Are We in an AI Bubble?
📖 9 min read
Quick Answer
In 2026 the question is everywhere: is artificial intelligence a world-changing revolution, a financial bubble, or both at once? The honest answer is uncomfortable — the technology is genuinely transformative, yet the money flowing into it shows classic bubble warning signs. Here is the evidence on both sides, without the hype.
💡 Think of it as…
A gold rush where the people getting reliably rich are the ones selling shovels. The gold (useful AI) is real — but thousands are buying claims at any price on the faith that the seam is endless, while spending vastly more digging than they are pulling out.
The spending-vs-revenue gap
Big Tech plans roughly $650–700 billion in AI capital spending for 2026, against an estimated $51 billion in direct AI revenue — a 10-to-1 gap with no historical precedent. Worse, an MIT study (2025) found 95% of organizations were getting zero measurable return on their generative-AI investments. That gap is the single strongest bubble signal.
Stretched valuations
OpenAI has been valued around $730 billion at roughly 56× its annualized revenue, and the AI-heavy part of the S&P 500 reached its most stretched valuation since the dot-com era. When prices assume near-perfect execution for a decade, any stumble is punished hard.
The dot-com comparison
Morgan Stanley notes AI capex-to-sales is projected near 34% in 2026 (heading to 37% by 2028) — above the dot-com peak of ~32% in 2000. Circular financing is back too: a handful of giants (Nvidia, OpenAI, Microsoft and others) increasingly buy from and invest in each other, which can inflate the appearance of demand.
Why this is NOT exactly 1999
One crucial difference: today’s hyperscalers are funding the buildout mostly from operating cash flow, not debt. The capex-to-free-cash-flow ratio sits below 1, versus nearly 4× in 2000. Less leverage means a correction is less likely to trigger a cascading financial crisis — painful, but more contained.
Signs the air is leaking
The "DeepSeek shock" showed an efficiency breakthrough can erase hundreds of billions in value overnight. AI-attributed layoffs hit roughly 55,000 in the US in 2025. Watch for: collapsing capex guidance, a major model that fails to justify its cost, or enterprises cutting AI budgets after poor ROI.
🔑 Key takeaway
Both the bulls and bears are partly right: AI is real and transformative, but the spending-to-revenue gap, valuations, and circular deals are textbook bubble risks. It is probably not 1999 — but it carries specific dangers that could cause major disruption in 2026–2028.
What it means for you
Asian investors are heavily exposed through chipmakers, tech indices and AI tokens. Knowing the difference between a durable trend and a priced-for-perfection mania helps you avoid buying the top — and explains why some savers diversify into scarce, non-correlated assets like Bitcoin and gold.
Байнга асуудаг асуултууд
Will the AI bubble definitely pop?▼
No one knows the timing. Bubbles can inflate far longer than skeptics expect. What is clear is that valuations price in flawless growth, so the risk of a sharp correction is elevated even if AI ultimately succeeds.
If AI is a bubble, is the technology worthless?▼
Not at all. Like the internet after dot-com, AI can be both genuinely revolutionary and massively over-invested at once. The bust would hurt prices, not the underlying usefulness of the technology.
How do I protect my savings from an AI crash?▼
Diversify, avoid leverage, and do not concentrate everything in a single hot theme. Some investors hold uncorrelated stores of value; this is educational information, not financial advice.