Alphabet Stock Could Soar as It Targets the $300 Billion AI Chip Market Dominated by Nvidia
Alphabet is positioning itself as a serious competitor to Nvidia in the $300 billion AI chip market by selling custom TPUs directly to customers. While Nvidia retains dominance due to its CUDA software ecosystem and flexibility, analysts project Alphabet's TPU business could generate over $100 billion in annual revenue by 2030. With current valuations considered cheap relative to growth prospects, investors may overlook this significant opportunity for Alphabet.
Nvidia (NVDA +2.56%) stock has advanced 1,300% since the artificial intelligence (AI) boom began in January 2023. The company's success, in terms of both financial results and share price appreciation, has been driven by its dominance in artificial intelligence accelerators, a market projected to top $300 billion this year.
Alphabet (GOOGL +1.11%) (GOOG +0.77%) has dabbled in AI accelerators for over a decade, but the company recently started selling custom silicon directly to customers, marking a more deliberate attempt to compete with Nvidia. Read on to learn more.
Image source: The Motley Fool.
Alphabet just positioned itself as a more serious threat to Nvidia
Nvidia invented the graphics processing unit (GPU) in 1999. Those chips, originally built to render realistic video game and computer graphics, have become the industry standard for accelerating complex data center workloads, such as artificial intelligence (AI), because they perform trillions of calculations per second.
Last year, Nvidia accounted for more than 80% of AI accelerator sales, according to Silicon Analysts. But its market share is likely to fall in the years ahead as more companies turn to custom silicon solutions. Three of its largest customers -- Amazon, Microsoft, and Alphabet -- have deployed chips purpose-built for AI.
Alphabet was the first to pursue a custom silicon strategy, and it remains the most significant threat to Nvidia. In 2016, Alphabet deployed its first tensor processing unit (TPU), a chip designed specifically for the matrix and vector-based math needed to build and run AI models.
Initially, Alphabet limited TPUs to internal use cases, but the company made its custom silicon accessible to Google Cloud customers in early 2018. In the second quarter, Alphabet began selling TPUs directly to clients for use in external data centers, positioning itself as a more direct threat to Nvidia.
Nvidia is unlikely to lose its market leadership in AI accelerators
Nvidia is unlikely to cede its dominance in the AI accelerator market for two reasons. First, the company has a significant competitive advantage in CUDA, a software platform comprising hundreds of code libraries and frameworks (building blocks) that help programmers write GPU-accelerated applications.
CUDA is the main reason Nvidia GPUs have become the dominant AI accelerators, and the proprietary nature of the platform constitutes a durable economic moat. "Once a team has built pipelines on CUDA, switching to another platform is prohibitively expensive," explains VentureBeat.
Second, TPUs run far fewer algorithms than GPUs because they are built for specific tasks. "While TPUs excel at specific deep learning workloads, they are far less flexible," according to VentureBeat. That means Nvidia GPUs could immediately run a new AI technology if one were invented tomorrow, but the same is not true of Alphabet TPUs.
NASDAQ: GOOGL
Key Data Points
Alphabet TPUs could become a $100+ billion revenue stream by 2030
Gil Luria, head of technology research at D.A. Davidson, sees custom silicon as a massive growth opportunity for Alphabet. Anthropic and Meta Platforms have agreed to spend billions of dollars on TPUs in the years ahead, and Alphabet recently announced a joint venture with Blackstone to build a TPU cloud business.
Looking ahead, Lura says Alphabet could eventually capture 20% of the AI infrastructure market, which would value its chips business somewhere around $900 billion. Meanwhile, Morgan Stanley analysts expect custom silicon (primarily Alphabet's TPUs) to account for 24% of AI accelerator sales in 2030, up from 15% today.
Those estimates suggest Nvidia will retain its dominance, but they also underscore the massive opportunity that sits before Alphabet. AI accelerator spending is projected to reach $600 billion in 2030, which means Alphabet could bring in more than $100 billion in revenue annually from TPU sales alone by the end of the decade.
I think investors are overlooking this opportunity. Alphabet trades at 19 times earnings, well below the three-year average of 25 times earnings. That valuation is cheap for a company whose earnings are projected to grow at 14% annually over the next three years. That's why Alphabet shares could soar as the company targets the AI accelerator market.
