Intel has introduced the Xeon 6 CPU and the Gaudi 3 AI accelerator in a strategic response to NVIDIA and AMD’s grip on the AI chip market and high-performance computing data centers.
The announcement of these chips comes at a critical time for Intel. Once a chipmaking legend, Intel is now reportedly facing potential takeover offers from Qualcomm, a development that underscores its perceived strategic value in the semiconductor industry. This also highlights the challenges that Intel has been facing as it navigates a market increasingly dominated by its rivals.
Despite these uncertainties, Intel’s launch of the Gaudi 3 accelerator and Xeon 6 CPU marks a significant milestone for the company division, which faces intense competition not only from NVIDIA and AMD in the enterprise sector but also from AMD and Qualcomm in the consumer space.
“Demand for AI is leading to a massive transformation in the data center, and the industry is asking for choice in hardware, software, and developer tools,” said Justin Hotard, Intel executive vice president and general manager of the Data Center and Artificial Intelligence Group. “With our launch of Xeon 6 with P-cores and Gaudi 3 AI accelerators, Intel is enabling an open ecosystem that allows our customers to implement all their workloads with greater performance, efficiency, and security.”
Xeon 6: A major leap forward
The Intel Xeon 6 processor represents a significant upgrade over its predecessor, offering double the performance for AI and HPC workloads. This enhanced performance is achieved through an increased core count, doubled memory bandwidth, and AI acceleration capabilities embedded in each core.
These features make the Xeon 6 particularly well-suited to handle compute-intensive workloads with exceptional efficiency, addressing the growing demands of AI across the edge, data center, and cloud environments.
The introduction of Xeon 6 with performance cores (P-cores) marks a pivotal moment in Intel’s ongoing efforts to deliver robust AI systems.
With this new processor, businesses can expect improved performance per watt, which translates into a lower total cost of ownership (TCO). This makes the Xeon 6 an attractive option for organizations seeking to deploy high-performance AI workloads without incurring prohibitive costs.
The emphasis on energy efficiency and cost-effectiveness indicates Intel’s strategy to cater to the needs of businesses increasingly looking to optimize their AI operations while managing expenses. The Xeon 6, with its enhanced capabilities, is positioned as a critical component in this strategy, offering businesses a robust solution for their AI and HPC needs.
Gaudi 3 AI Accelerator: The generative AI powerhouse
The Intel Gaudi 3 AI accelerator is a cutting-edge chip designed explicitly for large-scale generative AI applications.
It features 64 tensor processor cores (TPCs) and eight matrix multiplication engines (MMEs), significantly accelerating deep neural network computations. With 128 gigabytes (GB) of HBM2e memory dedicated to training and inference tasks, the Gaudi 3 offers seamless compatibility with popular AI frameworks like PyTorch and advanced models from Hugging Face.
Designed to power applications such as natural language processing (NLP) and image generation, the Gaudi 3 is optimized for large-scale generative AI deployments.
The Gaudi 3’s ability to deliver up to 20% more throughput compared to NVIDIA’s H100 for inference of large models, like LLaMa 2 70B, makes it a formidable competitor in the generative AI space. Intel claims that the Gaudi 3 offers twice the price-to-performance ratio compared to the H100, making it an attractive option for businesses focused on cost optimization without sacrificing performance.
The Xeon 6 and Gaudi 3 are designed to complement each other, providing a comprehensive solution for customers looking to deploy a wide range of AI workloads.
Pricing
Earlier in the year, Intel revealed that an accelerator kit comprising eight Gaudi 3 processors on a baseboard would be priced at $125,000, which translates to approximately $15,625 per processor.
This pricing strategy positions Intel as a cost-effective alternative to NVIDIA, whose H100 card is over $30,000.
However, NVIDIA’s Blackwell-based B100/B200 GPUs are also on the way and are expected to offer significant performance advantages.
Intel’s position in the market
Intel’s presence in the data center market remains substantial, with approximately 73% of GPU-accelerated servers in the cloud using Intel Xeon as the host CPU.
Data centers running massive AI workloads on NVIDIA chips still primarily rely on Intel’s CPUs as their brains. The new Xeon 6, with its powerful P-cores and improved performance, is well-positioned to maintain this dominance.
Gaudi 3, on the other hand, directly competes with NVIDIA’s H100 and AMD’s MI300X as Intel’s dedicated generative AI processor. By focusing on generative AI applications, Intel targets a growing segment of the AI market that is expected to see significant expansion in the coming years.
Intel’s strategic approach with the Xeon 6 and Gaudi 3, combined with their seamless compatibility with popular AI frameworks, suggests a robust offering for businesses aiming to deploy high-performance AI workloads. The timing of these launches, amid growing competition and reported interest from Qualcomm, driven by its need to diversify beyond the smartphone segment, where growth has slowed, adds another layer of complexity to Intel’s position in the market.
Intel has also announced collaborations with OEMs like Dell Technologies and Supermicro to create custom-built systems optimized for AI workloads. These co-engineered solutions are designed to enhance AI performance and efficiency, leveraging the Open Platform Enterprise AI (OPEA) platform.
Challenges and opportunities ahead
Despite the promising capabilities of the Xeon 6 and Gaudi 3, Intel faces significant challenges as it strives to regain its footing in the semiconductor industry.
Intel’s CEO, Pat Gelsinger, is trying to restore the company to its former glory by pushing for more advanced chips for data centers and consumer PCs while expanding manufacturing capabilities.
The Xeon 6 and Gaudi 3 AI chips launch represents a critical moment for Intel as it seeks to navigate a complex and competitive landscape. With their advanced capabilities and cost-effective designs, these chips offer a promising solution for businesses looking to deploy high-performance AI workloads.
However, Intel’s ability to maintain its competitive edge will depend on its ability to continue innovating in the face of formidable competition from NVIDIA, AMD, and potentially Qualcomm.
The future remains uncertain, but with the Xeon 6 and Gaudi 3, Intel has made a strong statement of intent while offering a solid alternative.
