Nvidia is reportedly raising prices on some AI servers by more than 15% as soaring memory costs make the infrastructure behind artificial intelligence increasingly expensive.
The increase is expected to be absorbed by technology companies currently spending hundreds of billions of dollars building systems designed to eventually make nearly every industry more efficient.
Executives stressed that the higher prices should not be viewed as evidence that AI is becoming costly, but rather as evidence that society is moving rapidly toward a future in which costs are lower.
“We’re in the investment phase,” explained one industry analyst, referring to the several-year period during which companies purchase enormous quantities of scarce chips, memory, electricity, land, cooling systems and power-generation equipment in order to produce software that may someday reduce the number of people required to answer emails.
The price hikes arrive as demand for AI infrastructure continues to strain semiconductor and memory supply chains, pushing up the cost of the physical equipment needed to operate increasingly capable models.
Technology companies nevertheless remain confident that artificial intelligence will generate historic productivity gains once the world finishes constructing enough data centers to contain it.
Investors welcomed the news, noting that the rising cost of building AI infrastructure demonstrates how valuable AI infrastructure must be.
At press time, the industry remained on track to make everything cheaper immediately after everything required to make it becomes dramatically more expensive.
Inspired by:
