Loading the Elevenlabs Text to Speech AudioNative Player...
Nvidia NVDA once again reported outstanding results for its fiscal third quarter and provided investors with a fiscal fourth-quarter forecast that exceeded our expectations. We raise our fair value estimate to $130 per share from $105, as we’re more optimistic about Nvidia’s growth over the next two calendar years, as the supply of the firm’s products is improving faster than we expected. We’re also pleased with management’s commentary around its Blackwell graphics processor’s gross margins once the products are fully ramped, which boosts our confidence that gross margins can remain in the mid-70% range in the long run.
Shares appear modestly overvalued to us, as growth will inevitably slow in the long run, and we still think Nvidia’s largest customers have plenty of incentive to reduce their reliance on Nvidia over time—whether it be with in-house chips or more efficient capital expenditure. We maintain our Very High Morningstar Uncertainty Rating for wide-moat Nvidia, as the artificial intelligence landscape is both secretive and rapidly evolving, while Nvidia’s revenue carries high operating margins so that any upside (or downside) in revenue has an outsize impact on cash flow.
Revenue in the October quarter was $35.1 billion, up 17% sequentially, up 94% year over year, and well above guidance of $32.5 billion and FactSet’s consensus estimate of $33.2 billion. This represents the sixth straight quarter of Nvidia’s revenue exceeding its quarterly guidance by $2 billion or more, although the negative stock market reaction after hours suggests that some investors were seeking an even greater beat. Revenue in the January 2025 quarter is forecast to be $37.5 billion, which would represent only 7% sequential growth. We’re unconcerned about this relatively lackluster forecast and instead expect to see Nvidia sell all the AI products it can build, whether they be current-generation Hopper or next-generation Blackwell.
Data Center Revenue Forecasts
For the past six quarters, Nvidia has increased its data center revenue by roughly $4 billion per quarter. We think this incremental revenue increase comes from Nvidia’s supply chain partners expanding capacity so that Nvidia can meet the insatiable demand for its AI data center products. We anticipate that Nvidia’s guidance for the January quarter will once again prove to be conservative, and we estimate that Nvidia will earn $4 billion more in incremental revenue for the seventh straight quarter, reaching nearly $35 billion in data center revenue next quarter. For perspective, this compares with the $15 billion Nvidia earned in its entire fiscal 2023 just two years ago before the AI boom.
Our near-term revenue estimates continue to be based on supply expansion that will allow Nvidia to catch up with demand. We add $4 billion to our quarterly revenue estimates for the following three quarters, reaching nearly $47 billion in data center revenue at this time a year from now. We then layer on another $1.5 billion-$2 billion in each of the following nine quarters through the end of fiscal 2028 (which is effectively calendar 2027). This math brings us to data center revenue estimates of $177 billion in fiscal 2026 (that is, calendar 2025) or 55% growth, $213 billion in fiscal 2027, and $240 billion in fiscal 2028. We surmise that investors have to be even more bullish than this to justify Nvidia’s stock price in the $145 range, where it recently traded.
Nvidia Networking Business
One of the blemishes on the quarter was Nvidia’s networking business, which was 14% of data center revenue for the past two quarters but came in at only 10% of revenue in the October quarter. In turn, networking revenue still rose 20% year over year, but not nearly as much as the 132% increase in compute (that is, graphics processing units) revenue, and sales were down 15% sequentially. Management was dismissive of the decline, and it foresees sequential growth in January and strong growth thereafter. We still anticipate that Nvidia’s networking business will see some nice attachment to its GPU business in the long run. However, it’s possible that Nvidia’s data center business may disappoint in the future if it can sell many GPUs but fails to attach its InfiniBand or Ethernet products along with some of these GPU sales.
We had modest long-term concerns around gross margins, as they are expected to dip into the low-70% range as the initial kinks are worked out of Blackwell production. However, management remains confident that margins can tick back to the mid-70% range, which likely stems from the durable competitive advantage Nvidia maintains in the software layer of its AI products.
Finally, our new model layers on a few billion more of gaming revenue per year, as we assume that Nvidia might bring a PC processor (CPU) to market to compete with Qualcomm, Apple, Intel, and AMD. We foresee many AI workloads being run on edge devices—such as PCs and smartphones—over time, and we anticipate that Nvidia might expand its business to capture this opportunity. We remain highly impressed with Nvidia’s ability to elbow out into new products and markets in recent years.
The author or authors do not own shares in any securities mentioned in this article. Find out about Morningstar’s editorial policies.
Keep Up to Date with the Most Important News
