Texas Instruments trades near 49x as a 90%-growth data-center sliver reprices the whole cycle

---

There is a particular kind of stock that becomes dangerous not because management is lying but because management is telling the truth and the market is doing the extrapolating. Texas Instruments is one of them. The first-quarter 2026 numbers it reported on April 22 are, as far as any outside reader can verify, exactly what the company says they are. Revenue of $4.83 billion was up 19% year over year, 9% sequentially, and above the high end of the company's own guidance. Earnings of $1.68 a share beat the $1.36 the Street expected by a wide margin. Net income rose 31% to $1.55 billion. Gross margin recovered to 58%, up 210 basis points sequentially. Analog revenue, the company's core, rose 22% to $3.92 billion at a 41.7% operating margin. The data-center line grew roughly 90% year over year. And the company guided the second quarter to $5.0 to $5.4 billion in revenue with EPS of $1.77 to $2.05 — another acceleration. None of that is in dispute.

The forensic case against Texas Instruments is therefore not a case against its honesty. It is a case against the arithmetic the market has stapled onto that honesty. Trading near a trailing price-to-earnings ratio of roughly 49 against current earnings of about $5.87 a share, with the stock up something like 58% in 2026, Texas Instruments is priced 99% above its own ten-year average multiple of about 24.65. The thesis here is narrow and falsifiable: a company whose largest end-markets are industrial and automotive — two of the most cyclical demand pools in the economy — is being capitalized as if a just-arrived cyclical recovery, supercharged by a small but fast-growing AI-data-center line, is a permanent secular ascent. That is the cyclical-priced-as-secular problem, dressed in 300-millimeter silicon.

The multiple is the whole story

Start where the risk actually lives, which is not in the income statement but in the price. A trailing P/E near 49 — with the forward figure still around 35 even on optimistic 2026 estimates — is not a number you hand to an analog chipmaker in the ordinary course of business. Analog and embedded semiconductors are good businesses: high-margin, sticky, long-lived, diversified across tens of thousands of parts and customers. But they are not hyperscale-software businesses, and the market has historically known the difference, paying Texas Instruments somewhere in the low-to-mid 20s on earnings across most of the past decade. The re-rating to the high 40s is recent, and it is overwhelmingly the work of one narrative shift: that the company has stopped being a proxy for the global industrial cycle and become a play on artificial-intelligence power and connectivity.

Multiples of this magnitude do something specific to risk: they invert it. At 24 times, a quarter of decelerating orders is a disappointment to be absorbed. At 49 times, the same quarter is a thesis break, because the multiple is not paying for the cash flows in hand — it is paying for the extrapolation of the cash flows to come. The market has, in effect, pre-booked the recovery, pre-booked the data-center ramp, pre-booked the 2026 free-cash-flow rebound, and pre-booked the idea that all three compound together for years. That is the asymmetry a forensic reader should fixate on. The bull must be right about the next several years of synchronized acceleration. The bear needs only one leg — automotive, China, the data-center growth rate, or the multiple itself — to wobble. When a stock has risen 58% in five and a half months, the easy money has already been made by someone, and the question is whether the next buyer is paying for a business or for a momentum.

A 90% number doing 11% of the work

The single most important sentence in the Texas Instruments bull case is "data center grew roughly 90% year over year." It is true, and it is the spark that lit the re-rating. It is also, on close inspection, a number carrying far more narrative weight than revenue weight. The data-center segment is, by the company's own framing, somewhere around eleven to twelve percent of total sales. Ninety-percent growth on eleven percent of the business adds roughly ten points of mix-shifted revenue at the very most — meaningful, accretive, real, but not transformational to a $19-billion-a-year revenue base in the way the share price reaction implies.

This is the denominator illusion in its purest form. A spectacular growth rate on a small base is being mentally annualized across the whole company, as if Texas Instruments were becoming a data-center company rather than an industrial-and-automotive company with a fast-growing data-center sliver. The other roughly 88% of the business — the analog and embedded content inside factory automation, motor drives, appliances, medical devices, automobiles, and personal electronics — does not grow 90%. It grows with the global capital-expenditure and auto-production cycle, which is to say it grows in the low-to-mid teens in a recovery and shrinks in a downturn. The forensic question is not whether the 90% is real. It is whether the market has quietly multiplied an eleven-percent segment's growth rate across an eighty-nine-percent base it does not describe.

Automotive is the dog that has not barked

If the data-center line is the part of the story the bulls repeat, automotive is the part they hurry past. And it is not a small part. Automotive has historically been one of Texas Instruments' two largest end-markets, alongside industrial, together accounting for the majority of revenue. On the first-quarter call, management characterized automotive as flat to slightly down sequentially — the one major segment without a clear re-acceleration signal. More pointedly, the company flagged that a China-driven softening could drag the overall mix lower into the third and fourth quarters.

