AMD's $17 billion fourth-quarter forecast lifts chip equipment and packaging outlook

AMD puts a number on AI demand

For chip buyers and investors, that is the clearest near-term number showing AI data-center spending has not stalled. For equipment, packaging and materials suppliers, the follow-on forecasts point to demand spreading into test, assembly and advanced packaging. What is still unknown is how much of the demand rests on a few hyperscale customers, and whether suppliers can add capacity fast enough to meet it.

AMD shares rose about 7% in late trading after the forecast. Bloomberg described the company as a key challenger to Nvidia in AI chips. The projections that follow it run to 2028 and 2032. They come from industry associations and research firms, not from chipmakers' order books.

DBS senior economist Han Teng Chua said AI-related stocks have seen considerable volatility recently amid questions over whether that spending can be sustained.

Equipment and packaging forecasts follow

SEMI expects worldwide semiconductor equipment sales of $165.9 billion in 2026, a 23.2% increase over 2025. The industry association sees the figure reaching an all-time high of $229.5 billion by 2028.

TrendForce, cited in the same Globe and Mail item, estimates capital spending by the nine biggest cloud companies will reach roughly $830 billion in 2026. That spending is what turns into orders for the tools. Test equipment sales are forecast to rise 31.0% in 2026 to $15.3 billion, after a 55.3% gain in 2025. Assembly and packaging tools are expected to grow 9.6% this year to $6.7 billion.

Materials sit further down the chain. ResearchAndMarkets puts packaging materials at $51.22 billion in 2026, growing at a 10.72% compound annual rate to $94.76 billion by 2032.

It ties that growth to heterogeneous integration and chiplet architectures, which it says are accelerating demand for high-density substrates, interposers and fine-pitch solder materials. Low-warpage substrates, low-loss dielectrics and thermal interface materials are listed as priorities. Compliance with RoHS and REACH rules is also shaping material selection, the report says.

Edge chips and teardown work

The Business Research Company projects AI on edge semiconductors at $8.13 billion in 2026, up from $6.89 billion in 2025. It expects the market to reach $15.5 billion by 2030, a 17.5% compound annual rate.

The same release puts growth from 2025 to 2026 at 17.9%, slightly faster than the 17.5% rate it projects for the rest of the decade. Edge chips run inference on the device instead of in a data center, and the release credits IoT deployment and demand for low-latency processing.

Teardown work is a smaller market pointing the same way. ResearchAndMarkets projects semiconductor teardown services at $795.34 million in 2026 and $1.23 billion by 2032, a 7.51% compound annual rate. It lists reverse engineering, intellectual property diligence, failure analysis and counterfeit detection among the applications. It links demand to advanced packaging, chiplets, through-silicon vias and growing supply chain scrutiny.

At 7.51% a year, that is the slowest of the growth rates above.

Singapore manufacturers see the order book

The Singapore Economic Development Board's quarterly survey was released on Monday 2 November. It found 29% of manufacturers expect business conditions to improve between October and March 2027, while 5% expect them to weaken. Optimists outnumbered pessimists by 24 percentage points, up from 12 points in the previous survey.

Precision engineering was the most positive cluster, with optimists ahead by 48 percentage points, supported by investment growth in semiconductors and AI. Electronics was also positive, driven by AI applications and semiconductor demand. The survey covered 397 manufacturers, with 83% responding, and ran between September and October. The chemicals industry went the other way, with pessimists ahead by 13 percentage points, and petroleum firms cited uncertainty over crude oil feedstock supply.

Capacity is the constraint

Chua said AI capital expenditure is likely to remain strong for the rest of the year, citing recent earnings reports from major US hyperscalers.

Chua also said limited capacity could remain a major challenge for firms seeking to expand production. That caveat runs against the supply chain forecasts above. The equipment and packaging numbers are third-party projections from SEMI and ResearchAndMarkets. None of them settles how concentrated the underlying demand is among a handful of hyperscale buyers.

Singapore's precision engineering and electronics clusters are the part of the survey most exposed to that question, since both depend on orders tied to AI data centres and semiconductor equipment.