Intel Signs “Multi-Billion-Dollar Framework” To Make Amazon AI Chips With Latest Technology

Ramish Zafar
Intel CEO Pat Gelsinger displays a test chip package built with Universal Chiplet Interconnect Express (UCIe). The UCIe standard will allow chiplets from different vendors to work together, enabling new designs for the expansion of diverse AI workloads. Gelsinger displayed the chip package on Tuesday, Sept. 19, 2023, at Intel Innovation in San Jose, California. (Credit: Intel Corporation)

This is not investment advice. The author has no position in any of the stocks mentioned. Wccftech.com has a disclosure and ethics policy.

Intel's shares jumped by 8% in after market trading to provide the stock with some much needed respite after enterprise computing behemoth Amazon AWS announced that it is teaming up with the firm to develop chips on the 18A process node. The deal should be part of Intel's fledgling Intel Foundry Services business, which has seen CEO Pat Gelsinger go all in as part of his bid to rescue the beleaguered chip manufacturing company.

Amazon's announcement shared that the deal has received the blessing of officials in Ohio, where the firm plans to invest $7.8 billion in data center operations. The deal will also see the pair 'explore' future chip manufacturing technologies such as Intel's 18AP and the 14A process.

Related Story TSMC Accelerates CoPoS Packaging to Replace CoWoS, as Glass Core Substrates Cut Costs 30% and Boost Wafer Utilization Past 90%

Amazon's Intel Deal Is The Latest In Big Tech's Bid To Develop In House AI Chips

Intel and Amazon's press release did not share the financial specifics of the new deal. Instead, they outlined a "co-investment" that is part of a "multi-year, multi-billion-dollar framework covering product and wafers from Intel" which saw the chip manufacturer tout its previous partnerships with Amazon to bring the latter's cloud computing instances on the Intel platform.

Amazon's Amazon Web Services (AWS), which led today's announcement, is the high margin, recurring revenue division of the business that originally started as an eCommerce firm. It, like other technology giants, develops custom chips to power up in house software, but these have mostly been  based on the Arm architecture.

Under the new deal, Intel will first use its Intel 3 manufacturing process to build a custom Xeon 6 chip for Amazon workloads. Previous Xeon Scalable processors have seen Amazon run the EC2 instances, which Amazon has claimed has allowed up to 20% higher performance than predecessor instances. In cloud computing, an Instance is a virtualization of a physical computer, and faster performance enables users to improve costs and computing.

A 2022 presentation of Amazon's Graviton processor. Image: AWS

Following the custom Xeon chip, Intel will also produce an "AI fabric chip" for Amazon's AWS on the 18A manufacturing process. 18A is Intel's all out effort to retake the chip manufacturing crown, and the firm is also working with Broadcom to use the technology. The process is slated to enter production next year, at par with the Taiwan Semiconductor Manufacturing Company's (TSMC) 2 nanometer node.

It is likely that the AI fabric chip will be based on ARM's chip designs and will be a successor to Amazon's current custom processors, but this remains speculation unless Intel and Amazon share additional details.

Intel CEO Patrick Gelsinger has repeatedly stressed that his firm is the only one with a diverse semiconductor portfolio capable of targeting all AI requirements. During Intel's first quarter of 2024 earnings call, Gelsinger shared that his firm has a "clear line-of-sight to becoming the largest system foundry for the AI era and the second-largest overall by 2030," and the Amazon deal might help achieve the objective.

Amazon, along with Google and Meta, is developing its in house chips for AI workloads. Google's tensor processors are already used by Apple to train its AI software, and Meta shared details of its training and inference accelerator for AI workloads in April. These developments come as NVIDIA remains the undisputed leader of AI processors, and firms rush to lay their hands on its latest chips that are in short supply.

Ramish Zafar Photo

About the author: Ramish is a seasoned technology writer and editor with more than a decade of experience. He specializes in semiconductor fabrication and market analysis. With a background in finance and supply chain management - via his bachelors in Finance and a micromasters in supply chain management from MIT - Ramish combines financial rigor with deep industry insight to deliver accurate and authoritative coverage.

Follow Wccftech on Google to get more of our news coverage in your feeds.

Button