At an AMD investor event, industry giants Meta, OpenAI, and Microsoft unveiled their plans to utilize AMD’s upcoming Instinct MI300X AI chip, marking a pivotal shift away from Nvidia’s pricey GPUs that have long driven AI innovation.
Key Highlights:
-
AMD's MI300X Arrival: Meta, OpenAI, and Microsoft announced their adoption of AMD’s Instinct MI300X at an investor event, signaling a shift in the industry’s reliance on Nvidia's high-cost graphics processors for AI applications.
-
Potential Impact: The launch of MI300X, set for early release next year, holds the promise of reducing AI model development costs if it proves both high-performing and cost-effective, potentially challenging Nvidia's market dominance.
-
AMD's Vision: Lisa Su, AMD’s CEO, anticipates the AI chip market to exceed $400 billion by 2027 and aims for a significant market share, projecting a formidable presence in the evolving landscape.
Who's On Board with MI300X?
Notable players like Meta, Microsoft, and Oracle have committed to using the MI300X. Meta plans to utilize these GPUs for AI inference tasks, while Microsoft intends to offer access to MI300X chips via its Azure web service. OpenAI also pledges support for AMD GPUs in its Triton software for AI research.
AMD foresees modest initial sales, projecting around $2 billion in total data center GPU revenue by 2024, contrasting significantly with Nvidia's recent $14 billion data center sales in a single quarter. Nonetheless, AMD remains optimistic, eyeing a potentially booming $400 billion AI GPU market by 2027.
Lisa Su maintains a confident outlook, stating that AMD doesn't necessarily aim to outpace Nvidia but believes they can secure a significant portion of the anticipated market growth.
In the market, Advanced Micro Devices Inc experienced a slight decline despite announcements from Meta Platforms Inc and Microsoft Corp regarding their adoption of AMD's latest AI chip, challenging Nvidia's dominance.
MI300X vs. Nvidia's H100?
AMD's Instinct MI300X, slated for imminent release, is anticipated to be more cost-effective than Nvidia's offerings, with Lisa Su directly comparing its 192GB HBM3 memory to Nvidia's H100 at a recent investor event.
Impact on Nvidia's Sales:
The potential success of AMD's MI300X among leading tech companies could disrupt Nvidia's sales, particularly after Nvidia's warning of an anticipated hit due to U.S. restrictions on chip exports to China. AMD's enhancements to its ROCm software suite aim to counter Nvidia's advantages in the eyes of AI developers.
© Copyright 2023. All Rights Reserved Powered by Vygr Media.
