Agentic AI M&A: what is happening now?

What are all the recent M&A deals in agentic AI?

When we look at all the M&A deals in agentic AI over the last 24 months, we find a market that has already moved from scattered experiments to real acquisition activity.

Date	Target	Acquirer	Value	Strategic rationale	Status and additional comment
2026-06-24	Seldon AI	TrueFoundry	Undisclosed	Add enterprise ML and agentic workflow infrastructure	Closed / announced as acquired. Infrastructure deal focused on combining predictive ML, LLMs, and autonomous agents
2026-06-24	Aampe	MoEngage	Undisclosed	Add autonomous marketing and personalization agents	Closed / reported. MoEngage’s first acquisition, focused on autonomous AI agents for individual users
2026-06-15	Fin / Intercom	Salesforce	$3.6B	Strengthen Agentforce with a proven AI customer-service agent	Announced / pending. Largest disclosed agentic AI deal in the 24-month window
2026-05-28	StackAI	Asana	Undisclosed	Add no-code AI agent building and cross-system execution	Closed. Strong horizontal agent-builder deal for human-agent workflows
2026-04-29	Sligo AI	SpendHQ	Undisclosed	Bring configurable agentic AI into procurement workflows	Closed / announced as acquired. Clear vertical workflow-agent deal
2026-04-23	Fragment	Sierra	Undisclosed	Strengthen Sierra’s AI customer-service agent stack	Closed. Small tuck-in by a leading AI-agent startup
2026-04-23	Ryft	Cyera	Estimated $100M–$130M	Secure and govern data access for AI agents	Closed / announced as acquired. One of the few deals with an estimated value
2026-04-08	Simtheory	Canva	Undisclosed	Add agentic AI and data infrastructure for workflow execution	Closed / announced as acquired. Relevant, but less pure than CX or workflow-agent deals
2026-04-07	Sportsbox AI	Bryson DeChambeau-led group	Undisclosed	Build SAMI, an agentic AI sports coaching product	Closed / announced as acquired. Very vertical consumer/sports agent case
2026-04-02	Katanemo Labs	DigitalOcean	Undisclosed	Build agentic inference cloud and data-plane infrastructure	Closed / announced as acquired. Infrastructure-for-agents deal
2026-03-11	Forethought	Zendesk	Undisclosed	Add agentic customer-service capabilities	Announced / expected to close. Important CX agent consolidation deal
2026-03-10	Moltbook	Meta	Undisclosed	Explore social networks where AI agents interact	Closed / confirmed. Small but strategically interesting agentic-web experiment
2026-03-09	Promptfoo	OpenAI	Undisclosed	Secure AI agents through testing, red-teaming, and evaluation	Announced / pending. Agent security infrastructure, not an end-user agent
2026-02-25	Vercept	Anthropic	Undisclosed	Deepen computer-use and agent capability	Closed. Anthropic’s first known acquisition, clearly agent-related
2026-02-19	WorkFusion	UiPath	Undisclosed	Add financial-crime compliance agents	Closed. UiPath’s second agentic vertical acquisition in the dataset
2026-02-10	Tavily	Nebius	Undisclosed	Add real-time agentic search to AI cloud infrastructure	Announced. Infrastructure deal for developers building agents
2026-01-14	Included AI	Phenom	Undisclosed	Add agentic people analytics for HR teams	Closed / announced as acquired. Clear HR vertical agentic AI deal
2026-01-13	Modella AI	AstraZeneca	Undisclosed	Add generative and agentic AI to oncology R&D	Closed / announced as acquired. Vertical healthcare agentic AI
2025-12-29	Manus	Meta	Reported around $2B	Add general-purpose autonomous agents to Meta products	Reported acquisition, later blocked / unwound by Chinese authorities. Strategically major but status is unusual
2025-12-23	Wobby	HCLSoftware	Undisclosed	Add AI data analyst agents	Announced. Clear data-agent acquisition
2025-11-20	General Agents	Project Prometheus	Undisclosed	Acquire agentic-computing talent	Closed / reported. Acquihire-style deal for Bezos/Bajaj AI venture
