Analysts expect Nvidia's Q2 revenue to be $28.705 billion, adjusted net income to be $15.912 billion, and adjusted EPS to be $0.645, according to Bloomberg's unanimous expectations.
AI darling Nvidia will report its second-quarter results of fiscal 2025 on Aug. 28, with significant attention expected on management's comments regarding potential delays in the release of its next-generation Blackwell (BW) chip, and the company's ability to meet demand.
Analysts expect Nvidia's Q2 revenue to be $28.705 billion, adjusted net income to be $15.912 billion, and adjusted EPS to be $0.645, according to Bloomberg's unanimous expectations.
Previous Quarter Review
Nvidia reported fiscal first-quarter earnings that beat expectations for sales and earnings. Its strong results suggest that demand for the AI chips Nvidia makes remains robust.
CEO Jensen Huang said the company would begin to see revenue from its next-generation AI chip, called Blackwell, later this year.
The company also forecasted quarterly revenue above estimates on Wednesday and announced a stock split.
What to Watch for Q2
Nvidia’s data center business is likely to remain the primary driver of the company’s earnings for the quarter, led by the Hopper GPU computing platform and the InfiniBand end-to-end networking solutions as customers across various industries have been investing in training and interfacing generative AI and large language models.
Investors will be focused on two key aspects in Nvidia's upcoming earnings report: the evolution of demand and the release of its new product, Blackwell.
The Evolution of Demand
AI revenue is coming more into focus, as investors are rightfully questioning whether companies will generate enough to justify existing AI spending.
After a meteoric rise in the past years, Nvidia stock lost its sheen somewhat after its recent slump of more than 20%. The selloff was triggered by concerns related to a possible capex slowdown by Nvidia’s top customers and hyperscalers Microsoft, Meta Platforms, Amazon and Google.
However, all these companies have pledged to boost their spending on AI-related infrastructure even higher. Google parent Alphabet said its future capex will be “at or above the first-quarter level of $12 billion.”
Microsoft, which is Nvidia’s largest customer, reported capital spending of $55.7 billion for the fiscal year (ended in June) up 75% from last year, and noted that capital spending in fiscal 2025 will top that level to meet the growing demand signal for its AI and cloud products.
Meta stated that it currently expects significant capex growth in 2025 as it invests to support its AI research and product development efforts.
JPMorgan sees the four hyperscalers–Google, Microsoft, Meta and Amazon–spending a collective $200 billion in capex this year with AI buildout dominating the capital spending, which in turn bodes well for Nvidia.
Blackwell’s Any Possible Ramp Delays
Another cause for concern centers around the rumored delay in the rollout of Nvidia’s next generation Blackwell AI chips, which are roughly two times faster than Nvidia’s current Hopper models, but with notable improvements in energy efficiency. The delay is reportedly attributed to the complexity of the chip-on-wafer-on-substrate (CoWoS) packaging technology used by Taiwan Semiconductor Manufacturing Company (TSM) that manufactures these chips for Nvidia. The Blackwell launch is scheduled for later this year.
However, Nvidia commented that Blackwell sampling has started, and production is on track to ramp in the second half. Nvidia's hardware partners, like Foxconn, Quanta, Wistron, Pegatron, and Asus, have already demonstrated their Blackwell-based servers at Computex, a computer expo held annually in Taipei, Taiwan. Investors are hoping that Nvidia may provide a more specific timeline for the Blackwell ramp in its upcoming earnings call, beyond an ambiguous reference to the second half of the year.
Even assuming there is a delay in a worst case scenario, it is good to know that Nvidia is working on fixing the design complexities rather than bringing out a flawed model that is bound to fail. The delay (if there is one) is unlikely to hamper Nvidia or its prospects in the longer run. Don’t forget, Nvidia is still the undisputed leader in the data center GPU market, and even if AMD brings out a rival product comparable to Blackwell, it will still have to catch up with the yawning chasm in market share.
Analysts' opinions
Stifel analysts said Monday their discussions with industry participants revealed that any delays are likely to be measured in months rather than quarters. In the interim, feedback on the demand for H-Series chips remains positive.
As a result, the investment bank again expects a beat-and-raise scenario for Nvidia’s upcoming results and guidance.
Separately, UBS analysts said they see upside to Nvidi’a July print, anticipating continued strength from the data center segment, which could reach as high as $26 billion, above the Street’s estimate of $25 billion.
For guidance, UBS expects strong demand for Hopper and H200 to drive Nvidia to guide toward $31-32 billion, with an implied data center revenue close to $28 billion.
Moreover, Goldman Sachs analyst Toshiya Hari, who reiterated his 'conviction buy' rating on Nvidia, as well as his $135 price target, heading into next week's earnings update, said the Nvidia demand story remains compelling.