Intel

Intel Corporation is an American multinational technology company and the world's largest manufacturer of x86 central processing units, headquartered in Santa Clara, California. Founded in 1968, Intel is a major participant in the artificial intelligence hardware market through its Gaudi accelerator line, its Xeon server processors with built-in AI acceleration, and its push into AI-enabled personal computers, while also operating one of the world's largest semiconductor foundry businesses.^[1]

History

Intel was founded on July 18, 1968, by Robert Noyce and Gordon Moore, who left Fairchild Semiconductor to build a memory company that went on to create the microprocessor in 1971. Through the 1980s and 1990s, Intel's x86 architecture and "Intel Inside" marketing made it the dominant supplier of processors for personal computers and, later, servers. The company's 2005 decision to exit the discrete graphics market left it poorly positioned when GPUs became the primary hardware for AI training, and Intel spent the late 2010s and 2020s rebuilding its AI and foundry capabilities under successive chief executives.^[1][2]

In 2021, Intel announced its "IDM 2.0" strategy, including the creation of Intel Foundry Services to manufacture chips for external customers, and in 2024 the company began shipping its first dedicated AI accelerators at scale. The board appointed Lip-Bu Tan, a veteran semiconductor executive, as chief executive in March 2025 to lead a restructuring focused on manufacturing excellence and AI competitiveness.^[3]

Key Technology

Intel's AI hardware strategy spans several product lines. The Gaudi accelerators, developed by the former Habana Labs team Intel acquired in 2019, are AI training and inference processors positioned as cost-efficient alternatives to NVIDIA GPUs. Gaudi 3, launched in 2024, was built on a 5-nanometer process and aimed to deliver competitive performance for large language model workloads at a lower total cost of ownership than comparable GPUs.^[4][5]

The Xeon server CPU line incorporates built-in AI acceleration through the Advanced Matrix Extensions (AMX) instruction set, which accelerates matrix operations used in deep learning inference directly on the CPU, allowing many AI workloads to run without a discrete GPU. In the client market, Intel's Core Ultra processors integrate a neural processing unit (NPU) for on-device AI, forming the basis of the company's "AI PC" category that it projects to ship in the hundreds of millions of units.^[6]

Intel is also pursuing leading-edge manufacturing through its Intel 18A process technology, based on RibbonFET gate-all-around transistors and PowerVia backside power delivery, and operates a foundry business that manufactures chips for external customers including Amazon Web Services.^[7]

Applications and Impact

Gaudi accelerators have been deployed in enterprise AI clusters, including a deployment by IBM Cloud announced in 2025, and are used by organisations seeking AI compute with an alternative to the dominant GPU suppliers. Intel positions Gaudi for inference-heavy workloads and price-sensitive deployments, particularly in the enterprise and sovereign AI markets where procurement policies favour multiple vendors.^[4][5]

The AI PC category, driven by Core Ultra processors with NPUs, has made on-device AI features such as real-time language translation, background blur, and local small language model inference available to mainstream laptop buyers. Intel's foundry business, meanwhile, underpins its long-term claim to be a supplier to the broader AI semiconductor ecosystem rather than only a designer of its own chips.^[6][7]

>See Also

• AMD
• NVIDIA
• AI PC
• GPU Cluster
• National AI Office Malaysia

References