2025-11-19	Sugarwork	Decidr	Undisclosed	Add knowledge capture for the agentic economy	Announced. Useful agentic-infrastructure signal, but not a pure agent product
2025-11-10	Velocity	BigPanda	Undisclosed	Accelerate agentic IT operations	Closed / announced as acquired. Strong IT operations agent deal
2025-11-04	Aisera	Automation Anywhere	Undisclosed	Add ITSM, HR, and customer-service self-service agents	Closed / announced as acquired. Major automation-agent consolidation move
2025-10-20	Metaculars	Skan AI	Undisclosed	Add enterprise operations agents	Closed / announced as acquired. Small but very relevant operations-agent deal
2025-10-16	Kodex AI	CUBE	Undisclosed	Build compliance, risk, and agentic AI platform	Closed / announced as acquired. Clear regulatory-compliance agent deal
2025-10-09	Kelley Austin	Perficient	Undisclosed	Add Salesforce AI, Agentforce, and data implementation capacity	Closed / announced as acquired. Services acquisition, useful signal but not core product M&A
2025-09-29	Qwiet AI	Harness	Undisclosed	Add agentic vulnerability detection	Closed / announced as acquired. Security-agent adjacent
2025-09-03	Solver / Laredo Labs	Nvidia	Undisclosed	Add AI coding-agent capability	Closed / reported. Coding-agent deal, based on reported acquisition
2025-09-03	OpenPipe	CoreWeave	Undisclosed	Add agent training and reinforcement fine-tuning	Closed / announced as acquired. Infrastructure layer for enterprise agents
2025-08-05	Regrello	Salesforce	Undisclosed	Turn business data into agentic workflows for Agentforce and Slack	Announced / expected close in Salesforce fiscal Q3 2026. Clear Agentforce tuck-in
2025-07-28	Cognigy	NICE	Undisclosed	Add conversational and agentic customer experience AI	Announced / pending. One of the clearest CX agentic AI deals
2025-07-23	Blueflame AI	Datasite	Undisclosed	Add agentic AI to dealmaking and investment workflows	Closed / announced as acquired. Strong vertical M&A workflow-agent deal
2025-07-14	Windsurf	Cognition	Undisclosed	Consolidate AI coding agents around Devin and Windsurf assets	Definitive agreement signed. Important coding-agent consolidation move
2025-07-08	Forwrd.ai	Pendo	Undisclosed	Add predictive analytics into product-experience workflows	Closed / announced as acquired. Adjacent enabler, not a pure agent deal
2025-06-25	Predibase	Rubrik	Undisclosed	Help enterprises build and customize AI agents	Intent announced. Infrastructure for AI-agent adoption
2025-06-10	unSurvey capabilities and talent	G2	Undisclosed	Add AI-led conversational reviews and research	Closed / capability acquisition. More capability acquisition than classic company M&A
2025-05-21	FullyAI	Circus SE	Undisclosed	Add agentic intelligence to robotics and food automation	Closed. Niche but explicitly agentic AI
2025-05-20	Numbers Station	Alation	Undisclosed	Add structured-data AI agents	Closed / announced as acquired. Data-agent infrastructure deal
2025-03-12	Peak	UiPath	Undisclosed	Add vertical agents for retail and CPG decisions	Closed / announced as acquired. Important vertical-agent move by UiPath
2025-03-10	Moveworks	ServiceNow	$2.85B	Extend agentic AI to employees across enterprise workflows	Closed. Largest closed disclosed deal in the dataset
2025-02-10	Agnostiq / Covalent	DataRobot	Undisclosed	Add distributed compute orchestration for agentic applications	Closed / announced as acquired. Infrastructure, not an end-user agent
2024-12-06	Frame AI	HubSpot	Undisclosed	Add conversation intelligence to CRM and AI workflows	Closed. Agentic rationale, but target is more AI data/intelligence than agent product
2024-09-03	Tenyx	Salesforce	Undisclosed	Strengthen customer-service voice agents	Closed. Early clear agentic customer-service tuck-in

Is agentic AI M&A really active now?

