Intel · one lens
Hamilton Helmer on Intel
Intel’s 18A ramp shows commercial progress, but the record still lacks evidence of a cost advantage that rivals cannot match.
Intel still does not clear Helmer’s Power Intensity test. The test requires a Superior and Significant cash-flow Benefit and a Sustainable Barrier against imitation. Intel moved 18A from roughly 5% yield—about five passing dies per 100 attempted dies—in Reuters’ August 2025 reporting to volume Panther Lake shipments, per the TEXXR dossier. Intel reported $16.1 billion in second-quarter 2026 revenue, up 25%, or a derived $3.22 billion, from a derived $12.88 billion in second-quarter 2025, per the dossier. The record supplies no foundry unit-cost, free-cash-flow or differential-margin series that establishes either the size or persistence of Intel’s advantage.
Intel’s D-series announcement adds a proposed data-center GPU for inference with up to 480GB of LPDDR5X per card; TechRadar supplies no earlier D-series capacity for comparison. Intel has no D-series shipment, customer-commitment or margin figure in this record. Intel therefore cannot treat a memory specification as evidence of Power. Intel’s fabs remain a Scale Economies candidate because higher volume could spread fixed costs, but the dossier establishes no foundry-volume or unit-cost lead over TSMC. Intel’s 18A ramp remains a Process Power candidate, yet reported Panther Lake and Wildcat Lake shortages leave its production performance unsettled. The shared High NA plan also gives Samsung and TSMC a path to the same equipment and larger photomasks. The record identifies no embedded Intel routine that rivals cannot copy.
Intel CEO Lip-Bu Tan told Nikkei that the Trump administration had no policy to use its minority stake to interfere with Intel’s strategy. Intel and Broadcom’s CEOs warned CNBC that borrowing costs and access to capital could constrain AI investment. The dossier says Intel funded its Fab 34 ownership buyback with new debt; the warning makes financing relevant to its scale attempt, not a Barrier. CTech reported a fatal gas-cylinder explosion at Intel’s Kiryat Gat fab, but supplied no measured effect on 18A yield.
The current TEXXR snapshot indexes 49 Intel articles in its completed July–September 2026 quarter, against an average of 43.2 articles per quarter over the preceding eight quarters. That measures coverage, not demand. Intel closed at $89.85 on October 16, down $5.52, or 5.79%, from the previous $95.37 close, ahead of its scheduled October 22 earnings. The price move does not establish a Power Barrier.
The seven Powers, one by one
| Power | Status | Why |
|---|---|---|
| Scale Economies | contested | fab volume could spread fixed costs, but the record shows no foundry-volume or unit-cost lead over TSMC |
| Network Economies | absent | not argued in the record |
| Counter Positioning | absent | the record shows no superior model that an incumbent refuses to copy |
| Switching Costs | absent | the record identifies no separate customer cost to leave Intel |
| Branding | absent | not argued in the record |
| Cornered Resource | absent | Samsung and TSMC can pursue the shared High NA equipment and photomask plan |
| Process Power | contested | 18A reached volume products, but shortages and missing differential-margin evidence leave durability unproven |
An editorial application of Hamilton Helmer's published framework — not their actual view, affiliation, or advice.
Where this lens comes from
Hamilton Helmer's framework is set out at the lens page, drawn from 7 Powers. This page applies it to Intel and nothing else — the company's full record is in the dossier.