In the fervently competitive environment of artificial intelligence (AI) chips, AMD (Advanced Micro Devices) finds itself in a familiar second place to Nvidia, the current market leader. AMD’s well-accustomed to being the underdog, having contended with industry behemoth Intel for years in the PC and server chips market, while still managing healthy stock performance. With AI’s rapid expansion, another tech duel is set for AMD and Nvidia, matching the reigning AI chips leader against a proven challenger.
Facing off to Nvidia, AMD plans to launch its MI300 AI chip as Nvidia’s major AI event unfolds in San Jose, California on March 18. Nvidia plans to reveal its latest product, the B100 GPU, its first of the Blackwell series, at the GTC 2024 conference, a seminal AI developers’ event.
AMD’s tenacity as a credible competitor is highlighted by its enduring rivalry with Intel. Nearing bankruptcy a decade ago, AMD bounced back, reclaiming market share in PCs and servers even with Intel’s continued dominance. Intel led the pack with 72.4% of total PC chip shipments in Q4 2023, followed by AMD’s 18.3%. Intel further commanded the server chip market with 68.7% over AMD’s 20.5%.
However, manufacturers’ aversion to sole vendor lock-in has always given AMD an opening. As AI stocks rose sharply, AMD’s stocks dipped 4% to 187.06 on profit-taking, while Nvidia’s retreated 3.2% to 879.44.
Experts consider AMD’s Instinct MI300 series a compelling substitute for Nvidia’s H100 GPU, while cloud service providers and companies appreciate having a solid secondary choice to maintain competitive pricing and stimulate innovation. AI customers, often hyperscalers, prefer multiple suppliers to prevent dependency on a single provider. Nvidia’s once biennial product introduction strategy has been adjusted to an annual one in response to AMD’s recent acceleration into the AI accelerator market.
Insiders predict Nvidia will announce the specifics of its upcoming B100 product at GTC, aiming for a launch later this year, with plans to supersede the B100 with its X100 GPU in 2025. Other players in the industry include Intel with its Gaudi AI processors and Marvell Technology and Broadcom’s custom application-specific integrated circuits, joined by AI chip startups like SambaNova Systems and Cerebras.
Even with Nvidia controlling over 90% of the GPU market due to its comprehensive solution for AI computing, AMD and Intel remain committed contenders, banking on their GPUs’ sheer computational power. Piper Sandler analyst Harsh Kumar views AMD as a competitive Nvidia substitute, explaining that AMD offers both value and performance without necessarily providing predefined libraries or the full stack of software, yet they’re dedicated to delivering the best performance.
Given the AI market’s enormous potential, Wall Street analysts see no sign of slowing investment in data center infrastructure to bolster generative AI applications. Despite increasing competitors in this segment, Nvidia’s integrated AI solutions continue to propel its significant market lead.
In conclusion, in the escalating AMD vs. Nvidia battle in AI chips, both AMD and Nvidia have been valuable bets for their shareholders. While AMD’s rise doesn’t match Nvidia’s exceptional twentyfold upswing in the past five years, both are indeed Wall Street stars, outperforming 99% of all stocks in terms of overall fundamental and technical strength.
