The numbers are stated judgments. The record began on 13 October 2026 with starting confidences that the system set from the previous three weeks of reports, and a published rule has made every change since. They are not measured probabilities. The question and the rule are fixed before the record begins, and each change shows its evidence and its reasons. The rule and what is not claimed are on the method page.
Larger rivals, among them the largest US AI and cloud companies and the French model maker Mistral AI, offer equivalent private and sovereign deployments and win the customers Cohere targets, so by the end of 2027 Cohere's growth stalls or a larger company acquires it.
32.5% stated confidence, Unlikely (20–45%). Starting point on the record’s first day: Unlikely (20–45%), 32.5 per cent. Attention level: For information. Trend: Steady. Activation: Dormant.
The judgment as stated before the record began
This is the leading alternative to the main assessment. Cohere sells the promise that buyers need not choose between capable AI and control over it, and larger rivals now make the same promise. Google, Microsoft and Amazon Web Services have each begun offering AI that runs inside a customer's own data centre or in a cloud kept within Europe, SAP and OpenAI have announced a service for the German public sector, and the French company Mistral AI, newly funded in September 2026, makes the same pitch to European governments. If buyers can have the most capable models under their own control, Cohere's place narrows, its growth stalls and a larger company may buy it. Evidence for this would appear first as contracts won by rivals in Cohere's markets, delays at Cohere and departures of its leaders.
How the record started this assessment
On 13 October 2026, before the record began, the system read the 3 developments below and chose Unlikely (20–45%). The record started the assessment at the middle of that range, 32.5 per cent. The reading is kept in the first day's record.
Its reason, machine-written by anthropic/claude-sonnet-4.6: Microsoft's sovereign AI offering supports C1 by showing a larger rival entering Cohere's space, and the ESA-Mistral letter of intent supports C2 by showing Mistral winning a government-adjacent contract. However, there is no evidence of contracts lost by Cohere, internal strain, or acquisition interest, and the PwC alliance and North 2 launch from the main assessment show Cohere actively winning enterprise business. The competitive threat is real but not yet sufficient to suggest Cohere's growth stalls by 2027.
- 1. First reported 13 October 2026, 1 outlet.“SBS Launches Banking’s First Sovereign AI Platform” · einpresswire.com · 13 October 2026 · Open the report
- 2. First reported 5 October 2026, 1 outlet. Cited by the reading.“Sovereign AI: Accelerate innovation with control and choice” · microsoft.com · 5 October 2026 · Open the report
- 3. First reported 25 September 2026, 1 outlet. Cited by the reading.“ESA and Mistral sign letter of intent on 'sovereign AI' for space” · mlllm.io · 25 September 2026 · Open the report
Causes of uncertainty
- A slowdown at Cohere could come from the work of absorbing Aleph Alpha rather than from rivals, and the two are hard to tell apart from outside.
- Rival offerings are often announced long before they are available or in use, so an announcement is weaker evidence than a contract.
- Cohere's North platform can run other companies' models, so a customer may choose a rival's model and still pay Cohere.
What would change the judgment
- Several government or large-organization contracts in Cohere's markets going to larger rivals that offer private deployment.
- Cohere's funding round or its merger with Aleph Alpha being delayed, reduced or abandoned.
- Reports that a larger company is in talks to buy Cohere.
The causal factors, and which way each one cuts
| Factor | Which way it cuts | What evidence looks like | Last moved |
|---|---|---|---|
| C1. Private and sovereign offerings from larger rivals Offerings from larger rivals, among them the largest US AI and cloud companies and the French model maker Mistral AI, that let large organizations and governments run capable AI under their own control. Each one narrows what sets Cohere apart. | Supports. A stronger factor supports the assessment. |
| Not yet |
| C2. Contracts won by rivals in Cohere's markets Government and large-organization contracts for private or sovereign AI that go to rivals in the markets Cohere targets, among them Canada, Germany, the United Kingdom, Japan and South Korea. | Supports. A stronger factor supports the assessment. |
| Not yet |
| C3. Strain at Cohere Signs that Cohere is under pressure, such as job cuts, departures of senior leaders, missed revenue targets, and delays in its funding round or in its merger with Aleph Alpha. | Supports. A stronger factor supports the assessment. |
| Not yet |
| C4. Interest in buying Cohere Reports that a larger company is considering buying Cohere, or that Cohere is open to a sale, which this scenario expects once growth stalls. | Supports. A stronger factor supports the assessment. |
| Not yet |
| C5. Cohere chosen over larger rivals Evidence that buyers choose Cohere even where larger rivals offer private deployment, such as contracts won after a competition, renewals and expansions by large customers, and new customers that name control as their reason. More of it weakens this scenario. | Weakens. A stronger factor weakens the assessment. |
| Not yet |
What the assessment assumes, and what would break it
- CA1. When a larger rival offers private deployment of a more capable model, large buyers choose the rival over Cohere.Linchpin.