Agentic AI M&A is already active now, and the pace is too high to describe it as occasional experimentation.

When we look at all the M&A deals in agentic AI over the last 24 months, we count 44 transactions. The recent concentration is the key point: 35 deals happened in the last 12 months, compared with only 9 in the previous 12 months. That is almost a 3.9x increase in deal count.

The acceleration also shows up when we split the market into shorter periods. The first half of the 24-month window had only a few visible deals, such as Salesforce / Tenyx, HubSpot / Frame AI, DataRobot / Agnostiq, ServiceNow / Moveworks, and UiPath / Peak.

The second half becomes much denser, with Salesforce, Zendesk, NICE, UiPath, Automation Anywhere, Asana, OpenAI, Anthropic, Meta, CoreWeave, DigitalOcean, Nebius, and TrueFoundry all appearing as buyers.

The type of acquirer matters as much as the number of deals. These are not only AI-native startups buying other startups. Large enterprise platforms, automation platforms, cloud infrastructure companies, cybersecurity vendors, and frontier AI labs are all buying agentic capabilities. That breadth tells us the category is already relevant across several software markets, not just one isolated niche.

Did agentic AI M&A suddenly accelerate recently?

Agentic AI M&A accelerated sharply recently, and the change looks structural rather than random.

The cleanest comparison is the 12-month split: 35 deals in the latest 12 months versus 9 in the previous 12 months. But the acceleration is not limited to one hot month. Deal activity stays high across the recent period: 8 deals in Q3 2025, 9 in Q4 2025, 8 in Q1 2026, and 10 in Q2 2026. That is four consecutive active quarters.

The timing also shows a change in buyer confidence. In early 2025, the market was still mostly about selective deals like ServiceNow / Moveworks, UiPath / Peak, Alation / Numbers Station, and DataRobot / Agnostiq. By late 2025 and early 2026, we see buyers acquiring across many layers of the stack: customer-service agents, coding agents, IT operations agents, procurement agents, compliance agents, agent security, agent search, inference infrastructure, and data governance for agents.

There is another useful clue: several acquirers came back for more than one move. Salesforce appears three times, UiPath twice, and Meta twice. Repeat buying usually means the acquirer is building a roadmap, not reacting opportunistically to one interesting startup.

Are the big agentic AI deals already happening now?

Big agentic AI deals are already happening now, but public deal values are still rare.

The clearest large transaction is Salesforce announcing Fin / Intercom for $3.6B. ServiceNow / Moveworks is another major one at $2.85B, and it is especially important because it closed. Meta / Manus was reported around $2B, although its regulatory path later became unusual. Cyera / Ryft is much smaller than those platform deals, but its estimated $100M–$130M value is still meaningful for an infrastructure-layer acquisition.

The pattern is more interesting than the absolute numbers. The biggest priced deals are not in vague “AI assistant” categories. They are tied to high-volume enterprise workflows: customer service, employee service, and agent data security. Salesforce wants stronger customer-service agents inside Agentforce. ServiceNow wants AI agents inside employee workflows. Cyera wants to secure the data layer around enterprise agents.

Only 4 of 44 deals have disclosed or estimated values. That means we should not pretend we can calculate a full market value from disclosed M&A prices. Still, the few prices we can see are large enough to prove that agentic AI is already producing strategic transactions in the multi-billion-dollar range.

All things considered, the agentic AI M&A market is not only made of small acquihires. The visible price points show that when the target owns a strategic workflow or platform-relevant capability, buyers are willing to pay real money.

If you want more recent data on this point, please see our latest agentic AI market report.

