Intel has announced its development of the Xeon 6 processor. A so-called AI chip designed for data centres, the new semiconductor succeeds the Xeon 5 released a mere six months ago. Intel said that its next-generation AI chip would be more energy-efficient and powerful than its predecessor. The announcement of the Xeon 6 at Taiwan’s Computex tech conference comes amid competing product releases from AMD and Nvidia, the latter of which announced its ‘Rubin’ chip on Sunday. Intel chief executive Pat Gelsinger, however, emphasised his company’s versatility across the entire chip supply chain compared to its rivals.
“Intel is one of the only companies in the world innovating across the full spectrum of the AI market opportunity – from semiconductor manufacturing to PC, network, edge and data centre systems,” said Gelsinger. “Our latest Xeon, Gaudi and Core Ultra platforms, combined with the power of our hardware and software ecosystem, are delivering the flexible, secure, sustainable and cost-effective solutions our customers need to maximise the immense opportunities ahead.”
Intel AI chips designed for data centre market
Intel said that the Xeon 6 processor range would be divided into ‘E-core’ and ‘P-core’ variants, designed to address consumer needs regarding efficiency and performance respectively. The Intel Xeon 6 E-core, the firm continued, are available for purchase now, while the P-cores will be launched next quarter.
The chipmaker also revealed pricing for its Gaudi AI accelerators, stating that its standard AI kit of eight Gaudi 2 accelerators equipped with a universal baseboard came to $65,000 – “two-thirds the cost of comparable competitive platforms,” Intel claimed. The firm added that it was collaborating with global system providers including Gigabyte, Asus, Quanta and Wistron, among others, to bring its successor Gaudi 3 accelerator to market.
New details about Intel’s Lunar Lake architecture were also unveiled, with the firm stating that the AI chip’s architecture includes both E-cores and P-cores on its compute tile, in addition to a new graphics processing unit (GPU), image processing unit and neural processing unit. “Lunar Lake is set to power more than 80 different AI PC designs from 20 original equipment manufacturers (OEMs),” said Intel, with the firm adding that it expected to deploy 40 million Core Ultra processors this year.
US chipmaker’s future unclear
Intel is one of only three companies – the other two being TSMC and Samsung – capable of manufacturing the world’s most advanced semiconductors in its foundries. Recent years, however, have seen the firm cede market share not only to these rivals but also to chip design firms like AMD and Nvidia, all of which have surfed a wave of market enthusiasm for all things generative AI.
There is evidence to suggest, however, that Intel’s fortunes may recover thanks to its enthusiastic leap into the AI chips market and generous subsidies from the US government to expand semiconductor production. The relaunch of its foundry business in February – which, in 2023, made an operating loss of $7bn – and its adoption of ASML’s High NA scanners may also reap dividends.
Content from our partners
The company’s understanding of the foundry side of its business had steadily improved, Tech Insights analyst Dan Hutcheson told Wired at the time. “The messaging is much more focused, and they are picking up customers,” said Hutcheson, “which proves they are doing something right.”
