Samsung, SK Hynix stand to benefit as Nvidia AI server costs surge
Server manufacturers notified major data center customers, including Microsoft, Google, and Oracle, that prices for Nvidia-based AI systems would rise by more than 15% in many cases for shipments starting in early 2027. The price adjustments affected systems built on Nvidia's next-generation Vera Rubin and current Grace Blackwell platforms, with the exact scale of the increase varying by chip…
Samsung Electronics and SK hynix are gaining greater leverage in the artificial intelligence supply chain as memory costs rise, according to a Korea Herald report. Server manufacturers have informed major data center customers, such as Microsoft, Google, and Oracle, that Nvidia-based AI system prices will increase by more than 15% starting early 2027.
The price hikes apply to systems built on Nvidia's new Vera Rubin and current Grace Blackwell platforms, with the increase's magnitude differing by chip generation and memory configuration.
Memory components have become a key factor in the cost pressures. Nvidia's AI accelerators rely on high-bandwidth memory (HBM), while AI data centers consume significant amounts of standard server DRAM. Supply constraints persist as cloud service providers expand computing capacity and secure memory supplies through longer-term contracts, giving a handful of memory manufacturers a stronger negotiating position.
SK hynix accounted for 58% of global HBM revenue in the first quarter, while Samsung led the broader DRAM market with a 39% share in the second quarter. Together, the two South Korean chipmakers represent 65% of global revenue.
SK hynix gained an early advantage in HBM, which is directly tied to AI accelerators, while Samsung held a larger portion of the conventional DRAM market and expanded its share in advanced HBM products. Counterpoint Research reported that the majority of first-quarter HBM revenue came from HBM3E, with shipments of HBM4 expected to become more significant in the second half of the year.
The cost pressures also impacted the wider DRAM sector. TrendForce forecasted that server DRAM contract prices would rise by an additional 13 to 18% in the third quarter compared to the previous quarter, driven by continuous AI-related absorption of manufacturing capacity. Hana Securities pointed out that the reported price increases for Nvidia-based servers drew renewed focus on memory suppliers, highlighting Samsung and SK hynix as the primary beneficiaries amid ongoing infrastructure spending projections through 2027.
Despite the rising demand for high-performance memory, companies like Amazon, Microsoft, Google, and Meta are developing their own AI processors, which also require advanced DRAM and HBM. Nonetheless, escalating component prices have led to overall infrastructure expenses growing. According to a The Information report, server suppliers warned customers of approximately 17% price hikes on major Nvidia-based systems, estimating that chip-system costs for a 1-gigawatt data center could rise by at least USD 5 billion.
LS Securities head of research, Shin Joong-ho, emphasized the constraints of rising component expenses, stating that while big tech can handle higher costs for some time, there is a limit to how high infrastructure spending can go without affecting returns. If memory inflation continues to push up Nvidia system prices, companies may need to reconsider the expected returns on new AI data center projects.
Written by urgent.news from Economic Times Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
