Brad Gerstner isn’t only talking about a “supercycle” of AI. He is softly revealing how pros are attempting to jump on it.
The founder and CEO of Altimeter Capital is now one of the most carefully watched people in the AI sector. This is partly because he was one of the first to believe in Nvidia and other infrastructure winners.
But in his most recent interview on CNBC, he did something very helpful for Main Street investors: He explained why he is buying more of a very controversial AI stock while ignoring its closest competitors.
CoreWeave is a fairly young public business that calls itself “The Essential Cloud for AI.” Altimeter supported CoreWeave as a private firm, bought more when it went public in 2025, and then bought more after the stock fell along with other “neocloud” stocks, Gerstner said.
Gerstner made it plain that this investment isn’t a basket trade. He said that Altimeter is not adding anything to other neocloud operators.
The market, he believes, has shifted from an “all in” approach to an “all out” approach regarding the entire group, which has led to the exclusion of less experienced investors. He sees an opportunity for stock pickers in that blanket of negativity.
This narrative isn’t so much a warning as it is a plan: how a professional money manager thinks about CoreWeave’s wonderful growth, high debt levels, and close ties to Nvidia’s upcoming Rubin platform, and how an average investor should use that information to make judgments about their investments.
Gerstner offered further detail to CNBC, calling the company’s software stack, execution, and strategic alignment with Nvidia a “really interesting opportunity.”
It helps to know what CoreWeave does to understand why Gerstner is prepared to lean into it.
CoreWeave is not a cloud that can do everything, like the big ones that operate email, streaming, and office software. Rather, it combines a lot of Nvidia GPUs into very dense clusters and wraps them in its own software and operations, so that clients can train and run big AI models on a massive scale.
In simple terms, it doesn’t only rent out computer capacity; it also rents out AI “factories.” That makes CoreWeave a really good method to bet on the future of AI as a whole, instead of on a single chatbot or consumer app that might go away.
The basic profile of CoreWeave for most investors looks like the following.
It is a specialized cloud that is built on Nvidia GPUs, not a general “do-everything” cloud.
Its customers are AI-heavy companies that need to train and run large models, not just handle simple web workloads.
It seeks to stand out by offering more than simply access to raw chips. It has its own software layer and operational know-how.
Altimeter focused on it early, because of the company’s own figures. CoreWeave made nearly $1.36 billion in the third quarter of 2025, which is more than double what it made in the same quarter the year before.
Management also noted a backlog of $55.6 billion in revenue. That includes long-term contracts from consumers, not just short-term tests. Michael Intrator, the CEO, hailed it “an exceptional third quarter,” citing “disciplined execution” and stronger ties with big AI customers.
More AI Stocks:
Some of the most important names in this cycle are in that backlog. CoreWeave has made multi-billion-dollar infrastructure arrangements with a major social network, a top AI lab, and Nvidia itself through a strategic partnership to increase GPU capacity, according to company disclosures and outside reports.
While the contracts may not guarantee a seamless journey, they demonstrate that CoreWeave is not merely renting surplus chips on an hourly basis. It is increasingly becoming central to the long-term AI goals of its top clients.
This is when Gerstner’s plan really takes shape. He isn’t just paying for a tale about “AI demand.” He is starting with demand he can see in an earnings release: money that is already coming in and tens of billions of dollars more that customers have promised to spend.
He will only address the next question after that base is in place: Is the market mispricing the risks by treating every stock in a specialty the same way?
That’s what he means when he says that CoreWeave “stands alone.” Gerstner thinks the company is not just another cloud operator that promises too much.
It is a specialty with genuine production workloads, a unique software layer, and direct strategic importance to Nvidia’s next generation of hardware.
The AI hype may be cooling, but CoreWeave stock is getting surprising attention.Photo by FREDERIC J. BROWN on Getty Images
Gerstner’s plan for CoreWeave comes down to a few key pressure points for Main Street.
Backlog versus balance sheet: The $55.6 billion backlog demonstrates significant demand, but the quarterly interest bill of more than 300 million dollars shows how hard it is to supply that demand. Over time, both lines will move the stock.
Advice and timeline for the project: One delayed data center project was enough to bring down full-year revenue forecasts once. In a business that needs a lot of money, timetable slips might be just as important as any one revenue beat.
A lot of customers: A few big AI companies have made multi-billion-dollar deals with the platform, which proves it works. But if one of these big clients cuts back on spending or changes their plans, it can substantially affect results. Over the next few years, investors should keep an eye on how diverse CoreWeave’s biggest contracts get.
Degree of risk you can handle: This stock can go up or down 20% based on a single headline. If that kind of swing would keep someone up at night, Gerstner’s own plan says to use more well-known AI names for core holdings and think of any CoreWeave stake as a small, risky piece.
These bullet points represent what Gerstner, a professional, considers as he evaluates whether the return is still worth the risk.
The most forward-looking part of Gerstner’s argument is CoreWeave’s part in Nvidia’s Rubin platform, which is the next phase in the chipmaker’s AI roadmap.
At a trade event in January, Nvidia CEO Jensen Huang formally announced the Vera Rubin AI platform and said Rubin systems are already in full production. Nvidia has pushed Rubin as a means to achieve up to five times the training performance of its Blackwell generation, while markedly decreasing the cost of operating large AI models in production.
Nvidia’s own release named a few cloud companies that will be among the first to use Rubin-based instances in the second half of 2026. It named a group of specialist partners, including CoreWeave, in addition to the usual hyperscalers.
CoreWeave promptly released a statement indicating it will integrate Rubin technology into its AI cloud platform. The business said Rubin will help with “agentic AI, reasoning, and large-scale inference workloads.”
It also said it would add Rubin systems to its Mission Control software, which handles several racks of hardware as one AI “factory.” Huang called CoreWeave an “essential partner” in developing AI factories of the future.
From an investor’s perspective, Rubin connects three parts of the story:
It’s not simply a chip update; it’s Nvidia’s plan for the next generation of AI data centers.
If Nvidia fulfills its commitments, it presents an opportunity for early partners such as CoreWeave to maintain a competitive edge in terms of performance and cost.
It gives a highly leveraged company a chance to grow into its balance sheet, instead of being crushed by it.
If Rubin keeps Nvidia’s promises and CoreWeave does its part in the rollout, the firm might strengthen its ties with important customers, get better deals on future GPU capacity, and expand into the big balance sheet it already has.
In that case, Gerstner’s opinion that the company has “fallen out of favor” and is now trading at a discount to its potential would seem smart in hindsight.
For Main Street, the other side is just as vital. If Rubin deployments are delayed, AI expenditure stops, or a few big customers change how they divide up their work, CoreWeave might find up with very expensive equipment, a hefty interest bill, and a market that has previously shown how quickly it can turn negative.
In such scenario, the same leverage that makes things better would also make things worse.
That is why Gerstner continues distinguishing his CoreWeave viewpoint from his core AI positions. He says Nvidia is still the most important AI infrastructure company in the ecosystem.
CoreWeave, on the other hand, is a more risky way to state his thesis: a focused bet that the AI supercycle lasts long enough and Rubin falls hard enough for a highly leveraged specialist to turn today’s backlog into tomorrow’s steady cash flow.
The practical lesson for the typical investor is clear. Gerstner’s own plan is that diversified titans like Nvidia should be the base of the portfolio if you want to get into AI without worrying about it all the time.
If you want something “spicier,” a stock like CoreWeave should go in a small, clearly labeled corner with money you can afford to lose while the AI build-out proceeds over the next few years.
