From $24 Billion Orders to $150 Billion Spend: Nvidia's Two-Front Push to Dominate AI

The clearest signal yet that Nvidia’s AI juggernaut is accelerating comes not from Santa Clara but from the server hubs of Round Rock. Dell’s blockbuster earnings have laid bare the concrete demand for Nvidia-powered infrastructure, while Jensen Huang simultaneously maps out a spending spree on Taiwan that will run into the hundreds of billions. The picture is one of a company operating at two speeds: harvesting immediate orders today and planting seeds for tomorrow’s manufacturing capacity.

Dell booked a record $43.8 billion in first-quarter revenue, up 88% year over year, and its backlog of AI-related orders swelled to $24.4 billion. The server maker now expects to sell roughly $60 billion in AI-optimized servers for the full year, a 144% surge. That is a far more tangible proof point than any industry forecast. Dell also raised its total revenue guidance to between $165 billion and $169 billion, up sharply from the prior band of $138 billion to $142 billion. The sheer magnitude of the revision underscores how swiftly enterprises and hyperscalers are committing capital to AI clusters.

For Nvidia, the numbers validate a thesis already confirmed by its own books. In its fiscal first quarter of 2027, Nvidia posted revenue of $81.6 billion, an 85% increase, with the data center segment alone contributing $75.2 billion, up 92%. Dell’s infrastructure unit — the direct conduit for Nvidia’s chips — reported a record $29 billion in revenue, a 181% jump. Within that, AI-optimized servers accounted for $16.1 billion, while traditional servers and networking added $8.5 billion. The two reports together form a seamless demand narrative: the server cycle is real and accelerating.

Yet even as Nvidia collects the profits from this cycle, it is already attacking the next bottleneck. Co-packaged optics — optical connections placed directly next to GPUs — are critical for linking thousands of processors in giant AI clusters. According to a Rosenblatt Securities research note, Nvidia has demanded a 20-fold increase in capacity for indium-phosphide lasers from its suppliers by 2030. The suppliers themselves consider a 12-fold increase more realistic. The gap illustrates how optical networking has become the pinch point for scaling the next wave of AI.

Should investors sell immediately? Or is it worth buying Nvidia?

That is why Taiwan is getting the lion’s share of Nvidia’s capital. Speaking at the TD Cowen conference on May 28, Huang disclosed that Nvidia will spend $100 billion to $150 billion annually with its manufacturing partners on the island. The funds will churn out Blackwell Ultra chips and the upcoming Vera Rubin platform, with first Vera Rubin systems still on track for delivery in 2026. The investment underlines a strategy of locking in supply for generations to come, even as near-term demand soaks up every available chip.

On the financial side, Nvidia’s margins remain the envy of the semiconductor world. GAAP gross margin came in at 74.9% last quarter, with the non-GAAP measure at 75%. Analysts flagged rising memory prices as a potential headwind, though the high-margin software ecosystem and CPU business offer some insulation. For the current quarter — the second of fiscal 2027 — Nvidia expects revenue around $91 billion, propelled by both hyperscale customers and a growing base of smaller enterprise clients.

The stock has been treading water after its long rally. At €184.64 on Friday, Nvidia shares are up 14.61% year to date and 51.34% over 12 months, but they sit about 8% below the mid-May all-time high. The 14-day relative strength index of 36.4 hints at a mildly oversold condition following months of consolidation between roughly $210 and $230 (in dollar terms). With a market capitalisation of roughly $5.23 trillion and a price-to-earnings ratio of 32.6 based on earnings per share of $6.57, the valuation remains steep, but operational results continue to justify the premium.

Nvidia at a turning point? This analysis reveals what investors need to know now.

The next catalyst arrives on June 1, when Huang takes the stage for the GTC keynote in Taipei. Investors expect detailed specifications for the Vera Rubin architecture, and possibly new products targeting physical AI and robotics. Meanwhile, Nvidia is preparing to restructure its data center reporting into hyperscale and ACIE segments — separating AI clouds from industrial and enterprise customers. The overhaul will give the market a clearer view of where growth is coming from. For now, every external data point, from Dell’s order book to Taiwan’s production lines, points in the same direction: the AI buildout is only just beginning.

Ad

Nvidia Stock: New Analysis - 29 May

Fresh Nvidia information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.

Read our updated Nvidia analysis...

en | US67066G1040 | FROM | boerse | 69441091 |