Intel (INTC) on Tuesday announced a new AI chip designed to compete with one of Nvidia's (NVDA) most popular AI processors as part of the Intel Vision event.
The company's Gaudi 3 chip matches and exceeds Nvidia's H100 AI processor When it comes to training and deploying generative AI models. The H100 is one of Nvidia's most popular chips, powering artificial intelligence applications at the world's biggest technology companies, from Microsoft to Google.
According to Intel, Gaudi 3 is 40% more power efficient and has 50% faster inference than Nvidia's H100. Training refers to the time it takes to train an AI-based model, such as OpenAI's GPT-3. Inference, on the other hand, is the process of actually using the model in your app.
Intel says Gaudi 3 is up to 1.7x faster than H100 when training typical large-scale language models. Intel even says that its chips are on average 1.3 times faster than Nvidia's more powerful Nvidia H200 when it comes to inferring certain language models.
In addition to Gaudi series chips and Xeon CPUs, Intel also sells computers with onboard neural processing units that can handle AI tasks, so-called AI PCs.
These AI efforts come at a critical time for Intel.
The company is working to reinvent itself as a leader in sophisticated chip technology while increasing its chip manufacturing capabilities in the U.S. and abroad. This includes foundry operations that make Intel's own chips and chips for third-party customers. The company has already agreed to make chips for Microsoft.
But this reinvention has come at a high cost, with Intel announcing earlier this month that it would post an operating loss of $7 billion in 2023 as part of a new financial reporting structure. Intel's stock price has fallen 25% since the beginning of the year. Conversely, rivals Nvidia and AMD (AMD) are up 15% and 74%, respectively, so far this year.
Intel has been teasing its Gaudi 3 chip for some time, with CEO Pat Gelsinger showing off a version of the platform at a press event in December. This is seen as Intel's biggest chance to compete with Nvidia in its AI market, an area dominated by the graphics chip giant thanks to its powerful hardware and software.
The company's new chip also comes just a month after Nvidia debuted its own next-generation AI platform called Blackwell. The system includes Nvidia's B200 AI chip, which is made up of his two Blackwell chips combined to work as a single processor.
Intel is pushing its AI systems as an alternative to Nvidia products, insisting on an open source approach that allows customers to use the services and software they want.
And while Nvidia is the leader across AI technologies, Intel believes that enterprises remain wary of sticking to a single source for high-value needs like AI hardware. .
For example, hyperscalers such as Microsoft (MSFT), Google (GOOG, GOOGL), Amazon (AMZN), and Meta (META) already offer their own AI chips or are developing them in addition to using Nvidia's chips. going.
Intel's Gaudi 3 will give these hyperscalers another option in the chip market.
Intel also announced on Tuesday a new form of Ethernet connectivity for Gaudi 3 nodes. This competes with Nvidia's InfiniBand connectivity technology, which ensures uninterrupted connectivity and processing. The company also debuted its Xeon 6 processor for use in AI systems.
Taking a page out of Nvidia's book, Intel is offering the entire setup as a reference design that partners like Dell, HP, Lenovo, and Super Micro can use to build their own server cabinets.
Intel says customers will be able to order Gaudi 3-powered systems in a variety of sizes, from individual nodes and small boxes to huge megaclusters that can fill entire data centers.
Email Daniel Howley at dhowley@yahoofinance.com. Follow him on Twitter @Daniel Howley.
Click here for the latest technology business news, reviews, and helpful articles on technology and gadgets.
Read the latest financial and business news.ROM Yahoo Finance.