Some of Nvidia’s largest customers have reportedly been told that prices for servers containing its AI chips will rise by more than 15 percent in many cases. The increases are being linked primarily to higher memory costs and are expected to affect systems shipped from early next year.
The development matters because the economics of AI are influenced by more than model pricing. Organisations running private models, large inference environments or dedicated AI infrastructure remain exposed to changes in the cost of chips, memory, energy and data centre capacity.
Key facts
- Some Nvidia-based AI servers are reportedly facing price increases above 15 percent.
- The increases are expected to affect systems shipped early next year.
- Higher memory costs are a major factor.
- The final increase will vary by chip generation and memory configuration.
- Server manufacturers supplying Microsoft, Google and Oracle have reportedly informed customers of upcoming increases.
Our take
There is an interesting split emerging in AI economics. Access to frontier models through APIs is getting cheaper, while owning and operating the infrastructure required to run advanced AI can remain expensive.
That strengthens the case for New Zealand organisations to assess whether they genuinely need dedicated AI infrastructure. For many, cloud or API-based services may continue to offer better economics, while highly sensitive workloads may still justify the additional cost of private deployment.
Sources
KiwiGen.AI 