What are the top agentic AI startups by total funding raised?

Top startups in the Agentic AI market ranked by total funding raised

Here is an updated table that ranks the top startups in the Agentic AI market based on the total amount of funding they have raised to date.

The table also includes the total number of funding rounds, the date and size of the latest round, the financing type (e.g. Series A, equity financing), key investors, the startup’s current status (active, IPO, acquired, or shut down), and a confidence score based on the data collected (we excluded startups with very low data confidence, to make sure everything is reliable).

If you need to dig deeper and get more detailed data, please check our report covering the Agentic AI market.

Which Agentic AI startups have raised the most money?

Cognition is currently the best-funded Agentic AI startup in our ranking, with more than $3.9 billion raised, putting it far ahead of every other company in the market.

Harvey comes next at more than $1.8 billion, while Poolside and Sierra are both around $1.6 billion. Mercor has reached roughly $1.1 billion and Magic about $1 billion. Lovable, Legora, Replit and Wonderful complete the top 10, and every company in that group has raised more than $800 million.

The gap at the top is unusually large. Cognition alone has raised roughly twice as much as Harvey and more than Poolside and Sierra individually. Its latest Series E added more than $2 billion in a single financing, after the company had already raised another $1 billion only a few months earlier.

There is also no single Agentic AI business model dominating the top of the table. Cognition, Poolside and Magic are centered on software development. Harvey and Legora focus on legal work. Sierra targets customer service. Mercor sits between human expertise, AI training and enterprise AI deployment. Wonderful is building a broader enterprise AI platform.

The common thread is the size of the work being targeted. Investors are backing companies that could take over meaningful chunks of expensive professional work, not just companies adding another AI feature to existing software.

Rank	Startup	Main focus	Total raised
1	Cognition	Autonomous software engineering	$3.9B+
2	Harvey	Legal AI	$1.8B+
3	Poolside	AI models for software development	$1.6B
4	Sierra	Customer-service AI agents	$1.6B
5	Mercor	AI talent and expert platform	$1.1B
6	Magic	Autonomous coding systems	$1.0B
7	Lovable	AI application building	$953M
8	Legora	Legal AI	$866M
9	Replit	AI application development	$852M
10	Wonderful	Enterprise AI operating system	$834M

How much money have Agentic AI startups raised altogether?

The 99 Agentic AI companies in our dataset have raised roughly $25.3 billion in cumulative funding.

That total is already large for a market in which many of the best-known companies were founded only during the recent generative-AI wave. More interestingly, most of the money is sitting near the top.

The first five companies account for roughly $10 billion. The top 10 have raised about $14.5 billion, or around 57% of all the capital represented in our dataset. Move down to the first 20 companies and the amount reaches roughly $18.9 billion, close to three quarters of the total.

The distribution gets even more extreme farther down. By the time we reach the 50th startup, around $23.5 billion has already been accounted for. The lower 49 companies therefore represent only about $1.8 billion combined.

So the $25.3 billion headline can be misleading if we picture that money being spread across a broad field. Most Agentic AI startups are operating with tens of millions of dollars or less. A small group is operating with hundreds of millions or billions.

Is Agentic AI funding dominated by just a few startups?

Yes. Agentic AI funding is heavily concentrated today, with roughly 57% of the capital in our dataset going to the top 10 companies and about 75% going to the top 20.

That concentration creates two very different markets under the same Agentic AI label.

At one end, Cognition, Harvey, Sierra, Poolside and several other companies can spend hundreds of millions on engineers, model training, infrastructure, sales teams and acquisitions. At the other end, dozens of startups are working with less than $50 million of total financing.

The gap is even clearer when we compare individual companies. Cognition's latest round alone was larger than the cumulative funding of every startup below Harvey in the ranking. Sierra's recent $950 million financing was larger than the lifetime funding of all but six companies in the dataset.

This helps explain why the market can feel crowded and concentrated at the same time. There are plenty of Agentic AI startups, but only a small number are being financed as future global platforms.

Group	Combined funding	Share of dataset
Top 5	~$10.0B	~40%
Top 10	~$14.5B	~57%
Top 20	~$18.9B	~75%
Top 30	~$21.3B	~84%
Top 50	~$23.5B	~93%
Companies 51–99	~$1.8B	~7%

Why has Cognition raised so much more money than other Agentic AI startups?

Cognition has pulled away because investors are currently treating autonomous software engineering as one of the biggest commercial opportunities in Agentic AI.

