Friday, June 26, 2026
EN·DarkSubscribe
AI Infrastructure · News & Analysis
HomeChips & HardwareReport
Chips & Hardware · Report

AMD defends the fundamentals of the AI chip market and attributes recent price increases to rising chip manufacturing costs.

Reveals cost-push inflation in semiconductor production rather than pure demand dynamics, affecting capital budgets and procurement planning for infrastructure operators.
Trade pressSlicast · September 4, 2025 · Global · Source: wccftech.com
importance 65

AMD CFO Jean Hu confirmed at a Citigroup conference that the firm is not starting new chip production for its China products despite receiving licenses from the Trump administration. Speaking alongside Matthew Ramsay, AMD's vice president of investor relations, Hu stated that AMD is not "starting new wafers for MI308," the company's China-specific GPU. Instead, AMD is focused on clearing out existing inventory and working to determine "if Chinese customers can be allowed to buy from U.S." Further investment in Chinese products depends on whether AMD "can get the license for our next generation," she added. The discussion also noted that AMD, like NVIDIA, was forced to write off its China GPU inventory due to US license restrictions.

The executives acknowledged significant market demand constraints in China. Ramsay stated that "I think inside of China, there's a larger demand for AI processing silicon than there is ability to manufacture that silicon in China." Chinese chip manufacturers face constraints from US sanctions preventing the production of leading-edge chips, and the country's largest tech firm, Huawei, cannot procure leading-edge chips manufactured with US-origin technology. Despite AMD's desire to serve this demand, Ramsay noted that "getting visibility in the short term as to what that looks like is -- given all the moving parts, has been the challenge."

When asked about potential AI chip over-ordering and an AI bubble, Hu cited "tremendous evidence for AI adoption" and second-quarter capital expenditure by hyperscalers as proof that firms "have improved their return on their investment across not only their platforms." She noted that "We hear a lot of other companies adopting AI," though she believes that "we are still at a very early stage for AI adoption." Regarding rising AI chip prices, Hu explained that AMD adds new features with each product iteration, increasing the bill of materials and selling prices, but the company focuses on delivering improved total ownership cost to customers while maintaining healthy gross margins. Hu emphasized that "in the longer term, there are ups and downs over each cycle," yet maintains that AI represents "probably once a lifetime opportunity we're seeing."

Ramsay addressed the factors driving AMD's $500 billion total addressable market estimate for AI, citing inference use cases, the size of customer data sets, and their extension into various industries. He stated that "I've said this for a long time that I think as T goes to infinity, right, more and more CapEx and OpEx dollars in basically every industry goes into high-performance computing, AI is a significant inflection of that." Ramsay characterized AI as "the biggest inflection in computing since the invention of the Internet," suggesting that AMD's TAM could extend beyond the $500 billion figure outlined by CEO Lisa Su earlier this year.

Read the original
AMD defends the fundamentals of the AI chip… · Slicast