The product and end-market color gives the bulls something real, but it is not broad enough to dismiss the margin problem. Ilan said analog revenue grew 16% year over year, while embedded processing grew 9%, which means the core analog franchise is recovering faster than the embedded business. Communications equipment was the eye-catching vertical, growing about 45% year on year and about 10% sequentially, but the size and sustainability of that contribution were not quantified in the data pack. The right interpretation is not that Texas Instruments has become a communications equipment story; it is that a sharp communications rebound helped the quarter while analog breadth improved enough to support the revenue beat. Because the company’s gross margin stayed at 57.4%, the communications lift did not change the margin debate. That keeps the burden of proof on whether industrial and auto can add volume without incremental price or mix pressure.

The cash-flow picture complicates the bearish margin read, because the company is still funding the investment cycle from a large operating cash base. Cash flow from operations was $2.2 billion in the quarter and $6.9 billion on a trailing twelve-month basis, while capital expenditures were $1.2 billion in the quarter and $4.8 billion over the last twelve months. That leaves free cash flow of $2.4 billion on a trailing twelve-month basis, but it also shows how much of the analog recovery is being reinvested before shareholders see the full benefit. CHIPS Act incentives helped, with $637 million included and a $75 million payment received in the third quarter, so the free-cash-flow optics are not purely organic. None of this means the capex strategy is wrong; it means the stock is being asked to capitalize future 300-millimeter advantage while current gross margin is still 57% of revenue on the company’s own call basis.

The balance sheet and capital return reduce financial risk but do not solve the equity-duration problem. Texas Instruments ended the quarter with $5.2 billion of cash and short-term investments, against total debt of $14 billion with a weighted average coupon of 4%. The company returned $6.6 billion to owners in the past twelve months, including $1.2 billion in dividends and $190 million of repurchases in the quarter. That return profile will keep long-only holders engaged, but it does not answer the central earnings-power question. If free cash flow is being suppressed by capex and gross margin is still below the old regime, buybacks are not the main lever for EPS acceleration. The more relevant lever is whether utilization and depreciation absorption can move gross margin out of the high-50s while revenue remains above the $4.7 billion area.

Inventory is the one operating datapoint that argues against a demand cliff, but it also argues against a snapback. Inventory ended the quarter at $4.8 billion, up only $17 million from the prior quarter, and days fell to 215, down sixteen days sequentially. That is the kind of datapoint that supports supply discipline: inventory dollars barely moved while days improved. But 215 days is still a large buffer, and it makes the recovery more about shipment normalization than scarcity pricing. For Arrow Electronics, the linked customer in the data pack, this print reads as a healthier flow-through environment rather than a shortage-driven distribution surge: revenue at Texas Instruments beat by +2.1%, but inventories remained at $4.8 billion. For National Instruments, UMC, and JCET, the supplier read-through is more capital-intensity than volume exuberance: Texas Instruments spent $1.2 billion of capex in the quarter, while plant closures tied to the last 250-millimeter fabs drove $0.08 of restructuring charges.

The peer context is where the bull case has its cleanest quantitative defense, even though it does not erase the cost issue. In the IDM peer set, Texas Instruments’ latest reported gross margin of 58.0% sits above Intel at 39.4% and above NXPI at 56.2%, while its revenue YoY of +18.6% is ahead of NXPI’s +12.2%. That matters because Texas Instruments is not losing relative profitability in analog-heavy semis; it is still near the top of the group on margin. The variant view is therefore not that the franchise is impaired. It is that the market may be assigning too much value to the eventual margin destination before there is evidence that the new depreciation base can support it. A high relative gross margin can coexist with an earnings estimate problem if incremental revenue is being absorbed by depreciation and restructuring rather than dropping through.

The call delivery supports that cautious interpretation, because management’s prepared script sounded better than the guidance tone. In the tone history, Q3 FY2025 prepared_sentiment was 0.46, the highest value in the shown series, while guidance_tone was -0.21. That split is exactly what the financials imply: management had plenty to say about recovery and cash generation, but the forward guide carried the burden of lower revenue and higher depreciation. The Q&A was not promotional either, with qa_sentiment at 0.02 and qa_evasiveness at -28.1. The negative evasiveness score suggests answers were relatively direct, which makes the depreciation language harder to wave away. When a direct Q&A tone accompanies a negative guidance tone, the market should not assume management is sandbagging a margin rebound.

That tone pattern also helps separate what investors wanted from what the company actually committed to. The market likely came in prepared to reward evidence that the analog cycle had bottomed, and the revenue beat plus 16% analog growth met that bar. The company did not commit to an immediate earnings inflection, and the EPS guide of $1.13 to $1.39 made that explicit. The reason this distinction matters is that Texas Instruments has historically trained investors to look through cycles because of manufacturing control, long product lives, and capital return. This print says those advantages are intact, but it also says the timing of the payoff has lengthened. The last 250-millimeter fab closures may support long-term efficiency, but they cost $0.08 in the quarter, and the depreciation path remains the nearer-term headwind.

The portfolio conclusion is to own the recovery only if the mandate can tolerate another quarter where revenue is the cleaner datapoint than EPS. For relative semis, the stock offers a higher-quality analog recovery than most IDMs, supported by a gross margin profile that still screens above major peers. For earnings-momentum capital, the print is not yet sufficient, because a +2.1% revenue surprise produced a -0.7% EPS surprise and management guided the next quarter below Q3 on its own reported basis. The more nuanced trade is that Texas Instruments may deserve a better multiple than cyclicals with weaker gross margin, but not a full normalization multiple until the company proves that revenue above $4.7 billion can coexist with gross margin meaningfully above the 57.4% area. That is the debate the print sharpened, not resolved.

What to watch next is concrete. The thesis is confirmed if Q4 revenue lands near the upper end of the $4.22 billion to $4.58 billion range while EPS lands near the upper end of $1.13 to $1.39, because that would show the depreciation burden is not overwhelming seasonal demand. It is broken if gross margin remains near 55.9% after the Q4 downtick or if inventory days stop improving from 215, because then the recovery is being bought with capacity cost and working capital rather than demand pull. On the next call, the key date is the Q4 FY2025 report following this 2025-10-21 event, and the key numbers are the depreciation framing of $2.3 billion to $2.7 billion for next year, the $4.8 billion inventory balance, and whether analog growth can stay near 16% year over year without communications equipment again doing the visible work.