AnswersPublished22 sources
AI’s Chip Crunch Has Moved Beyond GPUs to Power ICs, Memory and Phones
AI is creating a selective chip crunch beyond GPUs: demand for DRAM and mature node power chips is lifting prices and extending some PMIC lead times to roughly 35–40 weeks, while IDC expects 2026 smartphone shipments... The pressure is concentrated in components such as PMICs, display drivers and other specialty ICs...
Published byEdited with GPT-5.6 TerraImages generated with GPT Image 2
Research answer
Key takeaways
- AI is creating a selective chip crunch beyond GPUs: demand for DRAM and mature node power chips is lifting prices and extending some PMIC lead times to roughly 35–40 weeks, while IDC expects 2026 smartphone shipments...
- The pressure is concentrated in components such as PMICs, display drivers and other specialty ICs, not across every type of semiconductor.
- Powerchip’s July increases—45% for DRAM and 10–15% for power management and display driver chips—show how quickly the squeeze can reach consumer electronics supply chains.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: How is the AI boom driving shortages and price increases in mature-node semiconductors and memory—particularly power-management ICs, display. Article summary: The AI buildout is creating a two-layer supply squeeze: it consumes huge volumes of high-end memory and GPUs, while the power-delivery, control, and interface chips needed around those systems are mostly made on mature n. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
AI infrastructure requires more than advanced GPUs and high-bandwidth memory. It also needs large volumes of power-management, control and interface silicon—much of it produced on mature, specialty manufacturing processes. As AI-server demand absorbs capacity and memory supply tightens, the effects are showing up in the less visible chips that support data centers and in the prices and availability of smartphones. 6
44
53
The shortage is concentrated in mature-node and specialty chips
A power-hungry AI server needs extensive power conversion and regulation around its processors, memory, storage and networking. That increases demand for power-management ICs (PMICs), as well as management and related power components. TrendForce reporting cited by industry coverage says that constrained 8-inch BCD-process capacity is a central issue, with PMIC delivery times extending from roughly 21–26 weeks to 35–40 weeks. 44
These are not necessarily leading-edge chips. Mature-node capacity is still essential for analog, power-management and display-driver products, and its supply cannot simply be replaced by more cutting-edge logic output. As a result, AI-oriented server demand can compete with consumer, industrial, communications and automotive buyers for the same specialty production lines. 6
7
That distinction matters: this is not evidence that every semiconductor category is in a uniform shortage. One market outlook, for example, described mature logic more broadly as improving to 16–26-week lead times and stable to lower pricing, even as selected PMIC families were tightening. Component type, process technology, packaging and customer qualification all matter. 10
Foundries are raising prices selectively
The broad headline range of 5–20% should not be read as a single, industry-wide increase. Reported adjustments differ among foundries and products:
- TrendForce reported that SMIC had raised prices by about 10% for certain processes, while Vanguard International Semiconductor (VIS) was reportedly considering increases of roughly 4–8%.
6
- UMC reportedly planned second-half 2026 wafer-price adjustments, with earlier reporting suggesting an increase of about 10%; the company had not disclosed a universal rate.
- Counterpoint reported SMIC’s cumulative increases had pushed 8-inch mature-node pricing up 10–20%, while UMC was targeting 8–12% in specialty nodes.
13
- Nexchip was reported to have raised 8-inch foundry prices 10%, while broader 8-inch foundry increases were described as averaging 5–15% in the first half of 2026.
5
The common thread is improved pricing power where demand is strongest—especially for power-related and specialty processes—not a universal return to shortage-era pricing for every legacy node. 5
6
10
Powerchip illustrates how the squeeze spreads
Powerchip Semiconductor Manufacturing Corp. (PSMC) provided one of the clearest examples of the market shift. In July, the company said it was raising DRAM prices by 45% and lifting prices for power-management and display-driver chips by 10–15%. It also said structural supply shortages could extend into 2027. 18
Earlier in the year, PSMC reported separate price increases for display-driver chips: 30% for 12-inch display-driver production and 15% for 8-inch display-driver production, citing short supply. 26
Those moves matter because they connect two different parts of the electronics supply chain:
- Memory: AI computing and cloud deployments pull on DRAM capacity.
