Goldman Sachs has released a medium-term forecast for the opportunity of generative artificial intelligence for its club name NVIDIA (NVDA). The numbers are staggering. The boom in generative AI adoption could generate more than $90 billion in revenue for NVIDIA over the next three years, according to Goldman’s latest analysis revealed in a note to customers on Sunday. By the way, in the most recent fiscal year that ended in January, his Nvidia computing and networking division, which is based in selling data center GPUs used for AI applications, generated about $15 billion in revenue. . Goldman’s estimates underscore Nvidia’s huge growth opportunities, which is the main reason why Jim Cramer called Nvidia a “don’t trade, own stock”. Apple (AAPL) is the first and only stock to receive this designation from Jim. NVDA YTD Pile Nvidia YTD Performance The Nvidia stock has been performing great so far this year, thanks of course to its AI dominance. But our belief that there are brighter days ahead for the business (Goldman shares a similar view) gives us confidence in the stock’s multi-year trajectory. To estimate his AI revenue opportunity for Nvidia, Goldman has broken it down into his two computing categories: training and inference. Training will become a bigger revenue opportunity in the years to come, incorporating sales of Nvidia’s graphics processing units (GPUs). Nvidia’s best-in-class GPUs provide the fast computing power used to “train” large language models like OpenAI’s ChatGPT. At a high level, the training phase of an AI model feeds it a large amount of data to learn and make predictions. At the moment, inference has a smaller revenue opportunity. In the AI world, inference refers to making AI models work based on new user input or real-world scenarios not yet encountered, making predictions to complete tasks such as generating text and images. . Here’s a chart showing how Goldman sees his Nvidia AI revenue opportunity spread across each category from his 2023 calendar year to his 2025 calendar year. Please note that Goldman’s estimates are for calendar years, not fiscal years. Nvidia is one of the companies that does not traditionally follow a January-December fiscal year. The fiscal year usually begins in late January. For example, fiscal year 2024 he started on January 30, 2023. Those fiscal first quarter results were announced in May. To put Goldman’s AI computing revenue projections into perspective, his Nvidia company-wide revenues totaled $70.5 billion over the past three fiscal years. Just over half of this, $37.06 billion, came from the graphics division, which includes sales of GPUs for PC gaming. As previously mentioned, the rest of the revenue from that three-year stack comes from Nvidia’s other segments, computing and networking. As Goldman points out, the computing and networking sector stands to benefit greatly from the generative AI boom, but investors should expect these analysts’ forecasts to include a rebounding revenue contribution from gaming. It should be noted that it does not include Other smaller end markets like automobiles, which management says are feeling the momentum lately, are also not taken into account. Goldman’s estimates reflect his Nvidia’s position as a leading enabler of generative AI. Other revenue streams are just as important, though they have fallen out of the limelight as AI adoption has increased. Historically, the majority of inference tasks are completed by central processing units (CPUs). Goldman expects more inference workloads to be handled by accelerators (think GPUs) in the future. “In the short term, it will be small compared to training, but in the long term he believes reasoning will be a bigger and more regular opportunity for Nvidia and its peers,” the Wall Street firm said. says. Meanwhile, NVIDIA executives are also optimistic about the inference opportunity, with CEO Jensen Huang saying in his May that GPUs are well suited to handle inference for generative AI applications due to their “size and complexity.” claimed to be In March, Nvidia announced his four new inference platforms. Of the nearly $85 billion in training revenue Goldman estimates over the next three years (shown in his second column of the chart above), nearly $11 billion will come from his cloud service provider in China. Predicted. The company acknowledged that this forecast could be “in jeopardy” if the U.S. government tightens its restrictions on chip exports to China. In late June, amid reports that the Biden administration was considering tougher export controls, NVIDIA saw no short-term hit to its business given strong demand outside of China. said. However, NVIDIA’s CFO Collette Kress said that in the long run future business would be affected by such restrictions. Goldman’s new perspective on AI opportunities also served as the basis for Nvidia’s profit forecast hike. Analysts raised their price target to $495 from $440 and reiterated their buy rating on the stock. The new PT of $495 per share reflects a 17.4% premium over Nvidia’s closing price on Monday. Goldman now expects NVIDIA to post adjusted earnings per share of $11.99 in 2025 and $14.14 in 2026, 17.6% and 13% above FactSet’s consensus analyst forecasts, respectively. ing. Once this year, Wall Street significantly raised its EPS outlook for Nvidia in late May when it released its surprising second-quarter 2024 outlook, but Goldman said more ups from the analyst community are coming. suggesting that it is likely. “Our updated forecasts comfortably beat the street consensus and assume positive EPS revisions to support sustained share price outperformance through the remainder of the calendar year,” Goldman said in a statement to clients. rice field. (Jim Cramer’s Charitable Trust is a long NVDA. See here for a full list of stocks.) As a subscriber to Jim Cramer’s CNBC Investment Club, you’ll receive trade alerts before Jim makes a trade. receive. Jim waits 45 minutes after sending a trade alert before buying or selling shares in the charitable trust’s portfolio. If Jim talks about a stock on his CNBC television, after issuing a trade alert he waits 72 hours before executing the trade. The Investment Club information above is subject to our Terms of Use and Privacy Policy, along with our Disclaimer. No fiduciary duty or obligation shall exist or be created by your receipt of any information provided in connection with The Investment Club. No specific results or benefits are guaranteed.
Justin Sullivan | Getty Images
Goldman Sachs releases medium-term forecast Opportunities for generative artificial intelligence for club names Nvidia (NVDA). The numbers are staggering.
