Intel Gaudi
Latest about Intel Gaudi

Intel begins sunsetting Ponte Vecchio to focus on Falcon Shores, Gaudi 2 and 3
By Aaron Klotz published
Intel has begun the retirement process on Ponte Vecchio, despite the fact it has not released its successor. Existing customers will still be able to purchase these GPUs, but new customers will not.

Commodore 64 can run AI to generate images — takes 20 minutes per 90 iterations to make 64 pixels
By Christopher Harper published
Nickbild on GitHub creates and documents a Commodore 64 AI sprite generator.

AMD increases Instinct MI1300 sales guidance to $4 billion — pales in comparison to Nvidia's $40B projection
By Anton Shilov published
AMD starts looking at sales of datacenter GPUs more optimistically, expects sales of Instinct MI300, other processors to top $4 billion in 2024.

Intel Core Ultra now powers more than 500 AI models, the company says
By Jeff Butts published
In a recent press release, Intel pointed out how successful its AI development initiatives have been not just for the company but for the technology field itself.

Intel expects $500 million in Gaudi 3 AI sales for the rest of the year
By Anton Shilov published
Intel does not expect its Gaudi 3 AI accelerator to make big money this year.

Engineer creates CPU from scratch in two weeks — begins work on GPUs
By Mark Tyson published
An engineer has shared his experience of designing a CPU from scratch, over two weeks, 'with no prior experience.' He has just started the arduous task of making a GPU from scratch.

Intel's Gaudi 3 is coming to China: But what about AMD's MI300 and Nvidia's B100/B200?
By Anton Shilov published
Intel preps special versions of Gaudi 3 for China, both in OAM and PCIe cards form-factors.

AI may eventually consume a quarter of America's power by 2030, warns Arm CEO
By Christopher Harper published
Amidst ever-increasing power grid demands and AI industry ambitions, head of Arm warns of "insatiable" demands.

Nvidia's H100 AI GPU shortages ease as lead times drop from up to four months to 8-12 weeks
By Aaron Klotz published
Lead times for Nvidia's H100 GPU have dropped enormously from almost a year down to just 12-8 weeks. This will help companies like OpenAI attain more H100 GPUs to train their own LLMs.
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.