AI and data software providers are turning to acquisitions to ramp up their agentic AI capabilities

Glaring gaps in vendors’ product portfolios have become increasingly apparent; operators’ growing demands for products that include agentic AI are not yet being addressed. Many vendors are on the hunt for acquisition targets in an effort to fill this gap as quickly as possible. This trend is already leading to a marked uptick in the number of deals captured in Analysys Mason’s AI and data platforms M&A tracker.

This article analyses the acquired assets and the value they bring to the acquiring companies. It also shows how partnerships offer a viable alternative solution for addressing the full range of agentic AI requirements.

Agentic AI was responsible for the surging number of acquisitions by AI and data software providers in 1H 2025

An analysis of the data in the latest iteration of Analysys Mason’s AI and data platforms M&A tracker shows the increase in M&A activity on behalf of AI and data software providers that serve the telecoms market. The number of deals made during the first half of 2025 was the same as that during the full calendar year in 2024, and we expect many more deals in the second half of the year.

Agentic AI is the key driver for many of these acquisitions. Almost 70% of the deals announced in 1H 2025 were linked to agentic AI, and the majority of these were acquisitions of companies with agentic AI expertise. For example:

• IBM acquired AI agent builder, Seek AI, to support the development of IBM’s watsonx AI Labs in New York
• Salesforce acquired Convergence.ai, which specialises in building AI agents to handle complex, human-like tasks in digital environments
• ServiceNow acquired Moveworks’s front-end AI agents and enterprise search capabilities to extend the reach of its agentic AI platform to more users.

Other acquisitions announced in 1H 2025 targeted capabilities that enable AI and data platforms to meet the demands of agentic AI workflows, including the speed, scale, security and orchestration requirements that are essential to AI agent operations.

• DataRobot acquired Agnostiq to accelerate the development and deployment of agentic AI applications on DataRobot’s AI platform. Agnostiq has advanced compute orchestration and optimisation capabilities that ensures that AI agents are developed and deployed to multiple infrastructure environments and can be managed effectively using common tools.

• Databricks acquired Postgres database specialist, Neon, to meet the speed requirements for provisioning databases that are automatically created by AI agents.

AI and data software providers used acquisitions to tackle data challenges that could hinder agentic AI adoption

Agentic AI is generating lots of attention across the telecoms industry, but it will not be extensively adopted until several data-related barriers are addressed. The inability to access, process and analyse unstructured data assets effectively to ensure that AI agents can use accurate insights from data to perform their autonomous functions is one of the challenges that vendors aim to address using M&A. Hallucination also needs to be addressed, given the significant adverse effect that it can have on agentic AI workflows, especially those that handle critical business workflows or direct interactions with customers.

Vendors are using acquisitions to gain data-related capabilities including data lifecycle management for unstructured and structured data assets, data retrieval, labelling and synthetic data generation. Figure 1 provides a summary of the data-related acquisitions made in 1H 2025 and the value that they can deliver to the acquirer.

Figure 1: Examples of acquisitions by AI and data software companies in 1H 2025

Source: Analysys Mason

AI and data platform providers can also pursue partnerships to address the gaps in their portfolios

M&A is effective, but not the only option for vendors who have recognised the urgent need to enhance their existing agentic AI proposition. Partnerships are a relatively low-risk solution to filling gaps in existing product suites and require limited up-front investment. Identifying and negotiating a partnership can be a slow, careful process, but solutions can be implemented relatively quickly, and often entail less extensive integration efforts.

Several vendors that are targeting the AI and data software market for the telecoms domain have taken the partnership route to address the requirements for agentic AI. For example, Ericsson partnered with AWS to use AWS’s BedRock agentic AI framework to develop and manage its AI agents, while Amdocs and Prodapt have partnered with NVIDIA to develop telecoms-focused AI agents using NVIDIA’s AI platform.

Acquisition and partnership are not mutually exclusive: vendors can combine both strategies to fast-track their agentic AI ambitions. For example, ServiceNow is setting up partnerships to support its agentic AI proposition in addition to acquiring assets. Its new data ecosystem, the Workflow Data Network (WDN), is a group of partners that provide data platforms, applications and enterprise tools. These partnerships ensure that ServiceNow’s AI agents, or AI agents developed within its platform, can connect to, understand and take action within data sources that exist in the WDN data ecosystem.

This article provides an analysis of data in our AI and data platforms M&A tracker. For further information on this tracker, please get in touch with Adaora Okeleke.