This matters because of how cyclical recoveries actually unfold. Healthy, durable recoveries broaden: every end-market turns up together, and the rising tide is genuine. Narrow recoveries — where one or two end-markets carry the print while the largest one stalls — are exactly the kind that disappoint two or three quarters later, when the carrying segment decelerates off its easy comparisons and the lagging segment has not yet inflected. Texas Instruments is, on its own description, running a narrow recovery: industrial up 30%, data center up 90%, and automotive flat-to-down with China as a downside risk. The bull narrative treats automotive as a coiled spring that will inevitably snap back. The forensic reading treats a stalled largest-segment, against the backdrop of Chinese demand softness and a still-uncertain global auto-production outlook, as an unconfirmed assumption the multiple has already paid for.

The free-cash-flow rebound is a forecast, not a fact

The defense of the multiple leans heavily on free cash flow. The company has guided investors toward more than $8 per share of free cash flow in 2026 as the six-year elevated 300-millimeter capital-expenditure cycle winds down and capital spending moderates from roughly $5 billion in 2025 toward a $2-to-$3-billion range. On a trailing-twelve-month basis through March 31, 2026, free cash flow was $4.35 billion, or 23.6% of revenue — but that figure includes $630 million of CHIPS Act incentive cash. Strip the government money out and the underlying number is closer to $3.7 billion, with operating cash flow of $7.82 billion against $4.10 billion of capital expenditure.

Two forensic observations follow. First, the headline free-cash-flow figure is partly a policy subsidy. The CHIPS Act incentives flattering the 2026 cash number are a function of Washington appropriations and the political durability of domestic-fab support — a dependency, not an operating achievement, and one that does not recur on the same schedule forever. Second, and more important, the $8-plus-per-share figure is a projection, not a print. It requires three things to all go right at once: capital spending must actually fall to the bottom of the guided range, revenue must keep growing into the back half despite the automotive stall, and gross margin must keep climbing as 300-millimeter utilization improves. Even if the company hits $8 of free cash flow, the stock near $288 is being paid at roughly 36 times that figure — a rich price for a forecast that has not yet appeared in a single actual quarter. The market is not valuing the cash flow Texas Instruments has produced. It is valuing the cash flow it has promised to produce in a year that is barely a quarter old.

Inventory tells a quieter story than the headline

Hidden beneath the 58% gross margin is a balance-sheet line worth pausing on. Days of inventory stood at 209 at the end of the first quarter — down 13 days sequentially, which the company correctly framed as progress, but still an elevated absolute level for a chipmaker. Texas Instruments runs inventory deliberately high as a strategic choice: it wants stock on the shelf to capture demand spikes and to serve customers others cannot, a defensible competitive posture. But 209 days is also a reminder that the company built capacity and inventory through the downturn ahead of a recovery whose breadth is not yet confirmed.

There are two readings of high inventory into a recovery, and the gap between them is the whole investment debate. The bull reading is that the inventory is dry powder — finished goods ready to convert into revenue the instant demand broadens, which juices reported margins as those pre-built units sell without fresh production cost. The bear reading is that elevated inventory into a narrow recovery is the raw material of a future air pocket: if industrial and data-center demand cools before automotive inflects, that 209 days does not convert smoothly to sales — it sits, and it pressures pricing and utilization on the way back down. The same number supports both stories. What it cannot support is the assumption, embedded in a 49 multiple, that only the bull reading is possible.

The 300mm advantage is real, and already in the price

Texas Instruments' strategic bet — a multi-year build-out of 300-millimeter analog wafer fabs in Texas and Utah — is genuinely smart, and the bears should say so plainly. Manufacturing analog chips on 300-millimeter wafers rather than the industry-standard 200-millimeter lowers the unit cost of each chip by an estimated 30-to-40%, a structural cost advantage that competitors relying on outsourced foundries or older fabs cannot easily match. Owning the fabs also gives Texas Instruments control over supply, geopolitical resilience, and the ability to serve customers through shortages — a moat built of concrete and capital rather than patents.

But a moat that is real is not the same as a moat that is cheap. The 300-millimeter advantage has been the centerpiece of the company's capital-allocation story for years; it is not news, and it is not a secret. By the time a strategic advantage has been disclosed, modeled, and celebrated through six years of elevated capital spending, it is priced in. The forensic point is one of timing, not merit: the cost advantage is arriving on the income statement now, as utilization climbs and the capital-spending burden falls, which is precisely why margins and free cash flow are inflecting. The market is paying a 49 multiple for an inflection that is, in large part, the mechanical and well-telegraphed payoff of a capex cycle everyone has watched for half a decade. Buying the payoff at the moment of maximum visibility is the opposite of buying it cheap.

