Nvidia‘s (NASDAQ: NVDA) engendered breathtaking growth by pioneering high-end data center GPUs for AI and machine learning tasks. The AI realm, currently controlled by industry giants like OpenAI, Microsoft, Alphabet‘s Google, and Amazon, bows to the dominion of Nvidia’s GPUs.
Nvidia’s ascendancy as a forerunner in the AI market, undeterred by competitors, made it an investor’s Eden. Yet, Advanced Micro Devices (NASDAQ: AMD), the Goliath of gaming GPUs, now sets its sights on the AI realm. Can AMD unseat Nvidia from its celebrated throne as the AI leader?
Image source: Getty Images.
The Clash of Titans: AMD vs. Nvidia
AMD’s bold leap into the GPU arena, beginning with the acquisition of ATI for $5.6 billion in 2006, sent shockwaves across Nvidia’s empire. Leveraging ATI’s Radeon brand, AMD ventured into discrete GPUs for gaming PCs, culminating in APUs that bundled CPUs and GPUs on a single die. AMD’s strategy to undercut market leaders by offering more affordable chips tormented Nvidia, reminiscent of its tussle with Intel (NASDAQ: INTC) in the CPU arena.
While Nvidia held an iron grip over about 80% of the discrete GPU market last year, AMD lagged behind with a modest 17% share. Intel, returning to the discrete GPU scene in 2022, struggled to emulate Nvidia’s foresight, as it regrettably conceded market share in the PC CPU segment to AMD over the past decade. Nvidia, in a bid to keep AMD at bay, embraced cost-effective chips to lure budget-conscious gamers.
AMD’s Tactical Advancements in the AI Domain
Unlike Intel, reliant on its in-house foundries, AMD and Nvidia, with their fabless model, engage Taiwan Semiconductor Manufacturing (NYSE: TSM) for chip production. This reliance on TSMC imparts constraints on chip design, dictating size and power efficiency.
With the launch of its Instinct GPUs in 2017, AMD embarked on a relentless journey to match Nvidia’s prowess. The subsequent unveiling of 7-nanometer and 6nm MI chips, culminating in the latest MI300 Instinct series in 2023, positioned AMD as a worthy contender to Nvidia. In recent benchmarks, AMD’s flagship MI300X outperformed Nvidia’s H100 in raw processing power and memory bandwidth, an impressive feat in the AI landscape.
Moreover, AMD’s enticing proposition of four MI300 GPUs at the price of one H100, coupled with Nvidia’s chip shortages, could swing the tide in AMD’s favor, drawing data center operators towards the underdog.
AMD’s Strategic Footing in the AI Ecosystem
Industry behemoths like Microsoft, Meta Platforms, Oracle, Dell, and Hewlett Packard Enterprise are already eyeing AMD’s Instinct GPUs as viable alternatives to Nvidia’s offerings. AMD’s portfolio, spanning Epyc server CPUs, programmable chips from its Xilinx acquisition, and instilled with Instinct GPUs, promises cost-effective solutions for data centers.
During AMD’s fourth-quarter earnings call in January, CEO Lisa Su painted a rosy picture of the company’s future, highlighting the swift adoption and revenue growth of its MI300 deployments. AMD’s pronounced presence in the data-center and embedded chip markets, with a notable $1.2 billion revenue surge in 2023, positions the company as a formidable challenger in the AI realm.
Is AMD Poised to Usurp Nvidia?
While AMD’s trajectory in the AI domain seems promising, the path to dethroning Nvidia as the undisputed AI champion appears steep. Nvidia’s pricing equilibrium, coupled with evolving chip supply dynamics, may forestall AMD’s ascent to the apex of the AI throne.
Hence, AMD’s narrative in the AI landscape foresees a narrative akin to its conquest of the gaming GPU and data center GPU domains—leading a charge that carves a lucrative niche without toppling Nvidia’s reign.