Nvidia Company (NASDAQ:NVDA) might be reporting Q2 ’25 earnings on August 28 after market shut, bringing analysts to a divide between continued progress primarily based on the H100 demand vs. a slowdown as a results of the delays in Blackwell. Regardless of the challenges voiced by techniques integrators reminiscent of Tremendous Micro Laptop (SMCI), Nvidia’s efficiency might not completely be depending on the discharge of the B100 on condition that demand nonetheless outstrips provide of the H100 GPUs. Given these elements, I stay bullish on NVDA shares and reiterate my BUY ranking with a value goal of $165/share at 19.50x eFY26 value/gross sales.
Within the final 90 days, analysts have revised their EPS estimates up 19x and down 12x. Analysts are forecasting Nvidia to report income within the vary of $26.8-30.45b with $28.7b on the midpoint and EPS in the vary of $.06-0.71/share with consensus coming to $0.64/share.
I’m modeling eq2’25 income to come back in at $28,733M with an adjusted diluted EPS of $0.68/share.
Financial institution of America (BAC) VP of Derivatives Analysis, Gonzalo Asis, is suggesting to hedge the danger of an earnings disappointment. Different analysts, reminiscent of Fairness Strategist, Scott Rubner, of Goldman Sachs, Dan Ives of Wedbush, Oppenheimer Asset Administration, and Blayne Curtis of Jefferies all anticipate a robust efficiency for Nvidia at their Q2 ’25 earnings. Total, I facet with these analysts as Nvidia continues to understand energy throughout all segments as enterprises, hyperscalers, and shoppers are in search of increased compute speeds throughout their particular compute wants.
Nvidia Operations
Beginning with Information Heart, Nvidia continues to understand exceptionally sturdy sequential progress for the section pushed by each hyperscalers and enterprises for his or her Grace Hopper structure. Administration acknowledged of their q1’25 earnings name that demand for the H100 continues to outstrip the agency’s provide capabilities as extra enterprises flip to accelerated computing. Accordingly, 40% of the deployed AI infrastructure in q1’25 was pushed by giant cloud suppliers with the rest being enterprises. That is primarily pushed by the necessity to construct out the infrastructure able to supporting these high-powered, data-driven AI/ML fashions. The use circumstances have gotten an increasing number of obvious as giant corporations reminiscent of Tesla (TSLA) are leveraging a large 35,000 H100 GPU cluster to coach their AI platform.
Given the latest hints of Blackwell being delayed into CY25, I consider some traders could also be hesitant on Nvidia going into eq2 ’25 earnings. Regardless of the headwinds, I don’t consider it should lead to a fabric influence to income progress given the surplus demand for the H100 GPUs. I anticipate the H100 to stay as a driving pressure for Nvidia even after the discharge of the H200 as enterprises put money into scaling out their AI capabilities to ship cost-saving/money-making functions, reminiscent of these supplied via Palantir’s (PLTR) platform. One attention-grabbing level administration talked about on the q1 ’25 earnings name was that the Blackwell GPUs are backwards suitable. This means that an enterprise or hyperscaler can construct out their Grace Hopper system with the H100 or H200 and implement the B100 when launched. Given the one-year launch cycle, it shouldn’t be anticipated that end-customers will delay investments for the next-generation launch. I consider this may lead to enterprises averaging up their infrastructure as they scale out their AI factories.
As for the Blackwell structure, the B100 is predicted to carry out 30x quicker for inferencing when in comparison with the H100 with a 25x decrease TCO and vitality consumption. Administration steered that the Grace Blackwell structure will embody the fifth era NVLink, a multi-GPU backbone, and InfiniBand and Ethernet switches. This versatile structure will assist hyperscalers and enterprises for each coaching and inferencing.
Along with this, networking gear suppliers Cisco (CSCO) and Arista Networks (ANET) every steered that the subsequent networking gear buying cycle has begun and that enterprises and hyperscalers are every rising their investments for increased pace gear. This development might drive vital enchancment to shipments for Spectrum-X Ethernet and InfiniBand.
For eq2’25, I’m forecasting the Information Heart section to offer $25b in income, a 144% year-over-year progress fee.
Turning to the opposite segments, administration anticipates Gaming, Skilled Visualization, and Automotive to every generate sequential progress in eq2 ’25. Nvidia is experiencing sturdy reception to their GeForce RTX GPUs for players, creators, and operating AI functions. Microsoft (MSFT) not too long ago introduced their partnership with Nvidia to optimize AI efficiency for Home windows, which is predicted to run LLMs as much as 3x quicker when utilizing GeForce RTX AI PCs. Although I don’t anticipate client demand for AI PCs to point out materials enchancment in CY24, enterprises might start adopting these highly effective machines to work in tandem with their AI-enabled knowledge facilities.
Nvidia can also be progressing their automotive choices with the Nvidia DRIVE Thor, which might be powered by the Blackwell structure. Nvidia is presently working with numerous Chinese language electrical automobile makers and is anticipating Thor to enter manufacturing automobiles beginning in 2025.
Nvidia Financials
Trying to the financials, I anticipate complete income to develop by 113% to $28.7b in eq2’25, primarily pushed by the Information Heart section. Administration steered of their Q1 ’25 earnings name that they anticipate every section will expertise sequential progress in eq2 ’25. Given my income forecast and administration’s margin forecast, Nvidia ought to generate $19.2b in adjusted working earnings and $0.68/share in adjusted EPS.
Dangers Concerned With Nvidia
Bull Case For Nvidia
Demand for Nvidia’s H100 GPUs stays stronger than what the corporate is able to supplying. This means that the agency could also be minimally impacted by the potential delay within the launch of the B100 GPUs. Sequential enchancment throughout segments with a rising curiosity in AI PCs throughout enterprises will drive vital income progress and doubtlessly result in stronger year-over-year margin enchancment.
Bear Case For Nvidia
Nvidia might expertise some headwinds on account of the delay within the B100 launch. This may occasionally lead to hyperscalers delaying some investments, anticipating the quicker and extra power-efficient GPUs. The broader financial system can also be experiencing some sturdiness challenges, as steered within the latest jobs report revision for 818,000 fewer jobs. Slower employment progress or a better unemployment fee might result in decrease demand for AI PCs and doubtlessly a smaller AI footprint.
Valuation & Shareholder Worth
NVDA shares presently commerce at 39.45x trailing/value gross sales, making the shares priced to perfection going into eq2’25 earnings. Given the wealthy valuation, shares might react with vital volatility in both course, relying on traders’ sentiment in the direction of the delays in Blackwell and a possible slowdown in income progress. Given the sturdy demand for the H100, I consider NVDA will launch one other sturdy earnings report that will drive shares positively.
Regardless of the excessive buying and selling a number of when in comparison with its friends, I consider NVDA’s valuation is justified given the numerous progress outlook for the agency within the coming years. If extra corporations flip to remodeling their conventional compute knowledge facilities to accelerated AI factories, Nvidia might preserve their increased valuation in tandem with their progress.
Valuing the corporate primarily based on their historic buying and selling multiples and my forecast for gross sales era, I consider NVDA shares have vital room for extra progress going ahead. I reiterate my BUY ranking for NVDA with a value goal of $165/share at 19.50x eFY26 value/gross sales.