Are most agentic AI acquisitions still small tuck-ins?

Most agentic AI acquisitions are still undisclosed, but the market is better described as opaque than small.

Out of 44 deals, 40 have undisclosed values. That is roughly 91% of the dataset. If we stopped there, it would be tempting to say the market is mostly small tuck-ins. But that would miss the more useful interpretation, because the disclosed and estimated transactions include $3.6B for Salesforce / Fin, $2.85B for ServiceNow / Moveworks, around $2B reported for Meta / Manus, and $100M–$130M estimated for Cyera / Ryft.

The deal types are also mixed. Some transactions clearly look like capability or team acquisitions, such as Anthropic / Vercept, G2 / unSurvey, Project Prometheus / General Agents, Sierra / Fragment, and Meta / Moltbook. Others look like serious platform acquisitions, especially Salesforce / Fin, ServiceNow / Moveworks, Zendesk / Forethought, NICE / Cognigy, Automation Anywhere / Aisera, and Asana / StackAI.

The market therefore has a barbell shape. At one end, large strategic buyers acquire product, customers, and distribution-ready capabilities. At the other end, buyers pick up technical primitives, small teams, or vertical agent products before they become expensive.

Finally, the hidden prices are themselves a clue. In young AI categories, companies often keep prices private because the transaction is partly about talent, partly about product, and partly about strategic positioning. So it looks like many small deals from the outside, but the strategic importance is much larger than the disclosed-price count suggests.

Is customer service the first real battleground in agentic AI M&A?

Customer service is the clearest first battleground in agentic AI M&A.

The concentration is hard to ignore. Salesforce bought Tenyx, then announced Regrello, then announced Fin / Intercom. Zendesk announced Forethought. NICE announced Cognigy. Automation Anywhere bought Aisera, which includes customer service alongside ITSM and HR. Sierra bought Fragment to strengthen its customer-service agent stack.

The logic is practical. Customer service has huge volumes of repetitive conversations, clear before-and-after metrics, existing knowledge bases, and obvious labor-cost pressure. That makes it easier for buyers to justify agentic AI than in less measurable workflows. A service agent can answer tickets, escalate issues, pull context from systems, and reduce human workload in a way that the CFO can understand.

There is also a distribution reason. Salesforce, Zendesk, NICE, Intercom, Sierra, and Automation Anywhere already sit close to customer interactions. If they do not own strong AI agents, another vendor can sit between them and the end customer. That is exactly the kind of platform risk that pushes M&A.

So yes, customer service became the first serious agentic AI acquisition battlefield because the ROI is easier to prove, the data is already there, and the incumbents cannot afford to lose control of the workflow.

If you want more recent data on this point, please see our latest agentic AI market report.

Are software platforms buying agents to defend their workflows now?

Software platforms are clearly buying agentic AI to defend their existing workflow positions.

The buyer list makes this obvious. Salesforce, ServiceNow, Zendesk, HubSpot, Asana, UiPath, Automation Anywhere, NICE, Harness, BigPanda, SpendHQ, Phenom, Datasite, and CUBE are all workflow owners in some way. They already control where employees, support teams, HR teams, security teams, procurement teams, or deal teams work. Agentic AI threatens that position because an agent can become the new interface for getting work done.

The deals fit the acquirers’ existing products closely. ServiceNow bought Moveworks for employee workflows. Salesforce bought Tenyx and announced Fin / Intercom and Regrello around service, Agentforce, and business processes. Asana bought StackAI to support human-agent teams across work tools. UiPath bought Peak and WorkFusion to add vertical agents to automation. BigPanda bought Velocity to strengthen agentic IT operations.

Here is the less obvious point: the acquirers are not just adding AI features. They are trying to prevent agents from becoming an independent layer above their platforms. If a third-party agent can execute workflows across Salesforce, ServiceNow, Zendesk, Asana, Slack, and Google Workspace, the workflow platform risks becoming plumbing.

