AI coding startup Cognition is in early discussions with investors for a new funding round that could value the company at around $40 billion, according to Bloomberg. The potential deal would represent a sharp increase from Cognition’s most recent valuation and underline the extraordinary investor appetite for companies building AI agents capable of performing software-engineering work. The talks are still at an early stage, meaning the valuation, size of the round and other terms could change or the discussions could ultimately fall apart.
Cognition is best known for Devin, its autonomous AI software engineer, which is designed to handle software-development tasks from planning and coding to testing and validation. The company has rapidly expanded beyond the initial AI coding-agent market, including through its acquisition of Windsurf, while striking partnerships with large IT services companies such as Infosys. Its latest reported funding round, completed in May 2026, raised $1 billion at a $26 billion valuation, according to Rothschild & Co.’s growth-equity report.
Cognition Targets $40 Billion Valuation
The potential $40 billion valuation would represent a major jump for Cognition in only a few months.
In May, Cognition raised $1 billion in a funding round led by Lux Capital and General Catalyst, with other investors participating. The company was valued at about $26 billion at the time, according to a June growth-equity report from Rothschild & Co.
If the new discussions result in a deal at $40 billion, Cognition’s implied valuation would rise by roughly 54% from its May level.
| Cognition Valuation Milestones | Valuation |
|---|---|
| September 2025 funding | About $10.2 billion |
| May 2026 funding | About $26 billion |
| New reported funding discussions | About $40 billion |
| Increase from May valuation | About 54% |
The latest $40 billion figure should be treated as a potential valuation, not a completed transaction. Bloomberg’s report says discussions are in the early stages.
Devin Is at the Center of Cognition’s Growth
Cognition’s primary product is Devin, which the company describes as an AI software engineer.
Unlike traditional coding assistants that primarily suggest snippets of code inside an editor, Devin is designed to work on broader software-development assignments. It can take a task, reason through the problem, write or modify code, run tests and work through development environments.
That difference is central to Cognition’s business strategy.
The company is betting that software development will increasingly move from human developers writing code with AI assistance to AI agents completing entire engineering tasks under human supervision.
Traditional Coding Assistant vs Devin
| Capability | Traditional AI Coding Assistant | Devin |
|---|---|---|
| Code suggestions | Yes | Yes |
| Code generation | Yes | Yes |
| Understands larger projects | Increasingly | Yes |
| Runs development tools | Limited/varies | Yes |
| Handles multi-step tasks | Limited | Core focus |
| Autonomous execution | Limited | Core capability |
| Testing and validation | Supported by some tools | Designed into workflow |
| Human role | Writes and reviews | Delegates, supervises and reviews |
The distinction is increasingly important as companies experiment with AI agents that can take responsibility for complete software-development workflows.
Cognition’s Revenue Has Expanded Rapidly
The company’s valuation discussions are also being supported by rapid growth in its commercial business.
A June report from Rothschild & Co. said Cognition had reached an annualized revenue run rate of approximately $492 million, with Devin driving the growth. The same report noted that the company had raised $1 billion at a $26 billion valuation in its latest round.
That means the company is being valued at a very large multiple of its current annualized revenue.
Using the reported $492 million revenue run rate, a $40 billion valuation would represent approximately 81 times annualized revenue.
| Metric | Reported Figure |
|---|---|
| Latest funding | $1 billion |
| Latest valuation | $26 billion |
| Potential new valuation | $40 billion |
| Annualized revenue run rate | ~$492 million |
| Potential valuation / revenue run rate | ~81x |
| Increase from $26B to $40B | ~54% |
The comparison is not equivalent to a conventional public-market price-to-sales ratio because Cognition is a rapidly growing private AI company and the $40 billion figure is only under discussion. Nevertheless, it illustrates the premium investors are placing on fast-growing AI-agent companies.
Why Investors Are Betting on AI Coding Agents
Software engineering has emerged as one of the most commercially promising applications of generative AI.
Coding has several characteristics that make it attractive for AI automation.
Software exists in a structured environment. Code can be tested, compiled and evaluated. Developers can give agents access to repositories, issue trackers, terminals and testing systems.
That provides AI systems with feedback that is harder to obtain in many other knowledge-work categories.
