loading
loading
“When you are building a new type of technology company, there will be periods of time when the market misunderstands and acts irrationally.” — Mike Intrator, CEO, CoreWeave “Sometimes the technology will be ahead of what the market will accept, and sometimes vice versa, and we have to be cognizant of what the market will accept at a given time.” — Ted Pick, CEO, Morgan Stanley “Yikes” — Elon Musk It’s pretty clear that AI is now as much a financing story as it is a technology story. While operating cash flow at the largest tech companies more than covered capital investment through 2025, that ceased to be the case this year. Meta’s admission that it will “need to raise hundreds of billions of dollars [of equity] over a period of several years beginning as early as fall 2026” brings the scale of that shift into focus. It’s a shift Elon Musk understands well. His IPO of SpaceX in June was perfectly timed to lock in funding before the competition. By raising $85.7 billion of equity, he secured a vast war chest while strengthening the company’s balance sheet enough to obtain investment-grade credit ratings and tap the bond market for a further $25 billion. Few operators have instincts as sharp across both financial market and technology domains as Musk. He has made access to capital a competitive advantage, cultivating loyal investors and understanding how to put the machinery of markets to work for his companies. Others in the industry are less skilled – among them, CoreWeave. We’ve talked about the company here before. Founded in 2017 by CEO Mike Intrator, Chief Strategy Officer Brian Venturo and Chief Development Officer Brannin McBee, it was initially set up to mine Ethereum in Venturo’s grandfather’s garage but, following the collapse of crypto in 2019, it swiftly pivoted to GPU cloud services. Today, it operates 51 data centers across the US and Europe, drawing 1.5 gigawatts of power which it leases out to customers like Microsoft, OpenAI and Meta. When we looked at the company last November, it had secured contracts to deliver 2.9 gigawatts of computing power. Expansion since has pushed that number to 4.2 gigawatts. While the company positions itself as The Essential Cloud for AI™, its function really is as a conduit to capital markets. Growing a footprint from 1.5 gigawatts to 4.2 gigawatts takes cash. CoreWeave expects to spend between $35 billion and $39 billion in 2026 alone towards getting there. To access that kind of capital, it needs to borrow. If it can lease new GPU deployments at more than the cost of that capital, the model works; if the spread compresses, the model unwinds. The problem is that CoreWeave’s cost of capital is rising. Yields on its unsecured debt have hit 15% and its stock is falling. CoreWeave is, in effect, a non-bank financial. And as anyone who has followed non-bank financials through more than one cycle knows, relying on the kindness of markets works only for as long as they stay kind. It’s no coincidence that CoreWeave’s travails preceded those of the broader AI complex. Its spread conveys a lot of information about the state of the market. To explore the economics of the AI boom through the lens of CoreWeave, and to see where the barometer is currently pointing, read on.
document fetched 2026-10-24 02:57 · canonical https://netinterest.co/p/capital-wars