Finally, these acquisitions show that agentic AI is becoming a control-layer battle. The buyer wants to own the place where the instruction starts, where the task is executed, and where the data comes back.

Is agentic AI M&A moving beyond chatbots these days?

Agentic AI M&A has already moved beyond chatbots into vertical workflow automation.

Customer service is still a major theme, but the broader list is much more varied. UiPath bought Peak for retail and CPG decision agents. UiPath also bought WorkFusion for financial-crime compliance agents. SpendHQ bought Sligo AI for procurement. CUBE bought Kodex AI for compliance and risk. Phenom bought Included AI for people analytics. AstraZeneca bought Modella AI for oncology R&D. BigPanda bought Velocity for agentic IT operations. Datasite bought Blueflame AI for dealmaking and investment workflows.

That variety matters because it tells us agents are being bought where they can perform a specific job, not just hold a better conversation. A procurement agent, compliance agent, financial-crime agent, or IT operations agent maps to a defined budget owner and a defined pain point. That makes the acquisition easier to justify than a generic assistant with unclear ROI.

The shift also changes how we should understand the market. “Agentic AI” is not one clean vertical. It is a capability layer being inserted into many vertical software markets at once. That explains why the acquirers come from so many categories.

Are AI coding agents consolidating already?

AI coding agents are already consolidating, but this is still early land-grabbing rather than mature consolidation.

The strongest signal is Cognition signing a definitive agreement to acquire Windsurf. That deal matters because Cognition’s Devin and Windsurf both sit in the AI coding-agent conversation. Nvidia’s reported acquisition of Solver / Laredo Labs adds another signal from the infrastructure side: a chip and AI platform company wants coding-agent capability too. Anthropic’s acquisition of Vercept is not only a coding-agent deal, but its computer-use and complex agentic-tooling angle belongs close to the same frontier.

What makes this interesting is the timing. AI coding agents are still developing fast, and the product experience is not settled. Buyers are moving before the category has clear long-term winners. That usually happens when the asset is scarce: strong engineering talent, agentic tool-use capability, and product learnings from real developer workflows.

Compared with customer-service agents, coding agents may have fewer M&A deals in the table, but each one says something important about strategic urgency. Coding is a workflow where agents can create visible productivity gains, and it is also where model labs, devtool companies, and infrastructure providers all have reasons to care.

If you want more recent data on this point, please see our latest agentic AI market report.

Is agent infrastructure becoming a real M&A category now?

Agent infrastructure is already a real M&A category, and it may be the most underappreciated part of the market.

We count 9 agent-infrastructure deals in the 24-month window. They cover different pieces of the stack: DataRobot / Agnostiq for compute orchestration, Alation / Numbers Station for structured-data agents, Rubrik / Predibase for agent customization, CoreWeave / OpenPipe for agent training, Nebius / Tavily for agentic search, OpenAI / Promptfoo for agent security, DigitalOcean / Katanemo for agentic inference cloud, Cyera / Ryft for data security, and TrueFoundry / Seldon AI for enterprise ML and agentic workflows.

The spread of these deals is the key clue. Buyers are not only acquiring front-end agent apps. They are buying the layers that make agents usable in real companies: search, evaluation, security, training, orchestration, inference, data access, and governance. Those are the boring problems that appear after the demo works but before enterprise deployment scales.

This also explains why infrastructure buyers are entering the dataset. CoreWeave, DigitalOcean, Nebius, Rubrik, Cyera, TrueFoundry, DataRobot, and Alation are not all trying to own the same end-user application. They want to own the rails that agent applications run on.

Finally, agent infrastructure M&A suggests that the agentic AI market is becoming more mature than the hype cycle implies. When buyers start acquiring tools for security, search, fine-tuning, inference, and data governance, it means they are preparing for real deployment, not just demos.

Are the buyers mostly big tech companies now?

Agentic AI M&A is not mostly a big-tech story, even though big tech is clearly present.

