Nvidia Guides Fourth-Quarter Revenue to $130 Billion, Leaving China Out of the Outlook
- Nvidia's fiscal fourth-quarter revenue outlook excludes Data Center compute sales to China.
- Data Center revenue grew 110% year over year, supporting a comparatively modest earnings valuation.
- Falling gross margins and slower cash conversion complicate the case for the shares.
Nvidia (NVDA, $154.26, —) guided fiscal fourth-quarter revenue to $130 billion, compared with $68.13 billion a year earlier, without assuming China Data Center compute sales. That matters because its third-quarter results put the investment case on demand elsewhere, rather than a recovery in Chinese orders.
Nvidia sells computing chips, servers and networking equipment to cloud providers and businesses; Data Center generates most of its revenue.
Its quarterly financials showed GAAP net income of $65.29 billion and diluted EPS of $2.70, which means AI infrastructure demand already supports substantial earnings rather than speculative future profits.
Nvidia closed at $155.46 on 17 November. Does Data Center's $107.6 billion in AI computing and networking sales, its dominant business, up 110% year over year, justify a trailing P/E of 16.7x despite weakening margins?
That tension determines what the shares are worth.
Data Center expansion supports the earnings case
Data Center, which supplies AI computing and networking systems and accounts for the overwhelming majority of sales, generated $107.6 billion, up 110% year over year.
That scale supports the earnings multiple because Nvidia's AI opportunity is already producing revenue rather than depending on unproven future markets.
Total fiscal third-quarter revenue reached $115.8 billion, up 103% from $57.01 billion a year earlier. Operating expenses grew 57% year over year, substantially slower than sales, which explains why operating profit expanded faster than revenue.
Adjusted diluted EPS was $2.68, up 111% year over year. Its closeness to reported earnings strengthens the case because Nvidia's profitability does not depend on stripping out a large collection of expenses; its adjusted figures now include stock-based compensation.
Vera Rubin NVL72 systems have begun shipping, with racks running at partners including Microsoft Azure, Google Cloud and CoreWeave. Because these deployments put Nvidia's newest platform into operating infrastructure, they provide stronger evidence of commercial adoption than product announcements alone.
Nvidia is expanding across processors, networking and deployment designs, allowing it to capture more of each cloud infrastructure installation's spending. Vera CPUs and Spectrum-X networking broaden the opportunity beyond GPUs, although owners still need cloud providers to keep investing across that product family.
China's exclusion leaves margins and collections under scrutiny
Leaving China Data Center compute sales out of guidance makes the outlook less dependent on a reopening of that market. For shareholders, this means the projected expansion rests on Nvidia's business elsewhere rather than an assumed recovery in Chinese compute orders.
The fiscal fourth quarter also includes an additional week, which means its revenue outlook benefits from more selling time. Nvidia's cloud infrastructure demand may remain strong, but the headline comparison alone cannot establish how quickly customers' weekly purchases are growing.
Current GAAP gross margin was 74.8%, below the previous quarter's 75.0%. The direction deserves attention because even modest margin erosion across Nvidia's enormous sales base can absorb part of the benefit from higher shipments.
Cash conversion also restrains enthusiasm because Nvidia must collect payments to turn reported profits into spendable cash. The results filed with the SEC showed operating cash flow of $48.25 billion, while receivables absorbed $18.05 billion during the quarter, leaving cash generation behind accounting earnings.
Nvidia's peer discount outweighs the visible weaknesses
The peer discount looks excessive because Nvidia's Data Center expansion already produces substantial reported profits. Advanced Micro Devices trades at a trailing P/E of 109x and Broadcom at 43.6x; Advanced Micro Devices (AMD, $512.41, —) competes in AI accelerators, while Broadcom (AVGO, $350.25, —) supplies custom AI chips and networking silicon.
Nvidia's trailing price-to-sales of 10.4x means substantial profitability must persist to support its price. Compared with other semiconductor stocks, Nvidia's appeal therefore rests on converting cloud infrastructure spending into earnings, because sales growth alone would not protect the valuation if margins deteriorated persistently.
In our view, Nvidia is worth owning at today's price for AI-focused investors: Data Center's $107.6 billion of computing and networking sales, its dominant business, up 110% year over year, justifies a 16.7x trailing P/E despite margin pressure.
The view would change if deteriorating margins or collections undermined earnings.
This article is for informational purposes only and is not investment advice.