NVIDIA is seeing explosive growth, with revenue doubling year-over-year, driven by surging demand for its graphics processing units (GPUs) that power cutting-edge generative AI systems.
In its latest quarterly earnings report, NVIDIA revealed revenue of $13.5 billion, up 101% from the prior year. Profit skyrocketed over 800% to $6.2 billion. This staggering growth far exceeded Wall Street expectations. Additionally, the company announced a $25 billion stock buyback program, sending its shares soaring over 10% in after-hours trading.
Behind these numbers is NVIDIA's data center business, which includes sales of GPUs for AI research and applications. This segment grew an astonishing 171% year-over-year to contribute over $10 billion in revenue, making up the majority of NVIDIA's sales. The company's AI chips are not just gaining traction in the United States; they have also captured a significant market share in China, despite export restrictions, accounting for 20% to 25% of overall sales.
"Every single data center will be accelerated," said NVIDIA CEO Jensen Huang, referring to the shift to add AI capabilities across computing infrastructure.
The main driver is demand for NVIDIA's newest H100 GPU, tailored for large-scale AI model training. Major cloud platforms, startups, and enterprises are clamoring for access to the chips to power the latest AI innovations, like ChatGPT conversational models.
However, extreme scarcity persists. With supply chain issues and production constraints, NVIDIA is struggling to meet the voracious appetite for its AI silicon. This shortage may last well into 2024. Still, NVIDIA said supply will "substantially increase" in the coming quarters.
Rival chipmakers Intel and AMD lag far behind NVIDIA in generative AI. But they aim to make up ground, promising more focus on AI-enhanced chips. This mounting competition may eventually erode NVIDIA's dominance.
While NVIDIA has carved out an early lead and is riding high now, there are looming challenges. Major tech firms like Amazon, Google, and Microsoft, all of whom are NVIDIA's current customers, are in the process of developing their own AI chips. Additionally, the industry’s potential shift from AI training to inference could offer a window for NVIDIA's competitors like Amazon Web Services, which creates its own chips specifically for AI inference.
Moreover, the availability of high-quality open-source AI models could lessen the demand for Nvidia's specialized training chips, although the company has recently launched its GH200 chips designed for inference.
For now, NVIDIA is riding an enormous wave, with its next quarter's revenue projected to nearly triple year-over-year, exceeding $16 billion. Fueled by the ascendance of AI, NVIDIA seems poised for unstoppable growth. But it remains to be seen whether its rivals or supply limitations will disrupt this trajectory.