Is the AI bubble bursting now? How to read the signs

What does it mean when the AI bubble is bursting now?

A burst is a lasting reversal, and a wobble is a short one. In a burst, falling prices reach most AI names, the fall holds for several quarters, and the money that financed the climb stops arriving. In a wobble, a few names drop, buyers return within weeks, and funding continues. The two look alike on the first day and differ by the fourth month. Time is therefore your best filter: a reading taken after one quarter beats a reading taken after one headline.

This difference matters for decisions. A wobble asks for patience, while a burst asks for a changed plan. Because both start the same way, the grid below waits for several independent signs before it calls a burst, and it asks for a second reading one quarter later.

Which price moves show the AI bubble is bursting now?

Look at the breadth of a fall. When only the weakest AI names drop, investors are sorting winners from losers, which healthy markets do all the time. When the strongest names fall with them, buyers are leaving the whole theme. Compare each group’s price with its level three months earlier and write both numbers down.

Watch what happens after a drop. A bounce that retakes the previous level signals confidence. A bounce that stalls below it, repeated twice, signals fatigue. The second pattern is rarer and more informative than any single day’s headline, which is why a monthly note beats a daily screen. Add the trading volume to your note: a fall on thin volume tells a different story from a fall on heavy volume.

Which funding signs show the AI bubble is bursting now?

Funding is the lifeblood of a bubble, so it gives the earliest warning. Delayed rounds, rounds closed at lower valuations, new listings pulled from the calendar, and lenders who tighten terms all mean that outside money is cooling. Each sign is public and dated, which makes it easy to log.

The EQOS Asia AI study records about $200B a year invested in AI research and deployment (EQOS Asia AI study, April 2026). A fall in that flow would reach start-ups first, since they hold the least cash. Count the start-up rounds announced each month and compare them with the same month a year earlier. Add the size of the largest round, since a few giant rounds can hide a thin field.

What do orders and earnings say about an AI bubble burst?

Earnings answer the question that prices cannot: are customers paying? Watch the orders reported by chip and cloud suppliers, the guidance they give for the next quarter, and the share of AI revenue in their totals. Cut orders from one large buyer carry more weight than cut targets from ten analysts.

On the demand side, the EQOS Asia study finds that 95% of AI projects fail in production, and that ROI not demonstrated accounts for 22% of the causes (MIT 2025 cited by the study). When more buyers report weak returns, budgets follow. Read their statements as a leading signal for next year’s orders, and note whether the reasons given are about cost, quality, or fit with the business.

What does the EQOS Asia score say about a burst now?

The study scores the enterprise AI market at φ(S) = 0.337, in the fragile zone where the system works and a shock can tip everything. No dimension sits in the critical zone below 0.3. Communication, which measures the explainability of deployed models, stands at 0.305, only 0.005 above that threshold (EQOS Asia AI study, April 2026).

Read this as a position, not a prediction. The market sits close to a line, and a public incident, such as a biased decision or a data leak, could push one dimension across it. The score tells you how much margin remains and which signs to watch first, while the day of any crossing stays unknown. A second reading each quarter shows whether the margin grows or shrinks.

How do you separate a normal drop from the AI bubble bursting now?

Use a counting rule. Score 6 signs as red or green each month: breadth of price falls, failed rebounds, delayed funding, cut orders, weak buyer returns, and a score near its threshold. Three or more red signs in one quarter call for a review. Three red signs in two consecutive quarters call for a changed plan.

The rule protects you from both errors. It keeps you from reacting to a loud week, and it keeps you from ignoring a slow slide. Record the date of each reading, because your record becomes the evidence you show a board or an investment committee when someone asks why you acted.

What should you do this month if signs of a burst appear?

Start with exposure. List what you hold or plan to buy in AI, and mark what you could carry through a 12-month fall. Pause any new commitment that depends on rising prices. Keep the rest under review. This page is general information, not investment advice, and your own situation decides the final choice.

To test your own case, the Λ simulator returns a φ(S) estimate in about 10 minutes. The EQOS Asia method page explains how the score is built. For decisions in Singapore, ASEAN, or Hong Kong, the pricing page lists the diagnostic formats, and you can book a 30-minute call to choose one.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile