The street will stick with NVIDIA (NVDA) to next week’s marketing revenue report.
Despite China-based DeepSeek’s super bullish AI papers coming out earlier this year, Wall Street still sees Nvidia clean up the global AI infrastructure. This is emphasized by the radical 2025 CAPEX assumptions of high standards such as Amazon (AMZN) and Meta (Meta) in this income season.
“In the coming decades [in AI] It is happening.
However, this is not to say that Nvidia’s earnings report shows no signs of caution.
Yahoo Finance data shows that NVIDIA’s first-quarter earnings per share trend has slightly lowered in the past 30 days. The street has not yet improved its 2025 EPS estimate for NVIDIA, more than 60 days.
NVIDIA is also one of the cheapest AI stocks right now.
Yahoo financial data shows 32 NVIDIA transactions with multiple forward price-to-earnings ratios (PE). Broadcom (AVGO) and Marvell Technology (MRVL) have 35 and 41 respectively. The arm hold (ARM) clock is 76 times.
“I would say I don’t want to be Jenson [Nvidia CEO Jensen Huang] It must be because wow, other people are doing the same thing.
Here is the resonance on the streets a few days before Nvidia’s February 26 revenue.
“Although recent concerns over the long-term demand impact of DeepSeek’s lower-cost approach, most high-standards have actually increased the growth of 2025 CAPEX by 4%-26% in the income season, supporting our argument , i.e. overall AI GPU demand outlook remains strong. Nevertheless, we lowered our fiscal 26-year data center revenue estimates by 11% to $209 billion on the back of the low NVL rack [Nvidia data center product] Assuming, but our data center revenue forecast for fiscal 26 remained at $184 million compared to 14% of consensus. However, we believe that NVIDIA is more stressed on the back of its B300/GB300 roadmap to provide stronger second-half fiscal year momentum to see if there is room for further upside. ”
-
grade: Reiterate overweight
-
Price Target: $190, above $180
“While we do think that manufacturing restrictions are limiting the shipping of GB200 NVL server racks, we think this will be offset by the following points: 1) Given the low initial manufacturing output of GB200 NVL, we believe customers have been able to launch with X86 GB200 and backfill orders for HGX-based B200 servers of head nodes; 2) [custom chip]demand for H20 GPUs has surged since China CSPS; 3) We believe that NVIDIA customers, especially communications service providers, are in stock funds from EMS providers, so goods sold effectively from NVIDI to EMS are considered revenue . Therefore, we are improving our estimates and our target goals and believe that NVIDIA’s strong results should alleviate any concerns about DeepSeek’s potential derailment of the AI CAPEX intensity in the near future. ”
