SANTA CLARA — Every stock in the semiconductor complex fell Thursday, but Intel fell for two reasons where its competitors fell for one.
The first was Kevin Warsh. The Federal Reserve governor’s hawkish speech before the Chicago Council on Global Affairs pushed interest rates higher across the yield curve, compressing multiples throughout the technology sector and sending the Philadelphia Semiconductor Index down 2.7% on the session. Intel closed at $24.18, off 2.2%.
The second reason is structural and has nothing to do with the Federal Reserve. Intel is the only major chip company in the index that is simultaneously navigating a foundry business turnaround, integrating a new AI accelerator into a market already split between Nvidia and AMD, and managing a balance sheet that has accumulated debt faster than its free cash flow can retire it. Warsh made the rate environment less forgiving. Intel’s fundamental challenge predates Warsh by years.
https://www.youtube.com/watch?v=GcEMhiSHVpM
| Security | Close | Change | % Change |
|---|---|---|---|
| Intel (INTC) | $24.18 | ▼ | -2.2% |
| SOX (Semiconductor Index) | — | ▼ | -2.7% |
| NASDAQ 100 (NDX) | 19,840 | ▼ | -2.3% |
| S&P 500 | 5,580 | ▼ | -1.9% |
The 18A process node is the bet that matters most. Intel Foundry Services — the division Chief Executive Pat Gelsinger established to rebuild Intel as a manufacturing platform for third-party chip designers — has staked its external customer pipeline on 18A delivering competitive yield and performance characteristics. The node uses two technologies Intel considers differentiating: RibbonFET, its gate-all-around transistor architecture, and PowerVia, a backside power delivery system that its engineers say removes a significant bottleneck in chip performance.
Engineering reports circulating through the industry in August suggest 18A is progressing toward production readiness. The question is not whether the process is technically capable. The question is whether it is capable enough, at sufficient yield, to justify the switching cost a chip designer would incur by moving production from TSMC’s N2 or N3 nodes to an Intel facility. That is a higher bar than Intel’s promotional material acknowledges.
Gaudi 3, Intel’s AI accelerator, landed meaningful deployment at select enterprise customers this year, according to quarterly disclosures. Revenue numbers remain small relative to Nvidia’s data center dominance, and the product’s software ecosystem — the openness and maturity of its developer tools relative to Nvidia’s CUDA platform — remains the most cited friction point from potential adopters. AMD’s MI300X is making inroads in inference. Gaudi 3 is chasing a smaller slice of that same argument.
The valuation picture is different from most of the semiconductor sector. Intel trades at approximately 18 times the next twelve months’ consensus earnings estimate — a discount to AMD at 33 times and to the broader Philadelphia Semiconductor Index. That discount reflects the market’s skepticism about the turnaround timeline, not the potential outcome if 18A succeeds. If the foundry bet works, the current multiple is cheap. If it does not, the company’s capital structure and manufacturing overhead create a set of problems that are expensive to unwind.
Free cash flow was negative in Intel’s most recent fiscal year, as capital expenditure for the foundry buildout consumed revenue faster than the new business could offset it. The company has drawn on government support — CHIPS Act funding and tax credits — to partially offset that burn. Analysts who follow the stock closely estimate Intel needs 18A production revenue to begin scaling meaningfully by mid-2027 to avoid another round of capital raises or asset sales.
The S&P 500 fell broadly on Friday as Warsh’s remarks hit equity markets beyond the chip sector. Thursday’s session did not change any of Intel’s turnaround timelines. What it did was remind investors that the cost of capital for a company with Intel’s balance sheet and return profile is not theoretical. Warsh’s signal that rate cuts are not imminent means Intel’s interest expense on its outstanding debt remains elevated longer. For a profitable company running at high margins, that is a footnote. For a company in the middle of a capital-intensive turnaround, it is a variable that planning models cannot afford to round down.
Related Posts
