The AI funding window is closing, and everyone wants a seat on the train. Clearly, the rush to raise capital is on, as more and more artificial intelligence-focused companies are feeling pressure to continue building to keep up. We know the demand for everything AI — from models to agents to chips to data centers — keeps growing and growing. But Wall Street is now becoming more discerning about which companies it wants to bet on during the next phase of what is being called the fourth industrial revolution. The insatiable need for capital is the bottleneck. Unlike the initial push to spread money around to bet on would-be AI winners, there is not enough to go around at this stage. The smaller cloud companies in recent weeks have been at the forefront of this rationalization push. CoreWeave has become the face of the "do we have too many neoclouds" debate. The stock has lost 60% from its year-to-date high of nearly $139 on May 6. It hit a 52-week low of just above $51 on Monday before moving higher with the broader market.
Make no mistake about it, some of the biggest companies in the world need more money, too. Club name Meta Platforms , as well as SpaceX and Anthropic, are dominating the headlines on Monday with plans to raise their share of the shrinking capital pie for the next legs of their AI ambitions. Meta announced plans to raise as much as $95 billion in three steps: first, an additional $22.5 billion worth of Class A common (what we and most investors currently own). The stock was $642.62 per share at the time of the filing. (On Tuesday, the offering was upsized to $25 billion and was priced at $625. Meta stock closed Monday at $630.56). The second step announced was plans for another $12.5 billion of preferred shares and, lastly, $60 billion no sooner than next year via an at-the-market offering for Class C shares. We learned about the plans to create the non-voting Class C shares last week as a vehicle to raise money in the future. Who knew the future was only a few short days away?
Elon Musk said that SpaceX has closed a $40 billion debt financing deal for graphics processing units (GPUs). We previously learned that SpaceX would look to exclusively build on Nvidia hardware. Remember, while SpaceX is a rocket and satellite company, it is also a massive AI infrastructure company. Anthropic dropped its initial public offering prospectus. The frontier AI lab, whose IPO valuation could be as high as $2 trillion, filed its so-called S-1 with the Securities and Exchange Commission. The share pricing and initial raise were not disclosed. They will come later. However, it is probably safe to say it will be for tens of billions of dollars, if not $100 billion, given the level of financial commitments the company has disclosed in the filing.
These three are just the latest in a mad rush for capital that has been putting a strain on the AI trade. The bulls argue that the demand means the compute is needed and the investments being made now will provide positive returns. The bears argue that there is a grave misallocation of capital underway, as the AI hype won't be met with AI profits down the line, and the compute demand everyone is predicting won't materialize. In a post on X confirming SpaceX's debt raise, Musk put those challenges in context . "As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn't have substantial revenue independent of AI data centers. ... That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many."
The neoclouds have a real problem. There is little doubt about that. The bigger question for us: Is that going to be a problem for everyone else? We don't think so, as Jim Cramer wrote in his Sunday column , saying he is not giving up on the AI trade, especially for the companies on the other side of Musk's "guppy" trade. We're talking about the ones that do more than AI. Last week, we argued that the key strategic advantage for the Meta Super Intelligence Lab (MSL) is that Meta Platforms, the broader company, has a wildly profitable core advertising business that gives MSL a structural advantage versus pure-play AI labs like Anthropic and OpenAI. A similar advantage can be found at the infrastructure layer, with hyperscalers Microsoft , Amazon , and Alphabet . Microsoft's other business is enterprise software found most on most personal and work computers. Amazon has e-commerce. Alphabet, like Meta, has a strong ad business. If many lower-quality neocloud players go under due to lack of funding, then it should read as a positive for the survivors long-term. (However, there is no getting around the near-term headline risk of any neocloud blowups. Investors will be selling first and asking questions later.)
The reason why is that demand won't evaporate if a neocloud goes under. The companies doing business with CoreWeave, Nebius , or Iren aren't after doing business specifically with those providers, as much as they are after the gigawatts of compute they have stood up at prices and services they like. Because the industry has standardized around a few players, like Nvidia and a few custom chipmakers, tied into the larger hyperscalers that are the ones soaking up all the excess capital. It won't be an issue for the customers to move their workloads. It's a simple equation: fewer suppliers, with equal demand, amounts to more share for each supplier.
Bottom line As is usually the case with choppy periods in any given industry, if trouble hits and players start getting their legs taken out from under them, it's those that have the financial durability to come out standing on the other side and reap the rewards. That brings us back to the biggest names, like Club stocks Microsoft, Alphabet, and Amazon, which are all set to report quarterly results this week. We expect funding to be the hot topic of conversation during their post-earnings conference calls. During Monday's Morning Meeting , Jim thought that Microsoft and Amazon would be feeling the most heat to raise funds following Meta's move Monday and Alphabet's $85 billion stock sale over the summer. UBS suggests that Amazon's 2027 capital expenditures plan could exceed $300 billion, even as the analysts raised their price target on the cloud and e-commerce giant and reiterated their buy rating. The UBS note gets at the heart of the issue and speaks to the conclusion of Jim's weekend column. Caution is warranted in the near term. We've certainly lightened up on our AI exposure in recent weeks, including an exit of Broadcom on Friday, due to funding uncertainty. Long term, however, you have to stick with the trade, pick your spots like the hyperscalers and Nvidia, and take advantage of weakness in the highest quality names, "because the revolution is real." The question, as Jim puts it: "How many will be left?"
