
Semiconductor pricing pressure continues to spread across markets as suppliers contend with higher material, energy, and manufacturing expenses. Renesas is reportedly preparing another round of increases for January 2027 as cost inflation remains embedded in mature and widely used semiconductor categories.
In 2026, price increases expanded beyond components directly tied to AI-driven shortages with raw material inflation and tighter mature-node capacity pushing costs upward. TrendForce expects these pressures to remain difficult to avoid into 2027 as AI demand absorbs additional manufacturing capacity.
Furthermore, Chinese suppliers are raising prices by double digits as restricted access to HBM forces them to secure supply through more expensive channels. Huawei’s newest Ascend 950DT AI chip has reportedly risen as much as 50% from quotations issued only two months earlier, as memory availability influences the cost of finished AI hardware.
As cost pressures on the manufacturing side continue to mount, Tokyo-based Renesas Electronics is reportedly planning another round of price increases for the start of 2027. This latest move follows a hike that took effect in July.
According to STAR Market Daily, the January 1, 2027, increase is due to ongoing structural cost increases across raw materials, energy, advanced packaging, and global logistics.
Renesas organizes its business around four key segments, with a product portfolio spanning analog, power, embedded processing, and connectivity for the automotive, industrial, infrastructure, and IoT markets. Given this wide range of product lines, price increases expose numerous downstream manufacturers.
Renesas is not alone in facing cost pressures, however. Infineon raised its prices on power switches and related products by 5% to 15% in April before issuing a second adjustment in July, according to TrendForce.
Meanwhile, a TechNews report confirmed that Texas Instruments has informed customers of price adjustments across multiple product lines. Chinese suppliers, including CR Micro and Silan, have also raised prices across their portfolios by as much as 15%.
Notably, the pressure also extends beyond power semiconductors. Average 8-inch foundry prices rose 5% to 15% between the first and second quarters of 2026, according to TrendForce data. As foundries scale back older production lines, the broader mature-node market is already experiencing upward pressure. With AI-driven demand for power and analog products consuming a large share of wafers, capacity is expected to further decline.
Costs are compounding at the national level in Japan, with its manufacturing sector recording historically sharp input cost inflation in August 2026, according to S&P Global’s latest report. Driven by higher raw material and energy prices as well as a weak yen, these macro pressures have worsened the costs Renesas cited in its price increase announcement.
For semiconductor buyers, this pattern of overlapping rising costs has become a structural problem to navigate with little hope for near-term relief. Sourceability helps customers ensure supply continuity through real-time price and lead time monitoring. Combined with access to global inventory from a network of vetted suppliers, this visibility helps organizations keep production moving and reduces exposure to sudden price spikes and delays.
Huawei has raised the price of its Ascend 950DT AI accelerator to roughly $37,255, according to Reuters. That jump is reportedly 20% to 50% higher than quotes customers received just two months ago for the card expected to debut in Q4 of this year.
Since HBM accounts for a significant share of an AI accelerator’s total production cost, inflated input expenditures are feeding directly into finished card prices. That’s true of both Huawei’s Ascend 950DT and its earlier Ascend 950PR and Ascend910C, which have seen price increases of roughly 30% and 22% respectively since the start of 2026.
Huawei isn’t the only Chinese chipmaker raising prices. According to TrendForce, Cambricon has repriced its next-generation 690 chip ahead of its release by between 20% and 30% more than quotes indicated two months ago. Smaller suppliers MetaX and Iluvatar CoreX have made comparable increases.
The common thread connecting each of these price hikes is the global HBM shortage, which is compounded by U.S. export restrictions on advanced memory to China. With supply from primary channels unavailable, Chinese chipmakers have turned to gray-market sources to secure HBM. Doing so means paying several times the price available to buyers outside China.
In the face of these challenges, China is working to reduce its dependency on imported HBM. CXMT has reportedly begun small-scale production of HBM3E and plans for a larger ramp in 2027. Chinese chip designers Cambricon and Alibaba’s T-Head are reportedly testing the chips for their own processors.
Tight AI accelerator supply is also changing how existing chips are allocated. Suppliers are increasingly redirecting available GPUs to major customers, including Iluvatar CoreX. According to Reuters, the chipmaker doubled its shipments to ByteDance this year to 100,000 units, in part by reassigning chips it originally intended for internal use.
HBM scarcity and export controls are the main drivers behind this recent wave of cost increases, but AI demand itself can’t be overlooked. AI data center expansion is consuming leading-edge memory and packaging capacity at unprecedented rates while simultaneously increasing utilization of mature-node power management and supporting components. TrendForce notes that, with foundries redirecting more wafer capacity to higher-margin AI-related products, pricing pressure is likely to continue across mature technologies.
As the memory shortage extends into 2027, buyers relying on short-term purchases will likely see more volatile pricing than those with long-term deals or preferred allocation. Sourceability can help customers monitor memory market conditions and identify alternative supply channels before a shortage in one critical technology or major price spikes cascade into broader production delays.