Invalidated if Cohere's annual recurring revenue (the yearly value of its ongoing contracts) at the end of 2026, as stated by Cohere or reported by Reuters, Bloomberg, the Financial Times, The Globe and Mail, The Logic or BetaKit from its statements or investor documents, is at least twice the figure of about US$240 million reported for the end of 2025. Larger rivals already offered private deployment of capable models during 2026, among them Google's Gemini in a customer's own data centre and large models in Microsoft's disconnected private cloud, so growth of that size would show buyers choosing Cohere regardless. If no figure for the end of 2026 is published by 30 June 2027, the assumption is reviewed rather than invalidated.
Source: The Globe and Mail, Cohere beats forecast with $240-million in annual recurring revenue, 13 February 2026, Google Cloud, Gemini is now available anywhere, 28 August 2025, Microsoft, Microsoft Sovereign Cloud adds governance, productivity and support for large AI models securely running even when completely disconnected, 24 February 2026. The criterion’s fingerprint was fixed on 13 October 2026 and is part of the record’s first day. - CA2. Cohere needs new capital to keep pace with larger rivals, and it has not yet secured it.Bridge.
Invalidated if, by 31 March 2027, Cohere announces in its newsroom or a press release that its current funding round, known as its Series E, has closed, raising at least US$2 billion.
Source: Cohere newsroom, The Globe and Mail, report on Cohere's talks to raise up to US$3 billion, September 2026. The criterion’s fingerprint was fixed on 13 October 2026 and is part of the record’s first day. - CA3. The sovereign offerings of the largest AI companies reach the public-sector markets that Cohere is entering, starting with Germany.Bridge.
Invalidated if, by 30 June 2027, SAP has not announced on its news site that OpenAI for Germany, the service for the German public sector that SAP and OpenAI announced on 24 September 2025, is available to public-sector customers.
Source: SAP News Center, announcement of the partnership between SAP and OpenAI for a sovereign OpenAI service in Germany, September 2025. The criterion’s fingerprint was fixed on 13 October 2026 and is part of the record’s first day.
The rival explanations
Evidence moves the assessment only when it bears differently on the assessment than on at least one of these. Evidence that fits them equally carries no weight.
- CX1. Cohere slows because of its merger rather than because of rivals: absorbing Aleph Alpha takes its attention for a time, and growth returns once that work is done.No development was recorded on this assessment in the last 7 days.Favoured by:
- Cohere attributes slower growth to the merger.
- Customers stay while new contracts pause.
Ruled out by:- Customers move to rivals during the merger.
- CX2. Cohere's platform outlasts its models: customers run other companies' models on North, so Cohere keeps growing even if its own models fall behind.No development was recorded on this assessment in the last 7 days.Favoured by:
- Customers announce that they run rival models on North.
- Cohere reports growth in its platform while use of its own models flattens.
Ruled out by:- Customers leave North when rivals offer private deployment of their own.
- CX3. Cohere survives as a regional champion with public support: it keeps growing in Canada and Germany through government backing and contracts while losing ground elsewhere.No development was recorded on this assessment in the last 7 days.Favoured by:
- Most new contracts are with Canadian and German public bodies.
- Few new customers in the United States or Asia.
Ruled out by:- Cohere wins large contracts outside Canada and Germany.
A development is recorded when the rule applies it to the assessment or sets it aside. Each count comes from the readings sealed with the changes, and a development set aside as fitting every alternative equally tells no alternative apart.
Evidence behind changes in the last 30 days
Each item is a report or data series cited by a sealed change on this assessment. The update loop reads headlines and summaries, and no summary is shown as the evidence.
No evidence has been cited in the last 30 days.
Evidence considered and set aside
Developments the update loop read in the last 30 days and set aside because they fitted this assessment and each of its alternatives equally. Under the rule such evidence carries no weight, so none of it moved anything.
Nothing on this assessment was set aside in the last 30 days.
Changes to this assessment
No change to this assessment yet.
The rest of the judgment
The question’s text on this page is as sealed on the record’s first day. Fingerprint . Fingerprint pending. Records are fingerprinted each day at 23:59 UTC.