In the current era of the silicon gold rush, the spotlight has largely remained on the chipmakers—the Nvidias and AMDs of the world that design the essential hardware. However, a new front is opening in the battle for artificial intelligence supremacy: the infrastructure layer. As traditional cloud giants like Amazon Web Services (AWS) and Microsoft Azure grapple with the sheer scale and specialized requirements of large language model (LLM) training, a new breed of ‘AI-native’ cloud providers is stepping into the breach.
Two names have dominated the conversation among institutional investors and tech enthusiasts alike: CoreWeave and Nebius Group. While both companies aim to provide the massive GPU clusters necessary to power the next generation of AI, their paths to market and investment profiles couldn’t be more different. One is a venture-backed darling with deep ties to the industry’s most powerful hardware supplier, while the other is a rebranded public entity rising from the complex restructuring of a former tech titan.
Key Takeaways
- Specialized Infrastructure: Both companies offer ‘GPU-as-a-Service,’ providing more agility and specialized performance than general-purpose cloud providers.
- CoreWeave’s Dominance: Backed by Nvidia, CoreWeave has secured billions in funding and is widely considered the leading private contender for an upcoming IPO.
- Nebius’s Pivot: Formerly part of the Russian tech giant Yandex, Nebius Group has emerged as a purely international, AI-focused infrastructure play based in Europe.
- Investment Accessibility: Nebius is currently traded publicly (NBIS), whereas CoreWeave remains a high-valuation private entity accessible mostly to institutional players and late-stage secondary markets.
CoreWeave: The Nvidia-Backed Powerhouse
CoreWeave has quickly ascended from its humble beginnings as a cryptocurrency mining firm to become one of the most valuable private companies in the AI space. Its secret sauce? A symbiotic relationship with Nvidia. As an ‘Elite’ partner in the Nvidia Partner Network, CoreWeave often receives priority access to the latest H100 and B200 Blackwell chips—hardware that is currently the most sought-after commodity in the global economy.
The company’s growth is fueled by a staggering amount of capital. In 2024 alone, CoreWeave has secured billions in debt and equity, valuing the company at roughly $19 billion. For investors, the ‘upside’ here is tied to the company’s eventual IPO. CoreWeave isn’t just renting out servers; it is building a bespoke environment optimized for low-latency, high-throughput AI training that general clouds often struggle to replicate. However, the risk lies in its concentration: its fate is inextricably linked to Nvidia’s supply chain and the continued appetite for massive LLM training runs.
Nebius Group: The European Challenger
If CoreWeave is the well-funded incumbent of the specialized cloud, Nebius Group is the seasoned underdog with a point to prove. Following the divestment of its Russian business, the Nasdaq-listed entity formerly known as Yandex N.V. rebranded as Nebius Group. Headquartered in Amsterdam, the company has pivoted entirely to AI infrastructure, leveraging its existing technical talent and a significant cash pile from its corporate split.
Nebius’s advantage lies in its operational maturity. Unlike many startups, it already has a history of managing large-scale data centers and complex software stacks. By positioning itself as the ‘European champion’ for AI compute, Nebius offers a compelling alternative for companies concerned with data sovereignty and those looking for a publicly traded entry point into the GPU-as-a-Service market. With plans to expand its GPU capacity significantly by the end of 2025, Nebius is trading at a valuation that many analysts believe doesn’t yet fully reflect its pivot into the high-margin AI sector.
Which Stock Has More Upside?
Determining the winner depends on your investment horizon and risk tolerance. CoreWeave likely possesses the higher ‘ceiling’ due to its massive scale and direct pipeline to Nvidia’s innovations. If the company goes public in 2025 at a significant premium, early investors could see exponential returns. However, since it is currently private, most retail investors are locked out of this growth phase.
Nebius Group, conversely, offers immediate liquidity and a ‘turnaround’ narrative. Because it is still shaking off the complexities of its Russian divestiture, the stock may be undervalued relative to its compute assets and cash on hand. For investors looking for a ‘pure play’ in AI cloud that is already trading on the Nasdaq, Nebius presents a unique, albeit volatile, opportunity.
Practical Advice for AI Investors
When navigating the specialized cloud sector, keep these strategies in mind:
- Watch Capital Expenditure (CapEx): In this business, growth is tied to how many GPUs you can buy and house. Monitor quarterly reports for updates on data center expansions and hardware procurement.
- Monitor the ‘Secondary’ Market: For those interested in CoreWeave, platforms like Hiive or Forge Global can provide insights into where private shares are trading, even if you aren’t an accredited investor.
- Diversify Beyond the Chips: Don’t just own Nvidia. The infrastructure layer (like Nebius) and the networking layer (like Arista Networks) provide a hedge against a potential slowdown in chip sales.
Frequently Asked Questions
When is the CoreWeave IPO expected?
While no official date has been set, industry rumors and the company’s recent hiring of financial executives suggest an IPO could happen as early as the first or second quarter of 2025, depending on market conditions.
Is Nebius Group still associated with Russia?
No. Following a comprehensive restructuring approved by international regulators, Nebius Group sold its Russian assets and is now an entirely Western-focused entity headquartered in the Netherlands and listed on the Nasdaq.
Can these companies compete with AWS or Google Cloud?
They don’t necessarily need to compete on everything. While AWS offers a thousand different services, CoreWeave and Nebius focus exclusively on high-performance computing. This specialization often allows them to offer better pricing and more efficient configurations for AI researchers and developers.