The headline figures were dominated by two large acquisitions, but the wider deal mix shows an enterprise agent stack taking shape across data, model access, governance and operational execution. Investors backed platforms that give AI agents context, permissions and operational control. Buyers pursued the same capabilities through acquisitions in data management, model access, governance and implementation services.
BARC’s Deal Radar focuses exclusively on transactions in our core coverage segments: enterprise data, analytics, business intelligence, and the governance and operation of enterprise AI. From the August edition onwards, every transaction included is directly relevant to these markets.
Two platform acquisitions account for most of the recorded M&A value: NVIDIA’s Hugging Face agreement and Stripe’s OpenRouter agreement. The OpenRouter price remains based on media reporting because the companies did not disclose terms. Databricks led a strong month for enterprise data and AI funding, while smaller rounds covered agent governance, software delivery and process execution.
August at a glance
- 35 transactions in Data & AI that we found interesting: 22 investments and 13 acquisitions.
- $7.29bn in disclosed funding across this core group, including $5bn from Databricks.
- Approximately $21.22bn in recorded acquisition value, including NVIDIA’s $12.93bn Hugging Face agreement and the reported $7bn value of Stripe’s OpenRouter agreement.
Compared with July
BARC’s core coverage contained almost as many transactions as in July.
Disclosed funding more than doubled. That increase was not only a Databricks effect: without its round, August still produced substantially more funding than July on the same basis.
Acquisition values were far higher than in July. NVIDIA’s $12.93bn agreement for Hugging Face carried the largest disclosed price. Stripe also agreed to acquire OpenRouter; Bloomberg put the price at more than $7bn, while the companies left the terms undisclosed. Both deals were pending completion.
A US market by count and capital
US-headquartered companies accounted for 21 of the 35 core-coverage transactions. Europe contributed 12: nine from the EU, two from the UK and one from Switzerland. Canada and Singapore contributed one each.
The capital split was sharper. US companies attracted approximately $6.75bn, or 93 per cent, of disclosed investment value across the core group. Europe attracted approximately $541m. Databricks explains much of that gap, but not all of it. Without its $5bn round, US companies still raised about $1.75bn, more than three times Europe’s disclosed total.
Europe’s larger rounds came from different parts of the market. Sweden’s Lovable raised $400m for AI software development. UK companies raised $122m across Callosum and Prevalent AI. Smaller rounds in Denmark, Germany, Finland and Bulgaria focused on agent workflows, enterprise knowledge, strategy execution and sovereign AI infrastructure.
Data and model access draw the largest deals
August’s largest funding rounds backed platforms that provide data, context and infrastructure for enterprise AI.
Databricks raised $5bn, while River AI raised $1.1bn for an open platform to train, adapt and deploy models. Callosum raised $100m to orchestrate AI workloads across models and heterogeneous computing architectures, with a focus on tailored inference. Stripe’s agreed acquisition of OpenRouter extends this logic into the commercial layer.
OpenRouter routes requests across hundreds of models and records their usage, while Stripe provides invoicing, tax, payment and fraud controls. The companies had already worked together on token billing. Connecting model selection and token consumption directly with pricing, billing and settlement could give Stripe a central role in token economics.
Bloomberg reported a value of more than $7bn. The combination also illustrates why cost control increasingly depends on model sizing, routing, caching and infrastructure governance, as outlined in this practical cost-control ladder for AI inference.
Other acquisitions targeted adjacent control points. NVIDIA agreed to acquire Hugging Face, a central platform for models, datasets and AI applications, for $12.93bn. Bending Spoons completed its $1.285bn acquisition of Airtable, adding collaborative enterprise data and workflows. AWS acquired DuckLabs, the Amsterdam-based company behind DuckDB, an open-source analytical database that brings fast SQL processing directly into applications, notebooks and local data workflows.
The move echoes SAP’s acquisition of Dremio: query execution, openness and flexible data access are becoming strategic control points for analytics and AI platforms. The structure of the DuckLabs transaction is important. The DuckDB Foundation retains the project’s intellectual property, the software remains open source under the MIT licence, and the team continues its work from Amsterdam.
This makes the deal as much about engineering talent and ecosystem scale as product ownership. The test is whether AWS can expand DuckDB’s reach without narrowing the openness and broad platform compatibility that made it successful.
