AI Server Price Hike Hits Up to 45% After Contract Signing

0
15

AI server price hike of up to 45% is affecting buyers even after contracts are signed, as tech companies pass on soaring memory costs. Major vendors like Dell, HPE, Lenovo, and Supermicro have reportedly increased prices on AI servers built with NVIDIA chips just before delivery, catching large customers off guard.

Post-Contract Price Increases Shake AI Infrastructure Market

According to sources cited by the Financial Times, suppliers are raising the cost of AI servers months after initial deals are finalized. These servers, which include GPUs, memory chips, and networking components, are sold to companies building their own data centers. While agreements generally occur well before delivery, vendors have been demanding up to 45% higher prices once shipments are ready.

The surge in prices is linked to a sharp rise in memory costs—over 400% since early 2024—that vendors are passing on to buyers. Large financial services firms and trading companies, despite being significant purchasers second only to tech hyperscalers, have little leverage against these last-minute price hikes. One insider described the situation as “disturbing” and noted the vast gap between the contract price and the final payment after delivery.

Legal Disputes Highlight Pricing Conflicts

This ongoing pricing volatility has sparked legal battles. In June, the UK-based trading firm XTX sued Dell in an Irish court after Dell sought to increase the server price by tens of millions of dollars beyond a previously agreed fixed price of $549 million. XTX had also paid an extra $5 million to secure price certainty. Dell justified the hike by citing rising component costs, particularly memory, and asserted it could not absorb these higher expenses.

According to court documents, Dell told XTX that while other buyers initially resisted the hikes and explored alternatives, they eventually accepted the increased prices. The case remains unresolved, spotlighting the challenges buyers face in securing stable pricing amid explosive demand for AI computing power.

Source: https://www.ithome.com/1/030/717.htm

LEAVE A REPLY

Please enter your comment!
Please enter your name here