Soft landing scenario for the AI bubble: the signals to watch
The essentials in 30 seconds
- Signals to watch in a soft landing are changes between two readings, and the direction matters more than the level.
- Communication at 0.305 is the first signal: a rise toward 0.5 shows the weakest dimension recovering (EQOS Asia AI study, April 2026).
- Failure signals track the 95% share of AI projects that miss production.
- Three families of signals count together: dimensions, failures, and funding.
- A monthly checklist with green, amber, and red readings keeps the review short.
Signals show whether the AI bubble is moving toward a soft landing. This page names the signals to watch, gives each a green reading, and sets a rule for when several of them confirm the case, as of October 2026. It extends the soft landing scenario score and uses the figures of the EQOS Asia AI study.
On this page
Compare the EQOS Asia offers and request a 30-minute pre-diagnostic.
What makes something a soft landing signal for the AI bubble?
An input is a value at one date, and a signal is the change between two dates. A soft landing is confirmed by direction: the weakest dimensions rise, failures fall, and funding follows earnings. A level can look poor and still carry a good signal, as long as the trend points the right way.
Three families of signals cover the case. Dimension signals track the nine market dimensions of the study. Failure signals track how many AI projects reach production. Funding signals track who repays AI spending. Each family gets a green reading below, and the rule at the end of the page combines them.
Which dimension signals show a soft landing taking shape?
Communication is the first signal, because it sits at 0.305, only 0.005 above the critical threshold of 0.3. A move toward 0.5 shows that explainability and transparency are improving. The second signal is Care at 0.338, which rises when retraining plans spread beyond the 15% of companies that have one today (EQOS Asia AI study, April 2026).
Trust at 0.450 and Clarity at 0.487 follow the first two. The study reasons that teams trust what they understand, so a rise in Communication tends to lift Trust in the next reading. Watch the sequence: Communication first, Trust next, and φ(S) last, moving from 0.337 toward the 0.401 that the 4 corrections produce.
Which failure signals show the AI bubble easing toward a soft landing?
The study reports that 95% of AI projects fail in production (MIT 2025, quoted by the study) and that 47% of enterprise users took a major decision on hallucinated content. A soft landing needs both shares to fall. Track them in your own portfolio of projects, because the market figure moves slowly and your figure moves first.
The reference target is the 82% predicted failure rate after the study’s 4 corrections, down from 94%. A project count that shifts from pilots to production is a green signal. A rising count of decisions reviewed by a person before action is a second green signal, since it cuts the contamination by hallucinations.
Does AI Act readiness count as a soft landing signal for the AI bubble?
The study lists AI Act compliance on 2 December 2027 among the headline facts of the enterprise AI market. Companies that publish a compliance plan, name an owner for each automated decision, and document the data behind it send a green signal. They also raise Clarity and Communication, which are inputs of the score.
The study also projects productivity gains of +40% and 300 million jobs that change. Treat both as projections. The green signal is a documented gain per use case in your own company, measured in hours or revenue. Documented gains show spending that finds paying buyers, which is the core of a soft landing.
Which funding and earnings signals support a soft landing?
The study records about $200B a year invested in AI research and deployment and valuations up ×4 in 18 months. A soft landing shows up when earnings begin to catch the spending. Green signals include customer payments covering a larger share of AI costs, revenue per user that you can document, and valuation growth that slows to the pace of earnings.
Funding from new investors alone is an amber signal, because it can sustain spending for a while without testing demand. Compare the two sources each quarter. If the customer-paid share rises for 2 reviews in a row, the funding family turns green, and the case gains weight in your scenario table.
How many signals confirm a soft landing in the AI bubble case?
Use a simple rule. Give each signal a colour: green for movement toward its target, amber for a flat reading, and red for movement toward a threshold. The soft landing gains confidence when at least 2 of the 3 families show green in 2 consecutive reviews. One green signal in one family stays an observation.
Write the rule before you look at the data. A rule fixed in advance stops a good week from feeling like a trend. When the count falls short, the case keeps its current weight, and an amber or red family triggers a rescore of the three scenarios.
How often should you check soft landing signals in the AI bubble?
A monthly checklist suits most teams: one page, nine dimension signals, three failure signals, three funding signals, each coloured. The EQOS Asia method page describes a monthly recompute through Sentinel and real-time alerts when a threshold is approached (EQOS Asia method page, October 2026). Those alerts fit the red readings.
Keep a dated log of each review. After 3 months the log shows the direction without further analysis, and after 4 quarters it supports a board discussion. Rescore the scenarios whenever a family changes colour, and record the reason in one sentence next to the date.
Where do you begin a personal watchlist for the AI bubble soft landing?
The Λ simulator gives you the starting reading for the 9 dimensions in about 10 minutes, free and anonymous (EQOS Asia study page, April 2026). Run it today, save the result, and run it again next month. The difference between the two runs is your first signal.
For a decision with money or headcount behind it, a free 30-minute pre-diagnostic with EQOS Asia sets up the checklist for your context. This page is general information, not investment advice, as of October 2026.
See also in this guide
Frequently asked questions
Which signal should I watch first for a soft landing?
Watch Communication first. It reads 0.305, only 0.005 above the critical threshold of 0.3, and it holds Trust and the overall score. A rise toward 0.5 shows that explainability and transparency are improving, which is the clearest early sign of a soft landing.
How many green signals confirm a soft landing?
A working rule asks for green in at least 2 of the 3 families, dimensions, failures, and funding, across 2 consecutive reviews. A single green signal stays an observation. Fix the rule before you read the data, so a good week cannot pass for a trend.
Are the study projections signals?
The projections of +40% productivity and 300 million jobs that change are market-level projections. The signal is a documented gain per use case in your own company, measured in hours or revenue. Your measured gain shows spending that finds paying buyers.
How often should I review the signals?
A monthly review suits most teams, with a dated log of each reading. Rescore the three scenarios whenever a family of signals changes colour. After 3 months the log shows the direction of travel without extra analysis.
Test your knowledge
1. What is a signal in this method?
- A price quote
- A forecast date
- A change between two readings
Show the answer
A change between two readings.
2. Which dimension is the first signal to watch?
- Communication
- Energy
- Stability
Show the answer
Communication.
3. How many families of signals does the page use?
- 2
- 3
- 9
Show the answer
3.
4. What predicted failure rate follows the study’s 4 corrections?
- 94%
- 95%
- 82%
Show the answer
82%.
5. What colour does funding from new investors alone receive?
- Green
- Amber
- Red
Show the answer
Amber.
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
Sources: EQOS Asia AI impact study (eqos-asia.com/case-studies/ia-2026.html), EQOS Asia method page (eqos-asia.com/method.html), simulator page (eqos-asia.com/simulator.html) and contact page (eqos-asia.com/contact.html), consulted on 22 October 2026. Page updated on 22 October 2026.