The AI Infrastructure Race: Nebius vs. Supermicro – And What It Means for Investors
The artificial intelligence boom isn’t just about chatbots and algorithms; it’s a massive infrastructure build-out. Two companies at the forefront of this expansion, Nebius (NBIS) and Super Micro Computer (SMCI), represent fundamentally different approaches to capitalizing on the AI revolution. Nebius focuses on providing the cloud-based foundation, while Supermicro builds the specialized hardware. But which offers the better investment opportunity?
Nebius: The Full-Stack Cloud Provider
Nebius is carving out a niche as a “full-stack” AI infrastructure provider. Unlike traditional cloud services, Nebius doesn’t just offer computing power; it integrates managed software services directly into its data centers. This allows them to cater to specific, demanding applications like data training, educational technology, and robotics. They currently operate a data center in Finland, lease capacity in the US, France, and Iceland, and are actively expanding with a new facility in New Jersey and a colocation deal in the UK. This strategy differentiates them from competitors like CoreWeave (CRWV), which primarily focuses on GPU processing tasks.
Supermicro: The AI Server Specialist
Supermicro has positioned itself as a leader in AI-optimized servers. While historically an underdog compared to giants like Hewlett Packard Enterprise (HPE) and Dell (DELL), Supermicro was first to market with liquid-cooled servers designed specifically for the intense demands of AI workloads. A crucial partnership with Nvidia (NVDA) ensures a steady supply of high-end GPUs – the engines driving AI applications. Currently, Supermicro controls approximately 9% of the dedicated AI server market, according to ABI Research, a significant foothold in a rapidly expanding sector.
Growth Trajectories: A Tale of Two Companies
Nebius is experiencing explosive growth. In 2024, revenue surged an impressive 462% to $118 million. However, this growth comes at a cost; adjusted EBITDA remains deeply negative at -$266 million. Analysts predict continued rapid revenue growth of 373% in 2025, reaching $556 million, with EBITDA improving to -$75 million. Nebius projects an annualized revenue run rate of $7 to $9 billion by the end of 2026, and analysts forecast $7.8 billion in revenue by 2027, representing a staggering 305% compound annual growth rate (CAGR) from 2024.
Supermicro, while still growing rapidly, is further along in its maturity. Fiscal 2025 revenue increased by 47% to $22.0 billion, with adjusted EBITDA growing 6% to $1.6 billion. Analysts anticipate a 65% revenue surge in fiscal 2026 to $36.3 billion, with EBITDA rising 10% to $1.8 billion. From fiscal 2025 to 2028, revenue and EBITDA are projected to grow at CAGRs of 33% and 26%, respectively.
Valuation: Where Does the Value Lie?
Valuation is a critical consideration. Nebius, with an enterprise value of $24.2 billion, trades at seven times next year’s sales. Supermicro, with an enterprise value of $16.6 billion, appears significantly undervalued at less than one times next year’s sales and six times its adjusted EBITDA. Nebius’s losses are likely suppressing its valuation, while Supermicro is still often perceived as a traditional server manufacturer rather than a high-growth AI play.
The difference in valuation reflects the perceived risk and growth potential. Nebius is a higher-risk, higher-reward play, dependent on successfully scaling its infrastructure and achieving profitability. Supermicro, while facing increasing competition, has a more established track record and a clearer path to sustained profitability.
Beyond Nebius and Supermicro: The Broader AI Infrastructure Landscape
The demand for AI infrastructure extends far beyond these two companies. The need for specialized chips is driving growth for Nvidia and AMD. Data center REITs like Equinix and Digital Realty Trust are benefiting from increased demand for data center space. Even established cloud providers like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud are heavily investing in AI-specific infrastructure.
The Future of AI Infrastructure: Key Trends to Watch
Several key trends will shape the future of AI infrastructure:
- Specialized Hardware: The demand for GPUs, TPUs, and other specialized AI chips will continue to grow.
- Liquid Cooling: As AI models become more complex, liquid cooling will become essential for managing heat and maximizing performance.
- Edge Computing: Processing data closer to the source (at the “edge”) will become increasingly important for applications like autonomous vehicles and industrial automation.
- Sustainable Infrastructure: Reducing the environmental impact of AI data centers will be a major focus.
- Software-Defined Infrastructure: Greater automation and orchestration of infrastructure resources will be crucial for managing complexity and optimizing performance.
FAQ: AI Infrastructure Investing
- Q: What is EBITDA? A: Earnings Before Interest, Taxes, Depreciation, and Amortization. It’s a measure of a company’s operating performance.
- Q: What is Enterprise Value? A: A measure of a company’s total value, including debt and cash.
- Q: Is AI infrastructure a good long-term investment? A: The AI revolution is still in its early stages, and the demand for AI infrastructure is expected to grow significantly in the coming years, making it a potentially attractive long-term investment.
- Q: What are the risks of investing in AI infrastructure? A: Risks include intense competition, rapid technological change, and potential economic downturns.
Ultimately, both Nebius and Supermicro offer compelling investment opportunities in the burgeoning AI infrastructure market. Supermicro’s established position, robust growth, and attractive valuation currently make it the more prudent choice. However, Nebius’s innovative approach and potential for explosive growth shouldn’t be ignored, particularly for investors with a higher risk tolerance.
Want to learn more about the AI revolution? Explore our other articles on AI investing strategies and the future of artificial intelligence.
Worth a look