AI server suppliers seek higher prices after contracts
AI Weekly reports that Dell, HPE, Lenovo and Supermicro are seeking 25% to 45% more for Nvidia AI servers at delivery than the prices in signed contracts. The vendors cited memory and circuit-board shortages, according to the report.
Key takeaways · 4
- 01
Ask suppliers whether the contract allows repricing between quotation and shipment.
- 02
AI Weekly reports vendors cited memory and circuit-board shortages for the increases.
- 03
HPE expects elevated memory and storage prices to persist well into 2027, according to AI Weekly.
- 04
In the XTX dispute, Dell cited component-cost increases and proposed an additional $62 million to $70 million.
Reported changes to server pricing
AI Weekly reports that Dell, HPE, Lenovo and Supermicro are demanding 25% to 45% more at delivery for Nvidia AI servers than the prices fixed in signed contracts.[1] The report attributes the increases to memory and circuit-board shortages cited by the four vendors.[1] It also says quoting terms were amended to permit repricing existing orders when commodity costs rise between quoting and shipment.[1] For buyers reviewing a procurement or renewal, the key issue is whether their own signed terms contain a comparable adjustment provision; the report does not establish that every order is subject to repricing.
Cost pressures and supply outlook
The reported demands come amid sharp increases in component costs: AI Weekly says JPMorgan puts memory-cost inflation at more than 400% since early 2024, and server DRAM contract prices rose around 90% in the first quarter of 2026.[1] The same report says GPU-server prices rose at least 20% to 30%, while CPU-node prices tripled or more.[1] HPE expects elevated memory and storage prices to persist well into 2027, and a global circuit-board shortage is expected to last into 2028, according to AI Weekly.[1] These are reported market estimates and outlooks, not a guarantee of what any buyer will pay.
A separate Dell customer dispute
The Irish Times reports that XTX Finland OY brought High Court proceedings against Dell over a price increase involving 1,680 servers intended for a 15,000-square-metre data centre.[3] XTX says the agreed order price was about $549 million and that it paid an additional $5 million above the quoted price to avoid exposure to future increases.[3] After delivery times initially slipped by two months, Dell said component-cost increases meant XTX would owe another $62 million to $70 million, and suggested it might seek further increases under the agreement.[3] The judge admitted the case to a list, approved directions and adjourned it to December.[3]
Buyer responses and contract risk
Write-ups of the Financial Times report name Jane Street, Optiver and Hudson River Trading among trading firms affected, and say those firms have largely accepted the increases because they still need the equipment.[2] In the XTX matter, Dell said some customers facing increases switched suppliers before returning and agreeing to its adjustments.[3] Dell also said it was not willing to lose tens of millions of dollars on the transaction.[3] Together, these accounts suggest buyers may face a choice between negotiating, changing suppliers or accepting higher prices, while the contract wording and supply needs remain central to each decision.
For teams planning AI infrastructure, a quoted server price may not be the final cost if contract terms permit changes before shipment. Procurement, finance and legal teams should assess the price-adjustment language alongside delivery needs and alternatives.
Why it matters
AI training for your team
kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.
How this developed
22 November 2026
AI server suppliers seek higher prices after contracts