Meta, Nvidia, OpenAI, and Anthropic all appear in the deal list, so big tech and frontier AI labs are part of the market. But they do not dominate the table. The most active buyer is Salesforce, with three deals. UiPath and Meta each appear twice. The rest is spread across enterprise software, automation, customer experience, cloud infrastructure, cybersecurity, HR, procurement, healthcare, and data platforms.

That fragmentation is important. It tells us agentic AI is being absorbed by many software categories at the same time. In one market, the relevant acquirer might be Salesforce or Zendesk. In another, it might be UiPath, BigPanda, CUBE, SpendHQ, Phenom, Datasite, AstraZeneca, Cyera, or DigitalOcean.

The buyer diversity also changes the exit logic for startups. A strong agentic AI company does not need to be acquired only by a hyperscaler or frontier lab. If it owns a valuable workflow, the natural buyer may be the incumbent platform in that workflow.

So we can conclude that agentic AI M&A is broad, not centralized. Big tech is buying, but the deeper market signal is that many non-big-tech software platforms now see agents as strategically necessary.

If you want more recent data on this point, please see our latest agentic AI market report.

Are acquirers buying revenue, technology, or talent in agentic AI now?

Acquirers are buying revenue, technology, and talent, but the mix changes by deal type.

The large platform deals look more like product-and-revenue acquisitions. Salesforce / Fin and ServiceNow / Moveworks are not just talent grabs. They bring product credibility, customers, usage, and a clearer path to monetization. Moveworks reportedly had more than $100M ARR, and Salesforce / Fin came with a large customer-service footprint through Intercom.

Several smaller deals look more capability-led. Anthropic / Vercept, G2 / unSurvey, Project Prometheus / General Agents, Sierra / Fragment, and Meta / Moltbook seem closer to product, talent, or experimental capability acquisitions. Those deals still matter because agentic AI talent is scarce and product intuition is hard to recreate internally.

Infrastructure deals sit between those two categories. OpenAI / Promptfoo brings agent security and evaluation. CoreWeave / OpenPipe brings agent training. Nebius / Tavily brings agentic search. DigitalOcean / Katanemo brings inference and data-plane infrastructure. Cyera / Ryft brings security and governance for agent data access.

Finally, this mix tells us the market is moving past pure acquihiring. Talent is still important, but the stronger deals increasingly involve a product primitive, a customer base, or a strategic layer of the agent stack.

Is agentic AI M&A still mostly experimental?

Agentic AI M&A still contains experiments, but the center of gravity has moved into core product strategy.

The experimental side is visible. Meta / Moltbook is a small agent-social-network acquisition. Project Prometheus / General Agents looks like an agentic-computing talent move. Sportsbox AI is a niche agentic sports-coaching product. G2 / unSurvey is more of a capability acquisition than a classic platform deal.

But the serious side is larger. Salesforce is building around Agentforce. ServiceNow bought Moveworks to expand employee agents. Zendesk and NICE are buying into customer-service agents. UiPath is adding vertical agents. Asana bought StackAI for human-agent teams. OpenAI bought Promptfoo to secure agents. DigitalOcean, CoreWeave, Nebius, TrueFoundry, and Cyera are acquiring infrastructure for deployment.

The balance has changed because the same theme appears across product, infrastructure, security, and vertical workflows. If this were only experimentation, we would see mostly tiny talent deals. Instead, we see multi-billion-dollar transactions, repeat acquirers, enterprise platforms, and infrastructure buyers all moving in the same 12-month period.

All together, agentic AI M&A has crossed the line from experimentation into platform planning. Some deals are still exploratory, but the market itself is no longer experimental.

Is the agentic AI M&A market already consolidating?

Agentic AI M&A is consolidating early, before the category has fully matured.