Smaller transactions focused more directly on enterprise data foundations. Graphwise supplies semantic context through knowledge graphs, Gulp Data values proprietary data, and DryvIQ classifies and governs unstructured content. amber, Prevalent AI, Deepentix and fileAI turn enterprise information into usable context and workflows. DiscreteStack adds sovereign AI infrastructure for European deployments.
| Company | HQ Country | Value (US$) | Stage |
|---|---|---|---|
| Hugging Face | United States | $12.93bn | Acquisition agreement announced; completion not announced |
| OpenRouter | United States | >$7bn (reported; terms undisclosed) | Acquisition agreement confirmed; closing pending |
| Databricks | United States | $5bn | Strategic investment |
| Airtable | United States | $1.285bn | Acquired by Bending Spoons |
| River AI | United States | $1.1bn | Seed + Series A |
| Callosum | United Kingdom | $100m | Seed |
| Prevalent AI | United Kingdom | $22m | Growth round |
| amber | Germany | $8.12m | Series A |
| DiscreteStack | Bulgaria | $0.93m | Seed |
| Graphwise | United States | undisclosed | Majority acquisition by Oakley Capital agreed |
| Gulp Data | United States | undisclosed | Acquired by FirstParty |
| DuckLabs | Netherlands | undisclosed | Acquired by AWS |
| DryvIQ | United States | undisclosed | Acquired by Nasuni |
| Deepentix | Austria | undisclosed | Pre-seed |
| fileAI | Singapore | undisclosed | Strategic investment |
Governance and security move into the agent stack
July’s governance deals centred on inventories, compliance evidence and runtime policy. August extended that activity into the infrastructure used to operate, evaluate and secure enterprise AI agents.
Onyx Security raised $113m for discovery and policy enforcement across enterprise AI deployments. Sapiom combines agent routing and runtime infrastructure with permissions, budgets and audit trails. Naïve provides a broader operating stack with routing, budget controls and approval requirements for sensitive actions. Velatir governs AI usage and data flows across tools, employees and agents, while Oliver AI applies permissions, observability and budget controls to analytical data access.
A related group focuses on evaluation and security. Vals AI benchmarks models and AI applications before enterprises deploy them. Mindgard combines AI discovery, red-teaming and runtime protection. Anaconda’s acquisition of Enkrypt AI adds security and compliance controls for models, agents and MCP servers. These systems expose tools and data through the Model Context Protocol.
| Company | HQ Country | Value (US$) | Stage |
|---|---|---|---|
| Onyx Security | United States | $113m | Series B |
| Vals AI | United States | $40m | Series A |
| Sapiom | United States | $35m | Series A |
| Mindgard | United States | $30m | Series A |
| Naïve | United States | $28.5m | Series A |
| Velatir | Denmark | $5.8m | Seed |
| Enkrypt AI | United States | undisclosed | Acquired by Anaconda |
| Oliver AI | United States | undisclosed | Pre-seed |
AI development attracts growth capital
Investors backed tools that shorten the route from an idea to working software. The largest round went to software creation itself, while code review, testing and build infrastructure also attracted substantial capital.
Lovable raised a $400m Series C for its software creation platform. CodeRabbit raised $143m for AI code review and agentic change management. Blacksmith raised $45m for continuous integration infrastructure and agents that diagnose and resolve build failures.
The pattern reflects a direct consequence of AI-assisted development. Faster code generation increases demand for independent review, testing, build capacity and operational controls.
Cognition subsequently completed a $2bn Series E on 8 September at a $48bn valuation. Because the financing closed after the August reporting period, it is not counted here and will be included in the September Deal Radar.
| Company | HQ Country | Value (US$) | Stage |
|---|---|---|---|
| Lovable | Sweden | $400m | Series C |
| CodeRabbit | United States | $143m | Series C |
| Blacksmith | United States | $45m | Series B |
Agents move into operational execution
Investors backed platforms that put AI agents to work across complex enterprise processes. HappyRobot raised $150m to deploy agents across voice, email, documents, the web and existing enterprise systems. Skan AI raised $63m for a platform that captures how work is performed, turns that context into execution logic and uses it to operate AI agents. Palette raised $3.47m for an environment that integrates AI agents into workflows across sales, operations and other business functions.
| Company | HQ Country | Value (US$) | Stage |
|---|---|---|---|
| HappyRobot | United States | $150m | Series C |
| Skan AI | United States | $63m | Series C |
| Palette | Denmark | $3.47m | Pre-seed |
Strategy, performance and reporting draw capital
Strategy execution, performance management and reporting formed a distinct deal cluster. STRGY AI raised $1.15m for software that connects daily work with strategic priorities, identifies drift and automates management reporting.
