AI: Bubble, or the Boom You Don’t Fight

Download Article

AI today is both the most overhyped and the most consequential story in markets. It’s an industry built on staggering optimism — and for now, that optimism is the fundamental. Betting against it may be more dangerous than being wrong with it.

The Oracle–OpenAI Reflex Loop

Oracle’s stock is a perfect case study in narrative-driven valuation.

After announcing roughly US$300 billion in future infrastructure contracts with OpenAI — deals yet to be paid for or fully executed — the stock surged over 30 % and briefly pushed Larry Ellison toward the world’s richest-man title. Oracle’s “remaining performance obligations” ballooned to US$455 billion, most of which represent future commitments rather than booked revenue.

In parallel, Nvidia has pledged up to US$100 billion to OpenAI — its own major customer — to fund compute infrastructure that will ultimately purchase Nvidia’s chips. OpenAI itself signed up to US$22 billion in contracts with CoreWeave, which is simultaneously a supplier and a financing partner.

The structure is almost recursive: everyone finances everyone else’s demand, creating a closed circuit of optimism where capital begets valuation, valuation begets capital, and the story reinforces itself.

Meta’s Course Correction

Even the giants are recalibrating. Meta cut ~600 roles in its AI division this month, pruning non-core research teams while doubling down on its “Superintelligence” project. It’s less a retreat than a sign of maturity — a shift from open-ended experimentation to commercial focus. AI remains central to Meta’s identity, even as it trims excess ambition.

When Optimism Becomes Momentum

It’s easy to call this a bubble. But unlike crypto or meme stocks, AI’s infrastructure is tangible: servers, data centres, silicon, software, and talent. These assets exist — they just may be early, expensive, or overbuilt. That doesn’t make the trade fake; it makes it reflexive.

Capital is rushing in faster than monetisation can catch up. Yet if that monetisation arrives — through enterprise AI integration, consumer adoption, or productivity gains — the upside is still enormous. The cycle feeds on belief, but belief is what finances every technological revolution before it’s proven.

The External Fault Lines

The risks are still real:

Concentration: A handful of firms dominate compute, models, and distribution. A misstep in any could echo through the entire stack.

Funding Tightness: Higher rates or investor fatigue could slow expansion before revenue catches up.

Regulation: AI safety, privacy, and energy-usage rules could increase cost and delay scaling.

Energy & Capacity: Data-centre build-outs are running into power-grid limits in the U.S. and EU.

Technological Disruption: A new, cheaper compute architecture could suddenly obsolete trillions in sunk cost.

Any one of these could deflate expectations temporarily — but they may only delay, not derail, the trajectory.

Protege9's View: Cautious Bullishness

This is a sector built on overwhelming optimism — and it’s dangerous to go against that momentum. While the financing looks circular, the ambition is real. The gap between hype and delivery may be wide, but capital is the bridge being built in real time.

It’s not a clean story of fundamentals yet, but markets rarely wait for clean stories. Betting on rationality has been a losing trade in every early-stage technological boom. The smarter posture, for now, is cautious participation — to ride the wave without believing it’s infinite.

Bottom Line

AI may still be over-capitalised and over-narrated, but it’s also under-realised. The infrastructure being built today could either prove overdone or turn out to be the scaffolding of a generational shift. Either way, ignoring it would be the bigger mistake.