- Mature-node peripherals: The systems built around AI accelerators require power and control chips often made on 8-inch or specialty mature-node lines.
When both layers tighten at once, device makers face higher costs for core memory and for the components needed to operate, charge, display and manage their products.
Why the impact is now visible in smartphones
IDC attributes a sharp deterioration in its 2026 smartphone outlook to the memory crisis. It forecasts worldwide smartphone shipments will decline 16.7% to just over 1 billion units—the steepest annual contraction it has recorded—while average selling prices rise 27.6% to $581. Total market value is still expected to grow 6.3% to $613 billion, meaning higher prices rather than unit growth are carrying market value. 53
This is the downstream effect of a component squeeze. When DRAM and NAND become more expensive or harder to secure, handset manufacturers have limited options:
- absorb the extra bill-of-materials cost and accept lower margins;
- raise retail prices;
- concentrate supply on models with higher margins; or
- alter product specifications or sourcing plans.
The evidence supports the broad cost pressure, but not every claim circulating about individual phone markets or exact bill-of-materials percentages. Assertions about a universal memory share of premium-phone costs, specific India price increases or a fixed price effect in every market should be treated as market-specific estimates unless independently documented.
Lower-priced devices are particularly exposed because they have less margin available to absorb a memory or power-component shock. Reporting from South Korea, for example, said shipments of sub-$100 smartphones dropped about 60% year over year in the second quarter, alongside the wider increase in global smartphone pricing. 57
Why supply relief may take time
The tightest categories are difficult to resolve quickly because they depend on specific process technologies and production lines. TrendForce reporting points to ongoing demand for AI power-management components alongside reductions in some 8-inch capacity, a combination that has helped raise utilization and support price adjustments. 7
Memory pricing has also remained elevated. A TrendForce survey reported by Tom’s Hardware projected conventional DRAM contract prices to rise 13–18% quarter over quarter in the third quarter of 2026 and NAND prices to rise 10–15%, though at a slower pace than the approximately 60% jumps reported for the second quarter. 21
The most defensible outlook is therefore cautious: tightness in DRAM and selected mature-node power and display components can persist into 2027, as PSMC has forecast, but a shortage through 2028 or beyond is not established fact. It will depend on AI infrastructure spending, memory-output allocation, foundry capacity decisions and demand from phones, autos and industrial electronics. 18
What to watch next
The key indicators are more useful than a single “chip shortage” label:
- PMIC lead times: A continued range around 35–40 weeks would signal persistent pressure on AI-server power infrastructure.
44
- 8-inch and specialty-node utilization: This is the capacity pool relevant to many power and display components.
- DRAM and NAND contract-price direction: Rising memory prices flow directly into server and handset costs.
21
- Smartphone mix and pricing: IDC’s forecast of fewer units but higher average prices is a visible measure of cost pass-through.
53
The AI boom is not merely increasing demand for the most advanced processors. It is making power delivery, memory and mature-node support chips strategic constraints—and those constraints are increasingly being felt far beyond the data center.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "AI’s Chip Crunch Has Moved Beyond GPUs to Power ICs, Memory and Phones"?
AI is creating a selective chip crunch beyond GPUs: demand for DRAM and mature node power chips is lifting prices and extending some PMIC lead times to roughly 35–40 weeks, while IDC expects 2026 smartphone shipments...
What are the key points to validate first?
AI is creating a selective chip crunch beyond GPUs: demand for DRAM and mature node power chips is lifting prices and extending some PMIC lead times to roughly 35–40 weeks, while IDC expects 2026 smartphone shipments... The pressure is concentrated in components such as PMICs, display drivers and other specialty ICs, not across every type of semiconductor.
What should I do next in practice?
Powerchip’s July increases—45% for DRAM and 10–15% for power management and display driver chips—show how quickly the squeeze can reach consumer electronics supply chains.
Sources