Field report S-17 / R7 · sheet /strategy
Corporate StrategyFiled from the territory
Semiconductor revenue rose 150%, but unit shipments gained only 9.9%
September semiconductor revenue rose 150% year on year, while shipments increased 9.9% and average selling prices climbed 127.5%, exposing capacity-planning risks.
- Grid
- S-17 / R7
- Surveyed
- November 20, 2026
- Reading
- 4 min read
- Correspondent
- Elena Vasquez

Field notes
- Future Horizons’ November report put September semiconductor revenue growth at 150% year on year.
- Unit shipments rose 9.9%, while average selling prices increased 127.5%.
- Memory revenue rose 383%, with units up 52.4% and average selling prices up 217%.
- Five planning variables require separate review: revenue, shipments, product mix, customer concentration and capacity commitments.
By Elena Vasquez
September semiconductor revenues increased 150% year on year, but unit shipments rose only 9.9%, according to Future Horizons’ November report, exposing a widening gap between market value and physical demand. Average selling prices increased 127.5% over the same period.
Future Horizons framed the risk in direct terms: “Revenue measures what customers spend. Unit demand helps explain how much production capacity they need.” Its conclusion for factory plans, equipment purchases and supply commitments is that dollar growth can overstate the physical market.
The report acknowledged growing demand but asked how much of the AI boom comes from additional chips, how much from higher selling prices, and how heavily the industry depends on one concentrated source of spending. That composition determines how much revenue can persist when scarcity eases.
How much of the AI boom comes from price?
The arithmetic already indicates that price contributed more than volume to September’s revenue increase. The 127.5% price increase was nearly 13 times the 9.9% shipment gain. A company can interpret 150% revenue growth as evidence of a market requiring proportionately more wafers, devices and production equipment. The 9.9% shipment increase describes a far smaller physical expansion.
That mismatch can encourage manufacturers to build for a market that appears much larger in dollars than in wafers or devices. If selling prices weaken, revenue supporting investment assumptions can disappear faster than manufacturers can bring capacity online. The danger lies in mistaking a price cycle for durable volume.
Memory illustrates the exposure. Revenue increased 383%, combining a 52.4% increase in units with a 217% increase in average selling prices. Demand grew, but pricing contributed heavily to the headline result, leaving the segment vulnerable if supply expands, purchasing patterns change or customers resist higher costs.
Who can absorb higher component prices?
Future Horizons’ competitive analysis found no uniform recovery. Hyperscalers building AI infrastructure can tolerate component prices that create serious difficulties for more price-sensitive applications. Smartphone and PC manufacturers face different economics. Higher memory costs can squeeze margins, force specification changes or increase retail prices.
Those responses can weaken demand elsewhere in the electronics market, even while AI infrastructure spending remains strong. A powerful AI investment cycle can therefore coexist with fragile conditions among other customer groups.
The consequences extend across the supply chain. Equipment suppliers, packaging companies and test providers all commit resources against customers’ expansion plans. Aggregate market growth offers limited guidance about which technologies, production stages and applications will sustain orders, increasing the risk of misplaced capital.
Treating headline growth as evidence of universal recovery could obscure those differences. Future Horizons’ warning also carries a concentration risk: strong demand tied to a narrow spending stream does not necessarily translate into sustained orders for every supplier or every end market.
Can new capacity arrive after the pricing peak?
Semiconductor capacity takes time to develop, while customer confidence and purchasing intentions can change quickly. Companies committing capital during shortages may bring capacity online after the strongest pricing conditions have passed. Future Horizons’ downturn forecast remains a forecast, but the mismatch between investment lead times and market adjustment matters regardless of its precise timing.
The timing problem applies across the supply chain. A plant, packaging line or test facility commissioned at peak pricing can face weaker economics when it starts production. The report’s concern is structural rather than tied to one forecast date: investment commitments can outlast the conditions that justified them.
Which computing approach will investors favor next?
The report also compared current AI infrastructure with the historical contest between electrical architectures. That comparison raises a separate strategic question: will the technologies attracting the most capital today remain the preferred way to deliver AI tomorrow? It does not establish which architecture will win.
The analogy instead warns against assuming that current resource requirements will scale indefinitely. Better efficiency, different computing approaches or changes in workload economics could alter demand for particular devices without diminishing AI’s broader importance. Capital allocation tied to one current design could therefore be misplaced even if AI adoption continues.
What should semiconductor executives measure?
Future Horizons recommended more granular planning, with five variables assessed separately:
- Revenue, to show what customers spend.
- Shipment volumes, to indicate physical demand.
- Product mix, to identify which technologies and applications are growing.
- Customer concentration, to expose dependence on a narrow spending base.
- Capacity commitments, to test whether investment matches durable demand.
Investment cases should remain credible under less favorable pricing and slower infrastructure spending. That requires examining what customers actually need, what they can sustainably afford and which investments remain useful when scarcity no longer supports exceptional prices.
The practical test is not whether AI’s future remains intact. It is whether each company’s capacity, product and customer assumptions can survive a less exceptional pricing environment.
Future Horizons’ conclusion is blunt: transformative technology can produce both durable opportunity and cyclical excess.
Disciplined investment in what customers need and can afford will shape resilience across the semiconductor supply chain.
via semiwiki.com (Source)
More from Elena Vasquez
Show full bio
Staff writer covering companies, deals and strategy in corporate strategy and chief-executive leadership at CEO Frontier.
514 articles
Adjacent sheets
- Applied Materials Posts Record $10.41 Billion Quarter on AI Demand
- S&P 500 to Sit 17% Off Peak by End-2027, Capital Economics Warns
- China's Trade Surplus Hits $133.5 Billion, Second Largest on Record
- TSMC Posts 59.7% October Revenue Jump as AI Spending Accelerates
- Memory shortage may tighten through mid-2027, delaying price relief