The Rising Cost of AI Is Turning the Entire Tech Sector Into a Pressure Cooker
Building the future ain’t cheap — and investors are finally tallying up the bill. The race to dominate artificial intelligence is burning through cash faster than companies can prove the payoff, and patience is running out. If the infrastructure won’t generate returns soon, companies will need to dig up a better answer for where all this money is actually going.
Borrowing today, paying forever: Long-term debt loads are climbing, and credit default swaps are hitting record highs — a clear sign that Wall Street is losing patience with AI spending that lacks a near-term payoff. Even giants like AmazonAMZN, MicrosoftMSFT, and GoogleGOOGL are leaning on profitable cloud units to defend their AI bets, especially with multi-year compute contracts that may never fully pay for themselves. Oracle’sORCL stock dropping 11% after boosting its data-center budget to $50B was just the loudest warning shot.
- Oracle co-CEO Clay Magouyrk says cloud contracts will “quickly add revenue and margin to our infrastructure business,” arguing that demand from other clients can absorb unused OpenAI capacity.
- Still, Jefferies analyst Brent Thill warned that Oracle’s software cushion can’t fix the “timing mismatch between upfront capex and delayed monetisation,” which keeps near-term pressure high.
When AI Finally Pays the Rent
Not every legacy tech name is straining under the AI bill. CiscoCSCO just reclaimed its dot-com peak after more than 25 years, powered by demand for AI infrastructure. The networking player’s return to a record last seen in Mar. 2000 shows that AI spending can pay off — but mainly for businesses built to profit from it, not just finance it at any cost.
- Cisco’s latest rally followed a stronger revenue forecast projecting up to $61B this fiscal year, ~$1B above prior guidance, as it positions itself to benefit from AI infrastructure demand.
- SLC Management’s Dec Mullarkey called Cisco’s comeback “a quaint reminder that a recovery from a bubble can take a long time,” comparing it to Japan’s market taking decades to rebound.
Shifting gears: While Oracle fights to justify its AI bill and Cisco enjoys a rare win, MetaMETA is moving away from open-source models toward closed systems it can actually monetize. Wall Street pushed back on CEO Mark Zuckerberg’s plan to keep spending heavily into 2026, but sentiment warmed up when reports hinted at metaverse cuts. Meta says AI already strengthens its ad business, but its massive infrastructure bill may take years to pay off. And after Llama 4’s disappointing debut, Zuckerberg is now personally driving a new model — Avocado — set to launch next spring as a closed, revenue-ready product. It’s another sign that the AI boom is forcing every tech giant to rethink what “scale” really means.