The semiconductor rally that carried markets through early October hit its first real speed bump this week, as fresh questions about the pace of AI revenue pulled chip stocks off record highs. After the iShares Semiconductor ETF advanced for five consecutive sessions and the sector's largest company reached a record high, sentiment shifted sharply on Thursday — when the Financial Times reported that OpenAI had told investors its annualised revenue was nearing $50 billion at the end of September, below the approximately $70 billion figure that had been cited in media reports, according to a weekly market analysis.

Because semiconductor valuations are closely tied to expectations for AI infrastructure spending, the report weighed on the group: the PHLX Semiconductor Index finished the week lower, while the broader market proved more resilient. It was a reminder that after a historic run — the index was up 73% year to date as of late September, with several of its largest constituents posting triple-digit gains in 2026 — even a hint of a gap between AI investment and AI revenue is enough to unsettle the trade.

Why a revenue number rattled the market

The logic of the chip-stock boom is simple and fragile: semiconductor companies have traded at premium valuations on the premise that AI customers will keep spending ever-larger sums on data-centre hardware, and that this spending will eventually convert into sustainable revenue. A report suggesting the world's most prominent AI company is tracking below some publicised expectations struck directly at that premise — not because $50 billion is a small number, but because the market had priced in more.

The pullback was compounded by macro headwinds midweek. Higher long-term Treasury yields and a rise in crude oil prices added pressure, and higher discount rates tend to weigh most on companies valued on long-term growth expectations — a description that fits the 2026 chip sector almost perfectly.

A week of two halves

The week had started strongly. Two companies in the group reported record milestones early on, and the sector's momentum looked intact. But as the week wore on, company-specific concerns joined the macro narrative: open questions around a manufacturing partnership, ongoing litigation, and — most fundamentally — the sustainability of AI-related demand all shaped performance within the group.

Active names under scrutiny this week included Taiwan Semiconductor Manufacturing Company, Nvidia, Intel, Arm Holdings and Micron Technology, according to the analysis — a roster that spans the full AI hardware stack, from foundry capacity to memory and processor design.

The consolidation around them continues

The chip-stock wobble is unfolding against a backdrop of accelerating consolidation in the infrastructure layer beneath AI. As WebProNews reported, IBM transferred contracts for hundreds of enterprise customers running VMware Cloud Foundation workloads to 11:11 Systems, with the deal closing on September 30. The move comes ahead of a March 2027 cutoff imposed by Broadcom after its acquisition of VMware — a reminder that the platform economics of enterprise computing are being rewritten in real time, and that the chips at the centre of it all carry concentrated expectations.

Related coverage on this site has tracked the tension from the other side: the AI debt wave swelling around Broadcom, SpaceX and Oracle shows how aggressively capital is being deployed to build AI capacity, while a recent AP-NORC poll found many Americans believe AI is developing too fast — a signal that the political and social license for the spending is not unlimited either.

What investors are really asking

The deeper question behind Thursday's selloff is one the market has deferred for two years: when does AI spending start looking like AI profits? OpenAI's reported figure remains staggering by any historical standard, but markets in 2026 are no longer grading on a curve of amazement — they are grading on a curve of justification. Every chip shipped needs a corresponding stream of revenue somewhere downstream, and any evidence that the conversion rate is slower than hoped forces a repricing of the entire supply chain.

None of this undoes the fundamental demand story. Data-centre buildouts continue, and the structural need for compute is still expanding. But the October pullback has introduced a new discipline into the trade: the era when AI headlines alone lifted chip stocks may be giving way to an era when the numbers have to do the lifting.

Key takeaways

  • The PHLX Semiconductor Index fell in the week ended October 9 after five straight sessions of gains, while the broader market held up better.
  • The trigger was a Financial Times report that OpenAI told investors its annualised revenue was nearing $50 billion — below the roughly $70 billion figure cited in some media reports.
  • The index was up 73% year to date as of late September, leaving valuations exposed to any hint of a revenue-spending gap.
  • Higher Treasury yields and rising crude prices added midweek macro pressure on growth valuations.
  • Company-specific issues — from manufacturing-partnership questions to litigation — widened the dispersion within the group.

Frequently Asked Questions

What caused semiconductor stocks to fall this week?

Chip stocks retreated after the Financial Times reported that OpenAI had told investors its annualised revenue was nearing $50 billion at the end of September — below the roughly $70 billion figure that had appeared in media reports. Because semiconductor valuations are tightly linked to expectations for AI infrastructure spending, the news weighed on the sector and the PHLX Semiconductor Index finished the week lower.

How much had semiconductor stocks risen before the pullback?

The PHLX Semiconductor Index was up 73% year to date as of late September 2026, with several of its largest constituents recording triple-digit gains during the year — a run that left valuations vulnerable to any sign that AI revenue was not keeping pace with AI spending.

Which companies were most affected?

The weekly analysis named Taiwan Semiconductor Manufacturing Company, Nvidia, Intel, Arm Holdings and Micron Technology among the actively watched companies, spanning foundry capacity, processors and memory across the AI hardware stack.

Does this change the long-term AI demand story?

Not fundamentally — data-centre buildouts and the structural need for compute continue to expand. What the pullback changes is the market's tolerance: investors are now demanding clearer evidence that massive AI spending is converting into durable revenue, rather than rewarding the spending itself.