Nvidia projects 70% revenue growth for next fiscal year after Q2 sales reach $96.2 billion

Financial results and fiscal outlook

Nvidia posted second-quarter revenue of $96.2 billion, exceeding Wall Street projections on the back of $89 billion in data center sales, an increase of 117% over the prior year. For the third quarter, the Santa Clara-based chipmaker projected revenue of $108.0 billion, plus or minus 2%, surpassing the average analyst estimate of $104.19 billion compiled by LSEG. Nvidia stated that its forward guidance assumes zero data center processor sales to China. Alongside the quarterly numbers, management gave a forecast projecting a 70% revenue expansion for the next fiscal year, while cautioning that memory component shortages could limit how quickly output can scale.

Following the results and subsequent conference call, market strategists evaluated the durability of capital expenditures across the sector. Charu Chanana, chief investment strategist at Saxo Markets, described the operational constraints as evidence of sustained backlogs.

The biggest positive was management effectively telling investors that AI demand remains supply-constrained even at this scale... That is a powerful counter to the narrative that the AI capex cycle is already peaking.

Expanding client base and industry investments

Nvidia indicated that processor demand is broadening beyond hyperscale cloud providers to include specialized AI research laboratories, cloud infrastructure firms such as CoreWeave and Nebius, and an expanded commercial relationship with Amazon Web Services. Companies including Microsoft and Meta Platforms had previously reiterated expectations that aggregate Big Tech infrastructure spending will exceed $730 billion this year, up from $400 billion in the prior year. Additionally, Nvidia entered an arrangement with six financial institutions to support deals aiming to raise over $500 billion for AI infrastructure.

The company has also pursued intellectual property and platform acquisitions. In March, Nvidia licensed technology from startup Groq in a $17 billion transaction to pair Groq inference chips with its upcoming Vera Rubin architecture, and reports from The Information indicated an agreement to acquire open-source platform Hugging Face for $12.9 billion.

Chris Weston, head of research at Pepperstone Group, addressed the broader supplier implications during Asia-Pacific trading.

The business is in absolutely rude health. It's going to wake people up and think 'what have I missed?' It's a really bullish story and anyone supplying Nvidia through the supply chain will likely see the benefit today.

Premarket equity and derivative movements

Nvidia shares gained 7.4% in premarket trading on Thursday, having earlier fluctuated between a 1.2% decline in initial extended trading and a 4% to 6% advance during the earnings call. The forecast lifted related semiconductor and hardware manufacturers. In US premarket activity, Micron Technology advanced 4.2%, Marvell Technology gained 5.2%, and Broadcom rose 1.6%, while data storage providers Sandisk and Western Digital added 4.9% and 3.4% respectively.

Software companies also posted gains after releasing their own earnings updates. Salesforce climbed 11.8% after raising its annual revenue and profit guidance and introducing a plug-in integrating Anthropic Claude models, while CrowdStrike rose 8.8% following higher full-year revenue targets and a second-quarter earnings beat. In equity index futures, Nasdaq 100 E-minis advanced 300.25 points (1.03%) at 05:05 a.m. ET, S&P 500 E-minis added 32 points (0.42%), and Dow E-minis slipped 34 points (0.06%).

Comparative valuations and competitor multiples

Following the earnings release, at least 10 brokerage firms raised their price targets on Nvidia stock. As of Tuesday's market close, Nvidia shares had gained 14.2% year-to-date, trailing competitors Advanced Micro Devices and Intel, both of which had more than doubled over the same period. Nvidia shares traded at 17.9 times forward earnings estimates, compared with 37.2 times for Advanced Micro Devices and 46.2 times for Intel.

The forward multiple reflects ongoing competition in the inference processor market, where rivals AMD and Intel, as well as Chinese domestic developers such as Baidu, are deploying alternative architectures.