CNBC Today On NVIDIA Stock After NVIDIA Earnings - NVDA Update
FinVid
Nvidia Beats on Revenue and Earnings 0:00
Nvidia reported adjusted earnings of two dollars and twenty two cents a share on revenue of ninety six point two billion dollars, beating Wall Street's expectations of two dollars ten cents and around ninety two point three billion dollars. Data center revenue came in at eighty nine billion dollars, slightly above the eighty six billion analysts expected. Despite the beat, shares initially dropped about two percent after hours, largely because the guidance and margin numbers came in just under what the more optimistic buy side investors were hoping for.
Guidance Falls Short of Buy Side Hopes 0:30
Nvidia guided third quarter revenue to one hundred eight billion dollars, comfortably above the official Wall Street estimate of about one hundred four billion, but short of the roughly one hundred nine billion dollar whisper number that buy side investors wanted. This guidance excludes any revenue from China. Analyst Gene noted the company needed at least a five billion dollar beat to fully satisfy the market, and it only delivered about four billion, which added to the early disappointment even though the underlying numbers remained historically strong.
Gross Margins Under Pressure 0:30
Gross margin for the quarter just reported came in exactly at the expected seventy five percent, but the guide for the next quarter fell to seventy four percent, just under the seventy four point seven percent the street wanted. Nvidia had promised margins in the mid seventies, and this dip reflects rising memory and component costs. Analysts on the panel noted this was the detail that most surprised them, since they expected a more gradual decline rather than an immediate step down.
Hyperscaler Growth and Customer Mix 3:00
Nvidia now splits its data center revenue between hyperscalers and other customers such as AI cloud, industrial, and enterprise buyers. Hyperscaler revenue more than doubled, up one hundred two percent, while the non-hyperscaler segment climbed even faster at one hundred thirty eight percent. Analysts pointed to SpaceX as an example of a customer that could grow from about three percent of Nvidia's business to ten percent, given plans for eight gigawatt hours of data center capacity next year, a shift that alone could push hyperscaler growth from an expected thirty eight percent to the low sixties.
Supply Commitments and Balance Sheet Concerns 5:01
Nvidia's purchase commitments jumped sharply from one hundred nineteen billion dollars to two hundred seventy nine billion dollars, mostly tied to securing memory supply. Analysts framed this as evidence that while many companies can design a chip, only a few can lock up the supply chain at this scale. Nvidia also disclosed fifteen year data center leases starting in fiscal 2028 and 2029 that it plans to eventually hand off to third parties, a sign of how deeply involved it has become in the physical buildout beyond simply selling chips.
Why the Stock Trades at a Discount 7:02
Panelists questioned why Nvidia trades at only about seventeen times next year's earnings despite revenue growing over one hundred percent and gross margins near seventy five percent, a valuation lower than the broader market. Analyst Chris Roland suggested this partly reflects Nvidia's sheer size, since even large fund managers can only allocate so much to a single stock, and partly reflects unresolved bear cases in the market that keep a cap on the multiple.
The Fiscal 2028 Growth Number That Moved the Stock 9:30
During the earnings call, Nvidia's CFO said the company expects to grow revenue by approximately seventy percent in fiscal 2028, far above the roughly forty four to fifty percent analysts had expected. Jensen Huang added that actual demand is even higher than seventy percent, meaning the guidance itself is conservative and constrained by supply rather than by demand. This clarity is what analysts pointed to as the main reason the stock moved higher after hours, even after the initial dip caused by the light margin guidance.
China, Competition, and Supply Constraints 11:31
Analyst Chris Roland described China as both an opportunity and a limited threat, since Nvidia is still waiting to ship meaningfully there, possibly due to an informal ban at the provincial level, while Chinese domestic chip capacity remains limited to around seven nanometer technology. On competition, panelists noted rivals like OpenAI developing their own chips, though Nvidia executives argued their approach differs by building a full AI factory platform usable across any cloud, rather than narrow chips built for one service.
Cash Flow Drop Tied to Timing, Not Concern 17:02
Nvidia's free cash flow fell to twenty one point three billion dollars from forty eight point five billion the prior quarter, which initially unsettled some market participants. This was driven by a large increase in accounts receivable, creating a roughly twenty two point three five billion dollar negative adjustment, meaning revenue had been recognized but the cash had not yet been collected by quarter end. Commentary framed this as a timing issue tied partly to higher cash taxes, not a sign of underlying weakness.
Addressing Circular Financing Concerns 18:00
Nvidia provided a table breaking down its future commitments, showing that the vast majority are supply and capacity commitments meant to help the company produce more product, not investments in cloud service agreements with customers. This was presented as evidence against the widely repeated claim that Nvidia is funding its own growth through circular financing, since its investments in cloud contracts are small compared to its commitments to suppliers.
Highlights From the Earnings Call 20:31
On the call, Nvidia announced an expanded partnership with AWS involving two million additional GPUs through the second quarter of fiscal 2029, and said non-hyperscaler customers will represent roughly half of its data center business going forward, reducing dependency on a small number of large cloud buyers. The company also said it expects to ship its Rock 3 LPX chip in volume later in the quarter, and that it has invested nearly fifty billion dollars in so-called Frontier Labs, which it believes will become the largest technology companies in history.
Rising Memory Prices to Pressure Margins Further 22:01
Nvidia's CFO said the company is experiencing extreme pricing conditions in memory, with price increases exceeding prior expectations and expected to climb even higher next year. Gross margins are expected to be seventy four percent in the third quarter, bottom out between seventy one and seventy two percent in the fourth quarter, and then settle between seventy two and seventy three percent in fiscal 2028, with further price increases taking effect in the first quarter.
Guarantees and Exposure Explained 24:01
Nvidia detailed its exposure related to financial guarantees, including a one hundred five billion dollar credit backstop tied to a data center campus in Ohio built with SB Energy, plus three point five billion dollars in land, power, and shell guarantees for AI clouds, totaling roughly one hundred eight point five billion dollars. A separate guarantee tied to OpenAI is capped at fifteen billion dollars, phased in as conditions are met, and declines as OpenAI fulfills its lease payments.
Upcoming Jensen Appearances 25:31
Jensen Huang is scheduled to speak at the Goldman Sachs Communicopia and Technology Conference on September 10th, and again at GTC Berlin on October 21st.
The Long-Term Growth Thesis 25:31
The speaker says he is confident Nvidia will be worth far more in future years, noting that Jensen has raised the possibility of Nvidia becoming a three trillion dollar revenue company, which could eventually push its market cap into the tens of trillions. He argues the world remains compute constrained through at least the first half of 2028, unlike the dotcom era, since today there is no unused capacity sitting idle. Hyperscalers report being supply constrained and are building data centers against signed contracts and prepayments rather than speculative bets, and surging revenues at leading AI labs like Anthropic and OpenAI reflect direct ties between compute and revenue.
Agentic AI And Physical AI Ahead 28:30
He points to 2026 as pivotal due to rapid adoption of Agentic AI and agentic systems in large enterprises, which is driving inference and compute demand higher. Beyond that, he expects physical AI, robots and machines performing real-world tasks, to fuel the next surge, calling it a multi-trillion dollar opportunity per Nvidia's CFO. Nvidia supplies training hardware, the Omniverse platform, and on-device inference through NVIDIA AGX, with over 3 million developers already building on its robotics stack. He also outlines a product roadmap through 2028, including Blackwell Ultra, Vera Rubin, Grok 3 LPX, Reuben Ultra in 2027 and Fineman in 2028, alongside Jensen's forecast of 3 to 4 trillion dollars in annual AI infrastructure spending by decade's end.
AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.
