Samsung Electronics has bumped up prices for a bunch of advanced contract chipmaking services by around 15% in some places, and it’s showing the sort of mounting pressure that artificial intelligence demand is putting on the world’s semiconductor manufacturing capacity. These price adjustments hit new orders as well, and they’re arriving just as demand for advanced processors keeps climbing rapidly.
Samsung’s decision shows how the AI boom is bending the cost logic of semiconductor manufacturing, especially as clients fight for scarce leading-edge production capacity. According to people familiar with the matter, Samsung raised prices for its 4-nanometre SF4 process by roughly 10% to 15% for customers in China and the United States, though there was some variation.
For customers in Taiwan, prices reportedly rose by around 5% to 10%. The company also increased prices for its 5-nanometre process by 10% to 15%; meanwhile, pricing for older 8-nanometre production rose by nearly 10%, give or take.
The increases seem to land at a pretty significant time for Samsung’s foundry business. Taiwan Semiconductor Manufacturing Co. continues to hold sway in the global contract manufacturing market, with more than 70% of foundry revenue in the first quarter of 2026, while Samsung sits at roughly 7%, according to Counterpoint.
Still, there’s tight capacity at TSMC, and that is basically opening doors for other players. Tech companies are sprinting to build AI accelerators, high-performance computing chips and a bunch of other advanced processors, so customers are starting to look around for extra manufacturing room, and not just the usual.
Samsung’s SF4 production line in Pyeongtaek, South Korea, has reportedly been running at full tilt since late last year. That plant makes logic chips for customers like Qualcomm and also supplies parts for Samsung’s high-bandwidth memory products.
This company expects those advanced process routes to account for more than half of its foundry revenue this year, and yes, AI and high-performance computing applications are forecast to add over 30%, up from around 15% to 20% in late 2025. Pricing power matters a lot, too, because Samsung’s foundry arm has been struggling financially for a few years. If prices go up, margins could improve, which might even help the unit move more quickly toward actual profitability.
At the same time, the semiconductor landscape is being shaped by geopolitical constraints. US rules on exporting advanced semiconductor manufacturing gear to China have pushed some Chinese technology firms to rely more on overseas foundries, which in turn adds strain to the existing capacity pool.
For chip designers and technology companies, higher manufacturing costs could make building AI hardware more expensive. And those higher development costs may, in the end, show up in how servers, data centre infrastructure, and AI services get priced.
For Samsung, though, the current situation gives it a chance to strengthen its competitive position against TSMC, and it could also improve the economics of its foundry operations.
This development kind of shows how the AI boom is creating ripple effects across the global tech supply chain. Artificial intelligence, of course, can be driven by software, but its actual expansion now leans more heavily on limited real-world resources, namely scarce advanced semiconductor manufacturing capacity.
