(SeaPRwire) – By: Reginald Vance
The NASDAQ just posted its worst session since June. Nvidia dropped 4.4 percent. CoreWeave sank 10 percent. AMD fell 8.1 percent. This wasn’t a routine pullback. Something shifted in how capital is priced for the AI infrastructure buildout. Jensen Huang said it bluntly on CNBC. He told reporters that demand for capital in AI is now far greater than the available supply. That one line changed the entire conversation.
Let me lay out what actually happened on Thursday. The S&P 500 fell 1.2 percent. The Nasdaq Composite slid 2.6 percent. The tech sector as a whole dropped about 3 percent. Nvidia, Alphabet, and AMD all sold off meaningfully. Meta shed nearly 7 percent after announcing a 2-for-1 stock split alongside quarterly results that the market did not embrace. CoreWeave collapsed 10 percent as the cost of capital hit a private neocloud provider with a heavy debt load. Tesla was volatile after an earnings beat, with investors parsing the shift from lower 2026 capital spending to higher 2027 and 2028 spend on Optimus, Cybercab, and Terafab. SpaceX gave back more than half of Wednesday’s post-earnings gain. Meanwhile, the 10-year Treasury yield fell 18.5 basis points to 4.961 percent as money fled equities for bonds. Blackstone, oddly, climbed 8 percent on upbeat investment sales guidance despite weaker third-quarter earnings from elevated borrowing costs.
The supply chain reality here is becoming impossible to ignore. AI training clusters require thousands of GPUs per data center. Each GPU demands memory, interconnects, power delivery, and cooling infrastructure. The capital stack to fund this has gotten expensive fast. Ed Yardeni called the AI trade a bubble on Thursday, though he noted it would not burst catastrophically. That is actually the more interesting take. This is not a liquidity crisis. It is a solvency anxiety at the edges. CoreWeave carries a heavy debt load. Private AI infrastructure companies are burning cash to build data centers that will not generate revenue for quarters. The public market is now pricing in the possibility that some of these bets do not land well. Tesla’s capex pivot tells the same story. Companies are stretching forward into 2027 and 2028 because they have no choice. The hardware lead times alone force that timeline.
The endgame here is consolidation. The vendors who control silicon, packaging, and foundry capacity will extract more margin. The neocloud providers with balance sheets intact will acquire the overleveraged ones. AMD’s steeper decline relative to Nvidia reflects the market’s judgment that the second-tier chip play is more exposed to this capital crunch. The question is no longer whether AI demand is real. It is which companies can survive the period between the capital raise and the revenue curve. The market just reset its expectations on that gap.
Author bio: Reginald Vance is a venture partner specializing in semiconductor valuation and advanced materials, with two decades tracking the intersection of hardware supply chains and capital markets.