An AI agent can make a change, run tests, observe the result and attempt another solution.
This creates a feedback loop:
Task → AI writes code → Tests run → Results analyzed → Code revised → Tests repeated → Human review
The more reliable that loop becomes, the more software-development work can potentially be delegated to AI.
The AI Coding Market Is Becoming Highly Competitive
Cognition is not operating alone.
The AI coding market now includes major AI companies, specialized startups and established developer-tool companies.
Microsoft-backed GitHub has Copilot, Cursor has built a major AI-native coding environment, Anthropic has expanded Claude’s coding capabilities and OpenAI has developed Codex for software engineering.
The competition is increasingly moving beyond autocomplete.
Major AI Coding Approaches
| Company/Product | Core Focus |
|---|---|
| Cognition / Devin | Autonomous AI software engineer |
| Cursor | AI-native coding environment and agents |
| GitHub Copilot | AI coding assistance and agents |
| OpenAI Codex | AI software engineering agents |
| Anthropic Claude Code | Agentic coding through Claude |
| Windsurf | AI coding environment |
Cognition’s strategy is distinctive because the company is positioning Devin as an autonomous software engineer rather than simply an assistant inside a developer’s editor.
Windsurf Acquisition Expanded Cognition’s Ambitions
Cognition has also expanded its position in AI coding through its acquisition of Windsurf.
The deal followed a highly publicized period in which major AI companies competed for Windsurf’s technology and talent. Cognition ultimately acquired the remaining Windsurf business after Google recruited key members of its leadership and research team.
The acquisition gives Cognition another important asset: an AI coding environment used directly by developers.
That creates an opportunity to combine two approaches:
Devin’s autonomous agent
Windsurf’s developer environment
=
A broader AI software-development platform
This could help Cognition compete not only as an AI agent provider but also as a developer-tools company.
Enterprise Adoption Is Becoming Critical
The next phase of Cognition’s growth is likely to depend heavily on enterprise adoption.
Large companies have thousands or even tens of thousands of developers working across legacy systems, internal applications and customer-facing software.
If AI agents can handle even a portion of routine engineering work, the potential productivity gains could be significant.
Cognition has already moved into enterprise distribution through partnerships.
In India, Infosys and Cognition have partnered to deploy Devin across Infosys’ internal engineering teams and global client engagements, with the goal of accelerating software delivery and integrating Devin into Infosys’ AI ecosystem.
This type of partnership could give Cognition access to large numbers of enterprise software-development projects without having to sell directly to every individual company.
AI Agents Could Change the Economics of Software Development
The traditional software-development model is heavily dependent on human engineering hours.
Companies hire developers, divide work into projects and assign teams to design, build, test and maintain software.
AI agents introduce a different possibility.
A company could potentially have a smaller number of senior engineers supervising a much larger number of AI agents.
Traditional Model
Product manager
↓
Engineering manager
↓
Developers
↓
Code → Testing → Review → Deployment
AI-Agent Model
Product manager
↓
Senior engineers / AI supervisors
↓
Multiple AI agents
↓
Coding + Testing + Debugging + Documentation
↓
Human review
The second model does not necessarily eliminate software engineers. Instead, it could change what engineers spend their time doing.
Developers could increasingly focus on architecture, product decisions, system design, reviewing AI-generated work and handling complex problems.
Cognition’s Valuation Shows How Fast AI Startup Values Are Moving
Cognition’s potential move from a $26 billion valuation to $40 billion would be another example of how quickly private AI company valuations can change.
The company has already experienced a dramatic valuation increase over a relatively short period.
Its September 2025 valuation was reported at about $10.2 billion, while the May 2026 funding round placed it at approximately $26 billion. The latest discussions could push that figure to $40 billion.
Cognition Valuation Growth
2025
$10.2B
↓
May 2026
$26B
↓
August 2026 talks
~$40B
From the September 2025 valuation to the potential $40 billion valuation, Cognition’s implied value would have increased by almost 292%.
That kind of appreciation would have been difficult to imagine for a software startup before the current AI investment cycle.
Why a $40 Billion Valuation Is Significant
A $40 billion private valuation would put Cognition among the world’s most valuable AI startups.
