Amazon Web Services, which is used extensively by retailers across Australia, notably by security solutions provider Swann, has confirmed that its new systems running Nvidia’s Blackwell chips won’t likely be online until early next year.

That schedule is later than what was earlier predicted. Nvidia introduced the new generation of processors in March, and said that partners like Amazon would have its Blackwell-based products ready to use by the end of 2024.

Nvidia’s timelines have “shifted,” AWS Chief Executive Officer Matt told Bloomberg Television. “Right now, we’re working with them and we have early samples, but as they get their yields up from the respin of the Blackwell chip, it’ll probably be early next year before we get production samples in real volume.”

AWS, the world’s biggest cloud-computing provider, will be the first to deploy computers based on the new chip in significant quantities, said Garman. Dell Technologies has also said that its Blackwell-based servers will be generally available at the beginning of 2025.

Nvidia products are becoming highly sought after by data centres which are using the chips to develop AI software and services.

In August, the company acknowledged that Blackwell proved more difficult to produce than anticipated. It said that it was making changes to increase its manufacturing yield.

As ChannelNews reported reported at the time, Nvidia reportedly informed Microsoft and at least one other cloud provider that production of its Blackwell B200 AI chip may be pushed by at least three months. The delay was attributed to design flaws discovered “unusually late in the production process.” Nvidia was reported to be working through a fresh set of test runs with chip producer Taiwan Semiconductor Manufacturing Company.

Nvidia Blackwell Architecture (Image: Sourced from Nvidia Newsroom)

Nvidia Blackwell Architecture (Image: Sourced from Nvidia Newsroom)

Nvidia added that it expects to bring in “several billion dollars” of revenue in the fourth quarter from the product.

The company’s CEO Jensen Huang said that supplies will be plentiful after manufacturing gains momentum. “We’re going to have lots and lots of supply, and we will be able to ramp,” he said.

About 40 per cent of Nvidia’s revenue stems from large data-center operators including companies such as Google and Meta — which are pouring tens of billions of dollars into AI infrastructure.