Acquirers targeted established applications in the same field. ZenaTech bought ESM Software for strategy execution, performance management and compliance. osapiens bought the Nasdaq Metrio platform and customer base to expand non-financial reporting and carbon accounting in North America.
| Company | HQ Country | Value (US$) | Stage |
|---|---|---|---|
| STRGY AI | Finland | $1.15m | Angel funding |
| ESM Software | United States | undisclosed | Acquired by ZenaTech |
| Nasdaq Metrio | Canada | undisclosed | Acquired by osapiens |
Service providers consolidate around AI delivery
Three August transactions expanded specialist delivery capacity in data and AI. Dataciders acquired DATANOMIQ, adding process mining, agentic engineering and software delivery expertise. Swiss investment firm arcApex acquired a majority stake in IT-Logix to support organic growth and further acquisitions. Ode acquired Casper Studios, adding Claude implementation experience, reusable skills, connectors and context.
Ode was launched in July as a standalone company backed by Anthropic and a consortium of investors. It was built on the Fractional AI team acquired by the consortium-backed services venture in May, working alongside engineers from Anthropic.
July had already shown the same pattern. Alexander Thamm acquired Birds on Mars, adding responsible AI, AI experience design and organisational transformation expertise to its Data & AI services group. Birds on Mars continues to operate as an independent brand.
The transactions follow different ownership models, but their commercial logic is similar. Enterprise AI requires specialists who can connect models to data, applications and operating processes. Strategic buyers and investors are securing that delivery capacity rather than relying only on partnerships.
| Company | HQ Country | Value (US$) | Stage |
|---|---|---|---|
| DATANOMIQ | Germany | undisclosed | Acquired by Dataciders |
| IT-Logix | Switzerland | undisclosed | Majority stake acquired by arcApex |
| Casper Studios | United States | undisclosed | Acquired by Ode |
Who is backing and buying
The tracker identifies a US-headquartered lead investor or buyer in 18 of the 35 core-coverage transactions. This includes the lead parties behind the largest funding rounds for Databricks and River AI, as well as the buyers in the Hugging Face and OpenRouter agreements.
Corporate venture investors concentrated on companies adjacent to their technology, customer or operational priorities. NVIDIA and AMD Ventures participated in River AI’s round. Datadog and BMW i Ventures joined CodeRabbit’s financing. Dell Technologies Capital, Citi Ventures, State Farm Ventures and Wipro Ventures backed Skan AI. SMBC Asia Rising Fund and Singtel Innov8 invested in fileAI as it expands in Japan and regulated industries.
The acquisitions fall into four patterns:
- Platform owners targeted model, data and workflow infrastructure. NVIDIA’s Hugging Face agreement, Stripe’s OpenRouter agreement, Bending Spoons’ Airtable acquisition and AWS’s DuckLabs acquisition address different control points across this stack.
- Existing software vendors added governance and data context. Anaconda acquired Enkrypt AI, Nasuni acquired DryvIQ, and FirstParty acquired Gulp Data.
- Buyers expanded into strategy, performance and reporting workflows. ZenaTech added ESM Software to its enterprise SaaS division, while osapiens acquired Nasdaq Metrio.
- Service providers and investors secured delivery capacity. Dataciders acquired DATANOMIQ, Ode acquired Casper Studios, and arcApex acquired a majority stake in IT-Logix.
These patterns create different integration questions. Platform acquisitions can change control over an ecosystem. Services transactions change who owns specialist implementation capacity. Workflow acquisitions broaden the buyer’s application portfolio.
What we are watching next
- If completed, the Hugging Face and OpenRouter acquisitions would place a major model and dataset platform and a widely used model-routing gateway under large platform owners. Their integration choices will affect openness, neutrality and how enterprises obtain models and model access.
- Agent governance is moving closer to routing, permissions, spend and runtime action. Buyers will expect these controls to work across models, data platforms and business applications.
- Data platforms are competing to supply context for agents. Knowledge graphs, unstructured-data governance, model routing and operational databases are becoming parts of the same architectural decision.
- AI development funding now covers code generation, review, testing and build infrastructure. The market may consolidate when customers prefer an integrated delivery chain over separate tools.
- Consolidation in D&A services continues. The IT-Logix and DATANOMIQ transactions show that buyers and investors are still building scale through specialist delivery teams.