That sounds strange, because many agentic AI companies are still young. But the deal pattern supports it. We see Salesforce buying multiple agentic capabilities. UiPath buying multiple vertical-agent capabilities. Meta buying or reportedly buying general-purpose and experimental agent assets. Zendesk, NICE, Automation Anywhere, ServiceNow, and Asana all moving into agentic AI through acquisition. That is a lot of platform positioning in a short period.

The consolidation is not traditional roll-up consolidation where one buyer buys dozens of similar companies. It is strategic gap-filling. Each buyer is trying to complete its own version of the agentic stack: customer service, employee workflows, vertical automation, coding, infrastructure, security, or data access.

This means the window for standalone agentic AI startups may compress quickly in certain categories. Customer-service agents, coding agents, and workflow agents are already attracting powerful buyers with distribution. Once an incumbent platform owns a good enough agent layer, smaller standalone companies may have to either specialize deeply or sell earlier.

At the end of the day, agentic AI consolidation is happening before the market feels mature. That is exactly what makes it important.

If you want more recent data on this point, please see our latest agentic AI market report.

So what is the latest update on M&A in agentic AI?

The latest update on agentic AI M&A is that the market has entered an early but real consolidation phase.

The biggest change is speed. In the last 12 months, we count 35 agentic AI M&A deals, versus 9 in the previous 12 months. That is a near-4x increase. But the more valuable insight is that the deals now cover the full stack: customer-service agents, employee workflow agents, vertical enterprise agents, coding agents, agent security, agent search, agent training, inference infrastructure, and data governance.

The buyer mix also matters. This is not only big tech buying AI startups. Salesforce, ServiceNow, Zendesk, NICE, UiPath, Automation Anywhere, Asana, OpenAI, Anthropic, Meta, CoreWeave, DigitalOcean, Nebius, Cyera, and TrueFoundry are all buying different pieces of the agentic AI market. That tells us agents are becoming a strategic layer across software, not just a new product category.

Finally, the market is still financially opaque. Most deal values are undisclosed, but the few visible prices are large enough to anchor the market: $3.6B for Salesforce / Fin, $2.85B for ServiceNow / Moveworks, around $2B reported for Meta / Manus, and $100M–$130M estimated for Cyera / Ryft. So we should not confuse limited disclosure with limited activity.

Check	Current status
Recent deal activity	Agentic AI M&A is active. We count 44 deals over 24 months, with 35 in the last 12 months
Market acceleration	Activity accelerated sharply. Deal count rose from 9 to 35 between the previous and latest 12-month periods
Large deals	Big deals are already visible. Salesforce / Fin reached $3.6B and ServiceNow / Moveworks reached $2.85B
Price transparency	The market is active but opaque. 40 of 44 deals have undisclosed values
Main buyer type	Enterprise software platforms are highly active. They are buying agents to protect workflows and distribution
Customer service	Customer service is the first clear battleground. Salesforce, Zendesk, NICE, Sierra, and Automation Anywhere all moved there
Vertical agents	Agentic AI is spreading into vertical workflows like procurement, compliance, HR, IT operations, oncology R&D, and financial crime
Coding agents	Coding-agent consolidation has started early. Cognition / Windsurf and Nvidia / Solver show strategic land-grabbing
Infrastructure	Agent infrastructure is now a real acquisition category. Search, security, training, inference, orchestration, and data governance all appear
Big tech role	Big tech is active but not dominant. The market is broader, with many enterprise and vertical software buyers involved
Deal logic	Buyers are not only acquihiring. Many deals add product primitives, enterprise customers, workflow depth, or infrastructure layers
Final status	Agentic AI M&A is consolidating early. Buyers are acquiring key capabilities now, before the category fully stabilizes

Who is the author of this content?

NEW MARKET PITCH TEAM

We track new markets so founders and investors can move faster

We build living "market pitch" documents for emerging markets: AI, synthetic biology, new proteins, and more. Instead of outdated PDFs or hallucinated LLM answers, our clients get a clean, visual, always-updated view of what's really happening: key players, deals, regulations, and signals that matter. Learn more about us.