The demand side was not really in dispute. Management guided Q2 revenue to $5.0 billion to $5.4 billion with EPS of $1.77 to $2.05, and the end markets were cooperating. In Q1, industrial revenue rose more than 30% year over year and data center revenue jumped roughly 90%, the latter driven by power-management chips feeding AI infrastructure. Analog revenue, TI's largest business, grew 22% to $3.92 billion. Traders had been pricing a high probability that Q2 analog would clear $4 billion.
So if the top line was close to a sure thing, why had the stock fallen about 15% from its high? Two reasons, and they are more interesting than the revenue.
The factory bet that defines the story
For most of the last decade TI made a contrarian wager. While much of the industry moved toward outsourcing production to foundries, TI spent heavily to own its manufacturing, sizing the buildout at more than $20 billion over roughly six years on new 300mm wafer fabrication.
The logic was sound and long-term: owning your fabs gives you cost advantages, supply security, and control that foundry customers do not have, and 300mm wafers are structurally cheaper per chip than the older 200mm standard. In an industry that periodically gets supply-shocked, controlling your own capacity is a genuine moat.
But capacity is paid for twice. First in cash, when you build it. Then again on the income statement, as depreciation, spread over years whether or not the factories are running full. That second bill is the one arriving now.
Here is why the timing bites. A company completing a massive capacity buildout starts depreciating those assets regardless of how busy they are. If revenue is soft, you have a lot of expensive, underused factory absorbing cost, which is exactly what happened during the inventory correction of late 2025 that pressured volumes and factory absorption. Now revenue is recovering, but the depreciation from the finished buildout is hitting at the same time, so the question is whether rising sales outrun rising fixed costs fast enough to expand margins.
That is why the whole quarter, for investors, reduces to a margin question rather than a revenue one. One analyst framed the bar precisely: watch whether gross margin approaches 60%. Strong revenue with disappointing margins would confirm the depreciation is eating the recovery. Strong revenue with expanding margins would mean the bet is starting to pay off. Same sales, opposite conclusions.
The demand that may have been borrowed
The second concern is subtler and specific to the number everyone is celebrating.
That 90% data center growth is the most exciting figure in TI's recent results, and skeptics argue part of it was pull-forward, orders that landed at TI only because rival analog chipmakers hit foundry bottlenecks. In other words, some customers may have come to TI not because they preferred TI, but because competitors could not supply them, and TI, with its owned capacity, could.
If that read is right, it connects directly to the factory story in a way that cuts both directions. TI's manufacturing independence is precisely what let it capture that demand when competitors were constrained, which is the moat working as designed. But demand captured because rivals were temporarily bottlenecked can leave when the bottlenecks clear. If competitors add capacity in the back half of 2026, some of that 90% surge could reverse, and share gained on supply scarcity is not the same as share gained on preference.
This is the genuinely hard part of reading the quarter, because a big data center number looks bullish on the screen and its durability depends on information that is not in a single release: how much was structural demand versus temporary share, and whether it holds when supply normalizes. The forward commentary matters more than the printed figure.
What actually resolves it
Because revenue is largely known, the load-bearing disclosures are the ones around it.
Gross margin trajectory is the headline test, the direct measure of whether the recovery is outrunning the depreciation wave. Sixty percent is the informal bar.
Free cash flow per share is the number TI's own management has pointed to, with CEO Haviv Ilan projecting $8 for 2026 as trailing free cash flow recovered to $4.4 billion from $1.7 billion. This is the real point of the factory strategy: the heavy capex years are supposed to give way to a period where the fabs are built, spending falls, and cash flow inflects upward. Evidence of that inflection is the bull case in one metric, and it is why free cash flow per share, not EPS, is the number that best tests whether the strategy is working.
And the durability of data center and industrial demand, which is about management's commentary on channel inventory, order trends, and whether the data center strength reflects lasting design wins or temporary competitor weakness.
One data point worth holding for perspective: even after 30%-plus industrial growth, industrial revenue still sits roughly 15% below its 2022 peak. The recovery is real and also incomplete, which means there is plausibly more runway, and also that the prior peak has not yet been reclaimed. Both framings are true, and which one an investor emphasizes says more about their priors than about TI.
Texas Instruments is close to the purest expression of a specific industrial bet: own the means of production, accept years of heavy spending and the depreciation that follows, and come out the other side with a cost and supply advantage that compounds. The revenue will probably be strong, and it is the least informative part of the report. The margin line tells you whether the depreciation is winning or losing against the recovery. The free cash flow line tells you whether the capex era is turning into the harvest era. And the data center commentary tells you whether the best number in the quarter is a foundation or a loan.