Gartner: 80% of neocloud firms face extinction by 2030 as AI capex outpaces returns
A top Gartner forecaster just handed neocloud investors a brutal timeline, and the stocks are already cracking under the pressure. Whether CoreWeave and Nebius survive as independent companies may come down to one number almost none of them can hit.
CoreWeave (NASDAQ:CRWV) fell 23.85% in the week ended October 19 and closed at $61.12. Nebius Group (NASDAQ:NBIS | NBIS Price Prediction) dropped 18.96% over the same stretch.
The selling came as rate-hike fears hit the AI debt market. It also followed a forecast from John Lovelock, Gartner’s chief forecaster, that more than 80% of independent neocloud companies will stop operating independently by 2030.
Neoclouds are independent GPU cloud providers that rent out AI computing capacity. They compete with hyperscalers such as Microsoft (NASDAQ:MSFT) and Amazon (NASDAQ:AMZN). CoreWeave and Nebius are two of the best-known listed neoclouds, and both have grown revenue at triple-digit rates.
That growth costs a lot. CoreWeave lost $626 million in the second quarter on $2.58 billion in revenue, and Nebius spent $8.13 billion on capital expenditures in the first half of 2026.
Lovelock’s forecast asks whether today’s prices can earn that spending back in time. For most of the group, he concluded that they cannot.
An 18% Return Hurdle Few Neoclouds Can Clear
Return on invested capital is the profit a company earns on the money tied up in servers, chips, and buildings. Lovelock set a high bar for it in a recent television interview.
“The asset base that is being built up through this massive spending on AI servers requires a certain level of cash flow in order to meet a reasonable return on invested capital. We pegged it at 18%. And the revenue that we see being attracted to companies with a simple business model are insufficient until 2032.”
He added that compute capacity exceeds demand today and that much of the processing power in use is given away inside premium products. The revenue pool has a ceiling too. Total enterprise and consumer AI revenue is forecast at no more than $3.6 trillion through 2030.
GPUs Lose Value Before the Payoff Arrives
GPUs lose value as newer generations ship, but Lovelock expects demand to exceed supply only closer to 2030. A company that borrows to buy chips today has to pay interest the whole time.
CoreWeave’s interest expense rose to $640 million in the second quarter from $267 million a year earlier. It expects $860 million to $940 million in the third quarter. It also plans $35 billion to $39 billion of capex this year, so it will keep needing new financing.
Management says older chips hold their value. It points to an A100 contract running into 2029 on hardware first introduced in 2020.
Consolidation Can Still Cost Shareholders
Lovelock separated failure from absorption: “We are not saying that these companies are going out of business. We are saying that they may cease independent operations. They may become part of mergers and acquisitions and consolidations that we’re seeing as part of the current AI cycle.”
A buyout can come at a premium or at a distressed price. Lenders get paid before shareholders, so the balance sheet at the time of the deal decides which outcome you get.
Nebius carries about $10.0 billion of convertible debt and a $775 million secured facility backed by GPUs. It also faces $12.1 billion in future lease payments, and three customers each make up more than 10% of quarterly revenue.
CoreWeave’s debt-to-equity ratio is 8.94. Both companies are cloud partners of NVIDIA (NASDAQ:NVDA), which leaves the group dependent on one chip supplier.
What Could Prove the Forecast Wrong
Contracted backlog is the best defense against the forecast. CoreWeave reported $104 billion of backlog, and Nebius holds $37.5 billion in remaining performance obligations.
Prices are also rising. CoreWeave raised prices about 25% across its products in July, and Nebius’s first capacity auction cleared 15% above its previous high. Those price gains don’t fit Lovelock’s view that capacity is oversupplied today.
The forecast could still miss. Pricing power may hold, long-term contracts may cover debt payments, or demand may exceed supply before 2030.
Considering the Bull and Bear Cases for Neocloud Stocks
The bull case rests on contracted backlog, the chance of a buyout premium, and AI demand that eventually exceeds supply. The bear case is the 18% return hurdle, oversupply today, heavy debt and lease commitments, customer concentration, and GPUs that lose value before the payoff arrives. Lovelock compared this cycle to the 2008 and 2009 cloud expansion, when scale decided the winners and Microsoft and Amazon came out on top.
For now, the risks facing neocloud stocks as a group appear to outweigh the protection their backlogs offer, while Microsoft and Amazon look better positioned to take on the AI expansion because they can fund years of low returns and are the most likely buyers. Microsoft spent $115.9 billion on capex in its latest fiscal year, and Azure revenue exceeded $100 billion. Amazon’s AWS revenue grew 37% in its latest quarter.
Contact [email protected] for any questions or corrections.
Omor Ibne Ehsan is a writer at 24/7 Wall St. He is a self-taught investor with a focus on growth, cyclical, and dividend equities that have strong fundamentals, value, and long-term potential. He also has an interest in high-risk, high-reward investments such as penny stocks.