It would also send a broader signal to venture investors: autonomous software engineering is being viewed as a potentially enormous market rather than simply a feature within existing coding tools.
Investors are effectively betting that AI agents can capture a meaningful portion of the economic value generated by software development.
That value could come from:
- Reducing software-development costs
- Increasing developer productivity
- Accelerating product launches
- Automating software maintenance
- Improving testing
- Handling legacy-code modernization
- Supporting enterprise application development
- Expanding the amount of software companies can build
The Biggest Challenge Is Reliability
The biggest question surrounding Cognition and Devin is not whether AI can write code.
Modern AI models can already generate substantial amounts of software.
The harder problem is whether an autonomous agent can consistently understand complex requirements and produce reliable software without creating additional work for human developers.
A coding agent that generates incorrect code may save time during the initial development stage but create larger costs during debugging and review.
That makes metrics such as task completion rates, code quality, production reliability and human intervention increasingly important.
For enterprise customers, security is another major concern. Giving an AI agent access to proprietary repositories, internal systems and development infrastructure requires strict permissions and monitoring.
AI Coding Could Also Create a New Bottleneck
There is a paradox at the heart of autonomous coding.
If AI agents become extremely effective at generating software, the amount of code requiring human review could increase dramatically.
That means the bottleneck could move from writing code to reviewing and validating code.
This is already influencing the broader developer-tool market.
Cursor, for example, is building Origin around code hosting and AI-oriented code review, while other companies are developing automated review agents.
The emerging software-development stack could therefore look like:
AI planning → AI coding → AI testing → AI code review → Human approval → Deployment
The companies that control several parts of that workflow could gain an important competitive advantage.
What Cognition’s New Funding Could Finance
If Cognition completes another major financing round, the additional capital could support several areas of expansion.
Model and Agent Development
The company can continue improving Devin’s reasoning, coding and autonomous task-completion capabilities.
Computing Infrastructure
Running AI coding agents at scale requires significant computing resources, particularly when agents perform long-running tasks.
Enterprise Sales
Cognition can expand its sales and customer-success organizations to target large companies.
Windsurf Integration
The company can integrate Windsurf’s developer environment with Devin’s autonomous capabilities.
International Expansion
Partnerships such as the one with Infosys could help Cognition reach software-development markets outside the United States.
The AI Startup Funding Race Is Intensifying
Cognition’s potential $40 billion valuation comes during a broader surge in AI startup financing.
AI companies are attracting enormous amounts of capital because investors believe the technology could reshape software, search, enterprise applications and knowledge work.
The scale of recent financing has also created a significant gap between leading AI startups and traditional venture-backed software companies.
Cognition’s latest $1 billion financing and potential new round demonstrate how investors are concentrating capital around companies that show both technological differentiation and commercial traction.
The question for investors is increasingly whether these companies can grow quickly enough to justify valuations that already anticipate enormous future markets.
What Happens Next for Cognition
The immediate next step is whether the reported funding discussions develop into a formal transaction.
Because the talks are still preliminary, Cognition could raise less money, negotiate a different valuation or decide not to proceed.
If a deal is completed at around $40 billion, it would provide the company with substantial additional resources at a critical moment for the AI coding market.
Cognition would then face pressure to demonstrate that Devin can convert its technological capabilities into durable enterprise revenue and long-term customer retention.
Looking Ahead
Cognition’s reported discussions around a potential $40 billion valuation highlight just how aggressively investors are betting on AI agents that can perform software-engineering work. The potential valuation would represent a roughly 54% increase from the company’s reported $26 billion valuation in May, when it raised $1 billion. Cognition’s Devin has become one of the most prominent examples of the industry’s shift from AI coding assistants toward autonomous software engineers, while its Windsurf acquisition and enterprise partnerships are expanding the company’s reach across the software-development lifecycle.
The bigger question is whether the economics can eventually support the valuation. AI coding agents have the potential to reduce development costs, accelerate software delivery and allow smaller teams to produce more software, but they must also solve challenges around reliability, security, code review and human oversight. If Cognition can turn Devin’s rapid revenue growth into a durable enterprise platform, a $40 billion valuation could look like another step in the expansion of AI-native software development. If adoption or reliability falls short, however, the company’s valuation would face the same scrutiny now confronting the broader AI startup market.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
