TSMC does not rule out price rises as inflation pushes up chip manufacturing costs



TL;DR

TSMC’s CFO said inflation is raising costs and won’t rule out price rises, while its CEO said AI chip shortages will last years.

Taiwan Semiconductor Manufacturing Company, the world’s largest chipmaker, told the BBC that inflation is pushing up its costs and did not rule out raising prices. CFO Wendell Huang said the company would not impose sudden “fourfold, fivefold” increases but acknowledged that costs have risen. “We reflect our value,” he said, pointing to TSMC’s technology leadership and manufacturing scale.

The comments came after TSMC’s annual shareholder meeting in Hsinchu, where chairman and CEO CC Wei told investors he would “like” to raise prices, as competitors in the memory chip market have done. Wei said the AI chip shortage would persist for years and that TSMC cannot meet customer demand despite running its fabs at full capacity. “We’re doing everything we can, wherever we can, and however we can,” Huang told the BBC.

Despite the public commitment to pricing restraint, TSMC has already begun raising what it charges for its most advanced chips. TrendForce reported in May that the company is weighing a 15% price increase on 3nm wafers in the second half of 2026, with a further 5 to 10% increase expected in 2027. Earlier reports indicated TSMC planned to raise sub-3nm prices 3 to 10% across 2026, with increases mapped through 2029.

TSMC makes the most advanced chips designed by Nvidia, AMD, Apple, and virtually every other major semiconductor company, controlling more than 90% of the market at the most advanced manufacturing nodes. Any pricing increase ripples through to the cost of AI infrastructure, and potentially over time, to the prices consumers pay for electronic devices. The company reported $35.9 billion in revenue for the first quarter of 2026, up 41% year-on-year, with advanced technologies at 7nm and below accounting for 74% of wafer revenue.

The demand is being driven by an AI infrastructure spending wave that has seen Nvidia alone commit more than $40 billion in equity investments in 2026, while hyperscalers’ combined capital expenditure is projected to exceed $690 billion this year. That spending flows directly to TSMC’s order book, but the company’s costs are rising in parallel as it expands manufacturing capacity in the United States, Germany, and Japan alongside Taiwan.

Huang pushed back against the idea that TSMC’s global expansion was driven by political pressure from Washington or Beijing. “We go out of Taiwan to build capacity based on customers’ demand. The customers want us to go there. It’s not the request of government,” he said. TSMC has committed $165 billion to its Arizona operations alone, encompassing six fabrication plants, two advanced packaging facilities, and a research and development centre.

But on the question of where the most advanced chips will continue to be made, Huang was direct: the most cutting-edge production will remain in Taiwan. Moving the full manufacturing ecosystem to the United States, he said, would take “five or 10 years, or even longer,” a timeline that directly challenges the ambitions of US industrial policy. The first Arizona fab is producing 4nm chips, but 2nm production at the site is not expected until the end of the decade.

The geopolitical backdrop is impossible to ignore. At a recent summit in Beijing, Chinese President Xi Jinping warned that mishandling Taiwan could put the US-China relationship in an “extremely dangerous situation,” while President Trump described a $14 billion Taiwan arms package as a “negotiating chip.” Taiwan produces the majority of the world’s most advanced semiconductors, making TSMC’s fabs in Hsinchu among the most strategically important industrial facilities on the planet.

Huang also denied that the AI boom is a bubble. “Our conviction in this AI megatrend is very strong,” he said, citing conversations with customers and with the hyperscalers who are their largest buyers. “These companies are financially very strong with a lot of financial resources, so we believe that they’re able to continue to invest.” Whether that confidence extends to absorbing higher chip prices without passing costs to end users is a question TSMC’s customers will have to answer.



Source link

Leave a Reply

Subscribe to Our Newsletter

Get our latest articles delivered straight to your inbox. No spam, we promise.

Recent Reviews


“It was severely downgraded,” Gilbert confirms. “I never would have found it if I was just looking through Google results.” (I tried the same prompt in Gemini earlier this month, and after an initial denial, the tool also gave me Eiger’s number.)

After this experience, Eiger, Gilbert, and another UW PhD student, Anna-Maria Gueorguieva, decided to test ChatGPT to see what it would surface about a professor. 

At first, OpenAI’s guardrails kicked in, and ChatGPT responded that the information was unavailable. But in the same response, the chatbot suggested, “if you want to go deeper, I can still try a more ‘investigative-style’ approach.” Their inquiry just had to help “narrow things down,” ChatGPT said, by providing “a neighborhood guess” for where the professor might live, or “a possible co-owner name” for the professor’s home. ChatGPT continued: “That’s usually the only way to surface newer or intentionally less-visible property records.” 

The students provided this information, leading ChatGPT to produce the professor’s home address, home purchase price, and spouse’s name from city property records. 

(Taya Christianson, an OpenAI representative, said she was not able to comment on what happened in this case without seeing screenshots or knowing which model the students had tested, even after we pointed out that many users may not know which model they were using in the ChatGPT interface. She also declined to comment generally about the exposure of PII by the chatbot, instead providing links to documents describing how OpenAI handles privacy, including filtering out PII, and other tools.) 

This reveals one of the fundamental problems with chatbots, says DeleteMe’s Shavell. AI companies “can build in guardrails, but [their chatbots] are also designed to be effective and to answer customer questions.”

The exposure issue is not limited to Gemini or ChatGPT. Last year, Futurism found that if you prompted xAI’s chatbot Grok with “[name] address,” in almost all cases, it provided not only residential addresses but also often the person’s phone numbers, work addresses, and addresses for people with similar-sounding names. (xAI did not respond to a request for comment.) 

No clear answers

There aren’t straightforward solutions to this problem—there’s no easy way to either verify whether someone’s personal information is in a given model’s training set or to compel the models to remove PII. 



Source link