Nvidia CEO Jensen Huang gives a keynote speech with Grace Blackwell NVLink72 in hand. … (+)
AFP (via Getty Images)
According to the Wall Street Journal, on January 6, Nvidia CEO Jensen Huang’s keynote speech at the 2025 Consumer Electronics Show will include products for industries such as robotics, automotive, and gaming. A series of announcements were made that appeared to be relatively small.
Are these mini announcements a distraction? One analyst believes the biggest news may be Nvidia’s strategy to compete with cloud services customers Amazon, Microsoft and Alphabet by building its own AI cloud, Investor’s Business Daily reported. I am.
Nvidia, whose stock rose about 4% in premarket trading, is excited about the opportunity to provide software to robot developers. “The ChatGPT moment for robotics is here,” Huang told CES attendees, according to IBD.
“Like large-scale language models, world foundation models are the foundation for advancing robot and AV development, but not all developers have the expertise and resources to train their own. “We created Cosmos to democratize physical AI and make common robotics available to all developers,” he added.
An NVIDIA spokesperson I contacted declined to comment.
Nvidia’s CES product announcements
If there’s one thing I learned while writing my latest book, Brain Rush, it’s this: That means Nvidia’s future relies heavily on working with companies close to the end users of its technology.
Because Nvidia, which designs chips and builds software for developers, allows companies close to end users, such as chip manufacturers, cloud service providers, enterprise software companies, and management consultants, to make Nvidia’s products part of their enterprise. Only by encouraging them can they enjoy rapid growth. It is a valuable system for businesses and consumers.
Nvidia’s CES product announcements aim to strengthen these partnerships. The industries we aim to strengthen with this announcement are:
Robotics. At CES, Huang demonstrated a “physical AI” simulation tool that helps robots learn. According to Reuters, these Cosmos-based models can “generate photorealistic videos” more efficiently than traditional data to train robots and self-driving cars (Uber, for example, is a Cosmos customer). ” Nvidia sees a $38 billion market opportunity “over the next few decades” by bringing robots into warehouses and factories, the magazine said. car. Nvidia, which currently partners with Mercedes and Volvo, announced a deal for Toyota to buy driver assistance chips and software for future vehicles. Huang said auto sales could rise 25% to $5 billion in the next fiscal year, which starts at the end of this month, the newspaper reported. game. Nvidia announced chips for desktop and laptop PCs based on the Blackwell architecture used for AI to provide better graphics and higher resolution for gamers. RTX 50 series chips launching in March will support a feature called DLSS 4, which uses AI to improve frame rates in games, CNBC reported. Nvidia’s gaming revenue for the company’s quarter ending in October was relatively small, “accounting for less than 10% of total revenue, compared with 88% from data center chips,” CNBC notes. did. AI researchers and data scientists. Nvidia has announced Project DIGITS, a desktop AI supercomputer (starting at $3,000) powered by the company’s latest Blackwell AI chip. This will allow AI researchers and data scientists to work on AI models without relying on “Nvidia’s most advanced AI chips housed in data centers,” the magazine reports. . Agent AI. Nvidia has released software to automate work for enterprises called Agent AI Blueprint. According to Brain Rush, Agentic AI performs a series of tasks. For example, plan your vacation, compare flights, hotels, excursions, book and pay for the best available. Nvidia’s AI blueprints, such as PDF to Podcast, can help researchers quickly understand key points in long PDFs, IBD said.
Nvidia’s AI cloud strategy
Nvidia didn’t announce this at CES, but the company’s biggest news may be its strategy to compete with cloud services customers, which account for most of Nvidia’s growth.
In fact, one analyst has noticed an increase in Nvidia’s demand for leased data center capacity with AI processing. Such AI data centers “rent servers with their own chips and host AI software development platforms,” Investor’s Business Daily noted.
The analyst sees NVIDIA’s cloud computing efforts as a surprise for the new year. “We believe Nvidia is looking to procure more capacity for the ‘AI cloud’ and may accelerate data center leasing in 2025. ,” TD Cowen analyst Michael Elias said in a report featured by IBD.
IBD reports that in addition to seeking additional data center capacity in other deals, Elias is involved in an “apparently large lease agreement between Nvidia and Digital Realty for data center capacity in Northern Virginia.” also mentioned.
Nvidia’s AI cloud will compete with so-called Neoclouds (AI processing cloud services such as CoreWeave, Lambda, Crusoe, Vultr, and Together AI), as well as Nvidia’s biggest AI chip customers, Amazon and Microsoft, Elias wrote.
By building its own neocloud, Nvidia will be able to reap more profits from the rapidly growing industrial sector. Nvidia currently leases and distributes its AI software via AWS, but operating its own cloud could increase profits. According to Pitchbook, the AI neocloud is expected to grow from $4 billion last year to $32 billion by 2027 at a compound annual rate of 100%.
Meanwhile, Nvidia’s cloud customers are buying more AI chips from rivals and looking to build their own custom chips. For example, in addition to buying Blackwell chips from Nvidia, hyperscalers are buying AI processors from Broadcom and Marvell Technologies, designing their own custom AI accelerators, and looking to sell them to OpenAI, Apple, and Meta Platforms. reported IBD.
Where will Nvidia stock go?
Nvidia’s stock price could rise. Based on 12-month price targets from 40 Wall Street analysts, NVIDIA’s stock would need to rise 18.5% to reach the average estimate of $135 per share, according to TipRanks. It is said that there is.
Below are comments from analysts who are less bullish and those who are more bullish.
Robotics is a niche opportunity. Bank of America analyst Vivek Arya said in a note to clients featured in Reuters: “The challenge we see… The goal is to make the product affordable and widely available.” There are still cool but niche opportunities like the Metaverse and self-driving cars,” Arya added. Nvidia’s market capitalization could reach $5 trillion by mid-2026. “He’s more bullish,” Wedbush analyst Dan Ives said. By building on the company’s technological advantages and expanding into new vertical markets, Nvidia is opening the door to a “$5 trillion valuation that could be achieved within 18 months,” MarketBeat said.
For me, the key question for investors is: How much revenue can Nvidia’s AI cloud generate? Will Nvidia’s hyperscaler customers buy less from Nvidia?My guess is that Huang expects Nvidia’s AI cloud services to help his company run better.
