Skip to main content

SMIC expects AI chip supply to stay tight through 2027

13 NOVEMBER 2026·2 MIN READ·4 SOURCES

SMIC said the AI-related semiconductor supply chain is likely to remain tight through 2027, citing strong demand and shortages of computing resources. Co-CEO Zhao Haijun made the remarks at the company’s third-quarter earnings briefing on November 13.

SMIC expects AI chip supply to stay tight through 2027

Key takeaways · 4

  • 01

    SMIC expects AI-related semiconductor supply to remain tight through 2027, with data-center and private-cloud demand contributing to the squeeze.

  • 02

    SMIC said some equipment orders take more than 24 months to arrive, affecting its capital spending this year.

  • 03

    SMIC expects depreciation expense to rise about 30% year over year in 2026.

  • 04

    For the fourth quarter, SMIC forecast sequential revenue growth of 4% to 6% and a gross margin of 26% to 28%.

What SMIC said about demand

SMIC, China’s largest chip foundry, said the AI-related semiconductor supply chain would remain tight through 2027.[1] Zhao made the remarks while answering a question about whether AI was in a bubble at the company’s third-quarter earnings briefing.[1] SMIC said the AI industry was still in its early stages and demand remained strong.[2] Zhao added that many customers seeking computing resources continued to face shortages.[2] He pointed to demand from data centers and private clouds as a reason the supply chain for AI-related applications would stay tight.[2]

Constraints may extend beyond 2027

The company’s timeline is not the only indication of pressure on supply: the industry broadly expected production capacity to be severely constrained in 2027, with the situation likely to persist until mid-to-late 2028.[1] Supply conditions may also differ by component.[1] AI-related memory supply and demand could broadly balance after mid-2028, but computing and other supporting chips could remain in short supply.[1] For teams planning around AI hardware, that distinction matters: an easing in memory would not, on this evidence, mean that all components needed for AI-related applications were readily available.[1]

Equipment lead times and spending

SMIC said some equipment orders were taking more than 24 months to deliver, affecting its capital spending this year.[2] Zhao also said the company’s depreciation expense would rise about 30% year over year in 2026.[2] These figures point to a capacity-planning challenge alongside the demand outlook: equipment delivery timelines can shape when expansion spending translates into added production, while higher depreciation is part of the company’s stated 2026 cost picture.[2] SMIC said it expands capacity in response to customer demand.[3]

Results and near-term outlook

SMIC reported third-quarter revenue of $3.19 billion, up 6.1% from the previous quarter and 33.9% from a year earlier.[3] Its gross margin was 29.9%, compared with 25.3% in the previous quarter and 22.0% a year earlier.[3] Capacity utilization rose to 97.3%, from 93.7% in the previous quarter.[4] Looking ahead, SMIC forecast fourth-quarter revenue growth of 4% to 6% sequentially and a gross margin of 26% to 28%.[3] The company also said it would adjust prices in line with industry trends and negotiate price increases with customers.[4]

For organizations planning AI infrastructure, SMIC’s outlook is a reason to test supply assumptions against a multi-year horizon rather than treating memory, computing and supporting chips as one interchangeable category. Procurement and finance teams can use the stated equipment lead times, capacity plans and company guidance to frame questions for suppliers and internal forecasts.

Why it matters
For teams

AI training for your team

kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.

How this developed

  1. 13 November 2026

    SMIC expects AI chip supply to stay tight through 2027

Sources

AI fluency, one session a day, built for your work.