Nvidia shares rise after Q4 revenue beats estimates

Nvidia’s stock rose about 1.3% in pre-market trading on February 26 after reporting fiscal Q4 revenue of US$68.1 billion, surpassing analyst estimates of US$66.2 billion from LSEG, with a 73% year-on-year increase.

The company’s data center unit, which accounted for 91% of sales, generated US$62.3 billion, above StreetAccount’s expectations of US$60.7 billion.

Nvidia provided an optimistic outlook for the fiscal first quarter, with revenue guidance of US$78 billion, plus or minus 2%, exceeding forecasts of US$72.6 billion.

Despite strong results, some analysts, including Richard Clode of Janus Henderson Investors and Dan Hanbury of Ninety One, remain cautious about the sustainability of AI infrastructure spending as hyperscalers face cash flow pressures.

Analysts note the market is focused on whether Nvidia can maintain its growth amid scrutiny of AI-related capital expenditures.

Markets have shown caution amid broader worries about AI investment trends.

🔗 Source: CNBC

🧠 Food for thought

Implications, context, and why it matters.

Beyond hyperscalers, new revenue streams are already taking shape

Some investors question whether hyperscalers will keep pouring money into AI infrastructure, yet Nvidia says the spend builds “AI factories” where token generation ties straight to revenue 1.
Revenue now comes from a wider mix of buyers since the top cloud providers together bring in a little over 50% of Nvidia’s data center revenue 1.
Sovereign AI, AI infrastructure purchased and run by national governments or state-backed entities, more than tripled year over year to over $30 billion in fiscal 2026 1.
Physical AI, AI that runs machines and robotics systems in the real world, added more than $6 billion of revenue in fiscal 2026 1.

A structural shift in computing economics is underway

Top hyperscalers have budgeted nearly $700 billion in capital expenditure this year, reshaping how computing gets funded 2.
CEO Jensen Huang says AI workloads need 1,000 times more computation than classical computing, which sets a higher long-term floor for spending 2.
Many companies now treat access to AI compute capacity as a condition for future revenue growth rather than a routine IT line item 1.
Nvidia projects its Rubin platform, its next-generation AI computing architecture, will cut inference token costs up to 10x versus Blackwell 1.