Soft landing scenario: scoring the AI bubble case
The essentials in 30 seconds
- A soft landing is the AI bubble case where prices ease toward earnings and the market keeps working.
- The EQOS Asia study starts the enterprise AI market at a φ(S) score of 0.337 and sets 0.5 as the viability threshold.
- 4 corrective actions lift the score to 0.401 and cut the predicted failure rate from 94% to 82%.
- The score covers the condition of the market and the direction of travel, and a date stays out of it.
- Two child pages cover the inputs and the signals to watch.
A soft landing is the AI bubble case in which prices ease toward earnings and the market keeps working. This page covers the scoring of that case as of October 2026, using the EQOS Asia study figures, and shows what the score needs and what confirms it. It is one of 3 cases in the guide to AI bubble scenarios.
On this page
Compare the EQOS Asia offers and request a 30-minute pre-diagnostic.
What counts as a soft landing scenario for the AI bubble?
A soft landing is the branch of the AI bubble in which valuations ease toward earnings over time, spending finds paying buyers, and the market avoids a broad break. Prices can fall, and they can also stall while earnings catch up. What defines the case is that the market keeps working through the adjustment.
The EQOS Asia study gives the enterprise AI market a verdict of bifurcation, with a score of 0.337 in the fragile zone (EQOS Asia study, April 2026). A soft landing is the branch where the shocks that could tip the system are absorbed, which is why the case can be scored instead of guessed.
How does a decision engine score a soft landing for the AI bubble?
The score reads 3 things. The first is the distance of the weakest dimension from the critical threshold of 0.3. The second is the level of φ(S) against 0.5, the viability threshold that the study sets for a successful AI deployment. The third is the direction of travel between two readings. A soft landing scores high when all 3 point up.
Today’s reading starts from 0.337, so the market sits below 0.5 and above the floor. The case scores as reachable and not yet reached: it needs the weakest dimensions to improve while no shock lands. The number gives you a baseline for the quarterly comparison.
Which actions move the AI bubble toward a soft landing?
The study lists 4 actions that raise a φ(S) score. Measure the data signal before deploying, require explainability above 0.5 for critical decisions, audit the alignment between AI strategy and ground reality, and fund a retraining plan with 2% to 5% of total AI investment. Each action targets a weak dimension.
After the 4 corrections the study computes φ(S) = 0.401 and a predicted failure rate of 82%, down from 94%. That is the movement a soft landing needs at company level: more projects reach production, and the recoverable value is about $80B a year for the world market.
Which inputs does the soft landing score rely on?
The inputs are the 9 dimensions of the market, the funding picture, and your own exposure. For a soft landing the weight falls on the dimensions that can still rise, such as Communication at 0.305 and Care at 0.338. Each input needs a value, a source, and a date so that another reader can recompute the score.
The page on soft landing inputs lists each one, explains how to measure it, and shows how a change moves the result. Read it before you run the simulator on your own case, because a score is only as good as the data it receives.
Which signals confirm a soft landing in the AI bubble?
Signals tell you whether the market is moving toward the case. Three families matter: dimensions rising, such as Communication moving away from 0.3; failures falling, such as the 95% share of projects that miss production; and obligations met, such as readiness for the AI Act date of 2 December 2027.
The soft landing signals to watch page turns each family into a monthly checklist with green, amber, and red readings. A soft landing earns confidence when several signals turn green in the same quarter, and a single green reading is a start, not a verdict.
What does a soft landing score leave open about the AI bubble?
A soft landing still carries cost. After the 4 corrections the predicted failure rate is 82%, which means most complex deployments still struggle. The score says the system survives, and it leaves open whether every investor or company gains. Gains and losses depend on exposure, and exposure belongs to your own inputs.
Timing stays open too. The score cannot say whether the adjustment takes 1 quarter or 3 years, and it makes no claim on prices. It gives you the condition of the market and the direction of travel. Pair it with the correction and crash scores to see the full range.
What can a decision maker do with a soft landing score?
Use the score to decide which commitments to stage. In a soft landing, spending that is staged and tied to results keeps its value, while spending sized for the best case carries the most risk. Rank each commitment by how it performs here and in the other 2 scenarios.
Set a review date and a trigger. If the weakest dimension rises for 2 quarters, you release the next stage. If it falls back toward 0.3, you pause and rescore. The rule is written in advance, so the decision follows the data and the mood of the week stays out of it.
How can EQOS Asia score a soft landing for your own system?
The Λ simulator estimates φ(S) for your system in about 10 minutes, free and anonymous (EQOS Asia study page, April 2026). Rate the 9 dimensions, read the radar view, and compare the projection without action with the projection when you activate the top 3 levers.
For a full structural read, the diagnostic starts at US$2,288 on the EQOS Asia pricing page, and a free 30-minute pre-diagnostic opens the conversation (EQOS Asia pricing page, October 2026). This page is general information, not investment advice, as of October 2026.
See also in this guide
Frequently asked questions
What is the soft landing scenario for the AI bubble?
The soft landing scenario is the case where AI valuations ease toward earnings, spending finds paying buyers, and the market keeps working. A decision engine scores it from the nine dimensions of the market. As of October 2026, the case is reachable and not yet reached.
What score does a soft landing need?
The study sets 0.5 as the viability threshold for a successful AI deployment and starts the enterprise AI market at 0.337. A soft landing also needs every dimension to stay above the critical threshold of 0.3. Reaching both conditions at once is what the case asks for.
Does a soft landing mean AI projects succeed?
A soft landing means the system keeps working. The study still predicts a failure rate of 82% after the 4 corrections, down from 94%. Companies that act on the weak dimensions improve their own odds, and the score shows which dimensions to start with.
Can the soft landing score predict prices?
The score reads the condition of the market and the direction of travel, and it makes no claim on prices or dates. Use it to rank your options and to set review dates. This page is general information, not investment advice.
Test your knowledge
1. What does a soft landing keep doing?
- Working through the adjustment
- Raising prices every quarter
- Stopping all AI spending
Show the answer
Working through the adjustment.
2. Which threshold does the study set for a successful AI deployment?
- 0.3
- 0.5
- 0.9
Show the answer
0.5.
3. What failure rate does the study predict after the 4 corrections?
- 94%
- 95%
- 82%
Show the answer
82%.
4. Which dimension is the weakest in the study?
- Energy
- Communication
- Clarity
Show the answer
Communication.
5. What does a single green signal give you?
- A start, not a verdict
- A final verdict
- A crash date
Show the answer
A start, not a verdict.
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 pricing page (eqos-asia.com/pricing.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.