Nvidia is aggressively utilizing its balance sheet to shape the future of artificial intelligence. Through major Nvidia AI investments—including a $2 billion stake in neocloud provider Nebius and significant financial backing for hardware startups—the $4.5 trillion chipmaker is expanding its role from a traditional hardware supplier to a foundational ecosystem financier. Despite these bold moves, Nvidia faces fresh tests as major clients like Meta develop in-house silicon, shifting the competitive landscape just days before the highly anticipated GTC conference.
Expanding Cloud Capacity with Nebius
Nvidia’s $2 billion investment secures an 8.3% stake in the Amsterdam-based cloud firm Nebius, with shares purchased at $94.94 each. This strategic partnership aims to rapidly expand global artificial intelligence data center capacity. Nebius has set an ambitious goal to construct more than five gigawatts of data-center infrastructure by the end of 2030, which is roughly enough energy to power 3.8 million homes.
As part of this massive infrastructure push, Nebius recently received city council approval to build a 1.2-gigawatt AI facility on 400 acres near Independence, Missouri. Power delivery for this site is anticipated in the second half of 2026. In return for the financial backing, Nebius will receive early access to Nvidia’s most advanced technologies, including its next-generation platforms, CPUs, and BlueField storage solutions. Commenting on the alliance, Nvidia Chief Executive Jensen Huang noted that Nebius is actively building an AI cloud specifically designed for the “agentic era.”
Backing Emerging AI Innovators
Beyond its cloud partnerships, Nvidia is placing heavy financial bets on emerging artificial intelligence startups. The company recently formed a multiyear strategic collaboration with Thinking Machines Lab, an AI startup established last year by former OpenAI executive Mira Murati. Thinking Machines aims to create artificial intelligence systems that are more comprehensible, customizable, and generally capable.
Under the agreement, Thinking Machines Lab has committed to deploying at least one gigawatt of Nvidia’s forthcoming Vera Rubin AI accelerators early next year. Reports indicate this arrangement could represent roughly $50 billion of computing power. Huang expressed enthusiasm for the deal, stating that the startup has assembled a top-tier team to push the boundaries of artificial intelligence.
Nvidia is also funneling billions into the hardware supply chain to resolve critical technical bottlenecks. Because specialized indium phosphide laser chips are not currently manufactured in sufficient volumes for AI data centers, Nvidia recently invested $2 billion each in laser manufacturers Lumentum and Coherent. Simultaneously, French startup Scintil Photonics, which is backed by Nvidia, began providing its own laser chips to customers for testing this week. Working with Israel-based Tower Semiconductor, Scintil uses co-packaged optics to transfer data via light pulses rather than electrical signals. Scintil CEO Matt Crowley confirmed the company is currently in talks with up to seven companies interested in deploying the technology by 2028.
The Meta Challenge and Custom Silicon
While Nvidia solidifies its underlying hardware ecosystem, its largest clients are increasingly diversifying their supply chains. Meta announced that it is developing four generations of its in-house MTIA chips over the next two years. The social media giant is steering many of these custom chips toward inference—the operational stage where artificial intelligence models actively answer user requests.
Meta engineering executive Yee Jiun Song emphasized this shift, stating that inference demand is currently exploding. Although Meta will continue to purchase hardware from Nvidia and AMD, market analysts warn that the competitive picture is shifting. Major cloud groups and hyperscalers are spreading orders across multiple vendors to secure supply and reduce reliance on a single provider. Analysts note that Meta is actively locking in supply as it diversifies its hardware sources.
Financial Outlook and the GTC Conference
Despite the evolving competitive landscape, Nvidia’s operating numbers remain highly robust. The company recently reported record quarterly revenue of $68.1 billion, largely driven by $62.3 billion from its data-center products. Looking forward, Nvidia forecasts its current-quarter revenue to reach approximately $78 billion.
Following the Nebius announcement, Nvidia stock held steady near its $186.03 close. However, market experts caution that broader industry dynamics are changing. Rising custom silicon investments mean Nvidia may face a tougher fight for the next wave of corporate IT budgets. Furthermore, competitors are signaling strong growth; Broadcom recently stated it has line of sight to more than $100 billion in AI chip sales in 2027.
Investors are now turning their attention to Nvidia’s upcoming GTC conference, which opens on March 16 in San Jose. Billed around AI factories, inference computing, and next-generation infrastructure, the event will serve as a critical checkpoint for Nvidia to address market pressures, validate its recent investments, and showcase its hardware roadmap to the world.