The peer comparison the multiple ignores

Place Texas Instruments next to its closest analog and embedded peers and the valuation looks even more stretched. Analog Devices, NXP Semiconductors, and Microchip Technology are all exposed to the same industrial-and-automotive demand pools, the same cyclical recovery, and in several cases the same data-center tailwinds. None of them carries a trailing multiple anywhere near 49 times. Texas Instruments has long commanded a premium to the group — its margins are higher, its capital returns more disciplined, its manufacturing more vertically integrated, and that premium is deserved. But "deserves a premium" and "deserves a doubling of its own historical premium" are different claims.

The re-rating has stretched the gap between Texas Instruments and its peers to a width that implies the company is not merely the best operator in analog but a fundamentally different kind of business — a secular AI compounder rather than a cyclical chipmaker. If that re-classification is wrong, the convergence trade is brutal: a reversion from 49 times toward even the high-20s or low-30s, with no decline in earnings at all, is a 30-to-40% drawdown delivered entirely by multiple compression. That is the mechanism by which expensive stocks fall even when the business does fine. The earnings hold; the willingness to pay for them does not.

The insiders are not buying the narrative they are selling

One quiet tell sits underneath any richly re-rated stock: what the people closest to the numbers do with their own shares. A 49-times multiple is, definitionally, a price at which management's equity is worth the most it has ever been, and at which the rational disposition for any long-tenured executive sitting on appreciated stock is to trim into strength rather than add. That is not an allegation of wrongdoing — selling appreciated stock at a multiple twice your historical norm is simply prudent personal finance, and it is what disciplined operators do. But it is a reminder that the people who know the cyclicality of this business in their bones are not the marginal buyers setting the price. The marginal buyer near the highs is, by construction, the investor with the least operating memory of how quickly an analog cycle can roll — the one extrapolating a 90% data-center number and a $8 free-cash-flow promise into a permanence the company itself has never claimed. When the most-informed holders are natural sellers into a re-rating and the least-informed are natural buyers, the burden of proof on the new money is heavy, and the price leaves no room to be wrong about who is right.

What the bulls genuinely get right

The bull case deserves a fair and specific hearing, because in important ways it is strong. Texas Instruments did not engineer this quarter — it earned it. Nineteen-percent revenue growth, a clean beat above guidance, a 31% jump in net income, and a 210-basis-point sequential gross-margin recovery to 58% are the marks of a real cyclical upturn, not a cosmetic one. Management was explicit that the strength came from the genuine return of industrial demand rather than tariff-driven pull-forwards — a crucial distinction, because pull-ins borrow from the future while a true demand return compounds into it. The data-center growth, even at eleven or twelve percent of revenue, is the right kind of exposure to have right now, and it gives Texas Instruments a credible foothold in the single most powerful capital-spending wave of the decade.

The strategic foundation is equally real. The 300-millimeter cost advantage is a durable, hard-to-replicate moat, and the company is exiting a punishing capital-expenditure cycle exactly as utilization and margins inflect upward — a sequencing that should drive genuine free-cash-flow-per-share growth from a 2025 base of $3.23 toward management's $8-plus target. Texas Instruments has one of the most shareholder-friendly capital-return records in the industry, decades of dividend growth, and a balance sheet that can fund both the buyback and the fabs. The Q2 guide — $5.0 to $5.4 billion in revenue with EPS up to $2.05 — points to continued acceleration, not a one-quarter spike. If automotive inflects on schedule, if the data-center line keeps its torrid pace, and if the $8 free-cash-flow target lands, the forward multiple compresses on its own as earnings grow into it, and today's price looks merely full rather than absurd. The bull is not betting on a fantasy. The bull is betting that a real recovery stays broad and that a real moat keeps widening. The disagreement is entirely about the price being paid for that bet, and about how much of it has already been assumed true.

The kicker

Strip away the narrative and the position reduces to a single, uncomfortable trade structure. An investor buying Texas Instruments near $288 is paying roughly 49 times trailing earnings and about 36 times a free-cash-flow figure the company has promised but not yet produced, for a business whose largest end-market is flat-to-down, whose re-rating is powered by a segment that is one-eighth of revenue, and whose cash flow is partly underwritten by federal subsidy. Every one of those facts is individually defensible and collectively combustible. The recovery is real; the moat is real; the quarter was real. What is not yet real is the future the multiple has already paid for in full — and in a cyclical business, the gap between the recovery you can see and the permanence the price assumes is exactly where the losses live.

The bet is no longer whether Texas Instruments can recover — it plainly has — but whether a market paying 49 times for a cyclical chipmaker has confused the first quarter of an upturn for the steady state of a secular machine, and pre-booked years of acceleration that automotive, China, and the back half of 2026 have not yet agreed to deliver.