Texas Instruments gave investors the kind of quarter companies usually want to frame, then watched the stock fall after hours. That tells you the semiconductor market isn't rewarding every AI-adjacent dollar the same way.
The numbers were clean. Texas Instruments reported second-quarter revenue of $5.46 billion on July 22, up 23% from a year earlier, with earnings of $2.14 a share and operating profit up 48% to $2.31 billion. MarketBeat put Wall Street's consensus at $5.24 billion in revenue and $1.91 in EPS, so this wasn't a miss dressed up as a beat. It was a beat. The stock still fell roughly 3% in extended trading.
That is the market's answer.
TI's own release said growth was broad, led by industrial, data center and automotive. It also guided third-quarter revenue to $5.65 billion to $6.15 billion, with EPS of $2.23 to $2.57. The company generated $8.7 billion in trailing 12-month operating cash flow and $6.5 billion in free cash flow, helped by CHIPS Act incentives. If you're looking only at the income statement, you struggle to find the problem.
The problem is where the money wants to go now. Texas Instruments makes analog and embedded processing chips, the parts that convert signals, manage power and sit inside industrial equipment, vehicles and data centers. They matter. You can't run the physical world without them. But they don't sit in the center of the AI trade the way Nvidia's GPUs do, or the way high-bandwidth memory sits beside them. Data center revenue doubling, as Barron's noted, sounds like an AI headline until you remember TI is selling power-management parts into that buildout, not the chip everyone is waiting in line to buy.
The market knows the difference.
AI adjacency is not AI control
Nvidia's stock moves on every hyperscaler spending plan because its chips are the scarce input in training and inference. Broadcom gets some of that treatment because its custom silicon and networking revenue are tied more directly to the AI buildout. TI benefits from the same construction cycle, but one step removed. If Amazon, Microsoft or Google builds another AI data center, it needs power chips and signal chains. It also needs concrete, cooling systems and electricians. Not every supplier gets the same multiple.
That isn't unfair. It's just how investors price control. If you own the bottleneck, you get paid for the bottleneck. If you own useful parts around it, you get paid like a strong cyclical supplier that happens to have a data center tailwind. TI can be excellent and still not be Nvidia. Investors just made that distinction in public.
There's also a valuation issue sitting under the reaction. GuruFocus showed TXN trading at a trailing P/E of about 50.5 on July 21, and its recent stock alerts flagged $88.9 million of insider selling over the prior three months with no insider buying reported. You don't need to overread insider sales, executives sell for plenty of reasons, but at that multiple the market gives you less room to be merely good. It wants acceleration. It wants direct AI upside. It wants the kind of story that makes next year's estimates look too low before analysts have finished updating this year's.
TI gave investors strength. It didn't give them that.
The factory bill is still coming due
The other issue is manufacturing. TI has spent heavily to expand its domestic capacity, including Sherman, Texas, and Lehi, Utah. That strategy gives the company more control over supply and more 300mm production, which management keeps pointing to as a long-term advantage. It also pushes costs through the model today. In the Q2 call transcript published by Seeking Alpha, CFO Rafael Lizardi said TI now expects depreciation of $2.2 billion to $2.4 billion in 2026, with further upward pressure in 2027 at a slower rate.
That bill lands now.
Capital spending has already been large. TI said it invested $3.3 billion in capital expenditures over the past 12 months, while returning $5.8 billion to shareholders. The company also kept 2026 capex guidance at $2 billion to $3 billion, according to the call transcript. That is serious discipline compared with the pandemic-era buildout, but it still means investors have to underwrite a manufacturing plan before they see the full demand payoff.
Then there is pricing. The MarketBeat transcript records management saying first-half pricing was flat, better than the normal low-single-digit annual decline TI often sees, and that customer-by-customer price increases have started. Some effect is expected in Q3, with more in Q4. Here's the thing: price increases can support margins, but they also ask customers to accept higher costs in markets that are not all moving at Nvidia speed. Industrial and automotive customers qualify suppliers slowly, but they also remember who raised prices when supply tightens.
This is still a good business. Analog chips have long product lives, broad customer bases and practical switching friction. TI's second-quarter analog revenue rose 26% year over year, while embedded processing revenue rose 16%, according to the company's release. Those are not weak numbers. They are the kind of numbers most semiconductor companies would gladly take.
The selloff says something narrower and more useful for you as an investor. In this chip market, a strong quarter is no longer enough if the stock has already priced in a lot of recovery and the AI link is indirect. TI proved demand is improving across its core markets. The market asked whether that improvement deserves a GPU-era valuation.
For now, the answer was no.
Also read: OpenAI's own AI models broke out of a test sandbox and autonomously hacked Hugging Face • Hyundai's 35,000 striking workers just forced the first real test of who controls humanoid robots on the factory floor • Substack's new AI scanner tells paying readers exactly how much of their newsletter a human actually wrote