The company has now raised more than $3.9 billion. Its latest Series E alone brought in more than $2 billion at a $48 billion valuation, led by Andreessen Horowitz and Accel. That followed a $1 billion financing at a $26 billion valuation only a few months earlier.

The fundraising acceleration came with a sharp increase in the company's reported business activity. Cognition said annualized run-rate revenue moved from about $492 million around its previous financing to nearly $900 million by the latest round.

Run-rate revenue should be treated carefully because it extrapolates recent performance rather than reporting a completed full year. Even with that caveat, almost $900 million is an unusually large number for a company founded in 2024.

Investors are betting that Devin and similar systems can absorb a meaningful share of global software-engineering work. Software development is expensive, digital from beginning to end and spread across almost every industry. A company that captures even a small share of that spending can become very large.

Cognition's funding lead reflects how aggressively investors are pricing the possibility that autonomous software engineering becomes a major software category of its own.

Which Agentic AI startups have already raised at least $1 billion?

Six Agentic AI startups in our dataset have reached roughly $1 billion or more in cumulative funding: Cognition, Harvey, Poolside, Sierra, Mercor and Magic.

Cognition is in a completely different tier at more than $3.9 billion. Harvey follows above $1.8 billion, while Poolside and Sierra are around $1.6 billion each. Mercor has crossed approximately $1.1 billion and Magic sits around the $1 billion mark.

Several companies are close behind. Lovable has reached approximately $953 million, Legora $866 million, Replit $852 million and Wonderful $834 million.

The billion-dollar club could get bigger quickly if those companies keep raising at anything close to their recent pace.

The composition of this group is also revealing. Three of the six billion-dollar companies are closely tied to software creation: Cognition, Poolside and Magic. Harvey automates professional legal work. Sierra goes after customer operations. Mercor has increasingly built around the human expertise needed to train, evaluate and deploy AI systems.

The largest pools of capital are gathering around expensive knowledge-work categories rather than lightweight consumer agents.

What are the biggest recent Agentic AI funding rounds?

Cognition's more than $2 billion Series E is the largest recent Agentic AI financing in our ranking, followed by Poolside at roughly $1 billion and Sierra at $950 million.

The next tier is already enormous by normal software standards. Harvey raised $550 million in its latest financing. Wonderful also raised $550 million. Mercor raised roughly the same amount in our dataset, while Lovable and Replit each raised $400 million.

Below them, EliseAI raised $300 million, Decagon $250 million, Rogo $160 million and Factory $150 million in its Series C before adding another large growth financing.

The striking part is the number of separate companies receiving checks this large. We are seeing repeated rounds of $250 million, $400 million, $550 million and more across coding, legal, customer service, finance and broader enterprise AI.

Large investors are clearly willing to finance several competing Agentic AI companies simultaneously.

Startup	Recent major round	Amount
Cognition	Series E	$2.0B+
Poolside	Strategic / Series C financing	~$1.0B
Sierra	Growth financing	$950M
Harvey	Growth financing	$550M
Wonderful	Series C	$550M
Mercor	Large growth round	~$550M
Lovable	Series C	$400M
Replit	Financing round	$400M
EliseAI	Series E Prime	$300M
Decagon	Series D	$250M
Rogo	Series D	$160M
Factory	Series C	$150M

Are $500 million Agentic AI funding rounds becoming normal?

No, a $500 million round is still exceptional, but it has become much less unusual at the top of Agentic AI.

Harvey, Wonderful and several other companies have recently raised around or above that level. Sierra went much further with $950 million. Cognition crossed $2 billion in one financing.

The speed is what makes this period unusual. Wonderful raised $150 million in a Series B and then $550 million in its Series C within the same year. Harvey announced $200 million at an $11 billion valuation and followed it with another $550 million financing at $15.5 billion. Cognition went from a $1 billion round to more than $2 billion within months.

Those financing cycles would normally be associated with companies much farther into their development.

There is still a huge gap between the leaders and everyone else. Most Agentic AI startups will never raise anything close to $500 million. Among companies that investors believe could dominate a large professional workflow, though, rounds in the hundreds of millions are no longer hard to find.

Are coding agents getting more funding than any other type of Agentic AI?

Yes. Software development is currently the strongest funding cluster in Agentic AI by a substantial margin.

Cognition has raised more than $3.9 billion. Poolside adds about $1.6 billion, Magic about $1 billion, Lovable $953 million and Replit $852 million. Factory has raised hundreds of millions more, while Augment and Windsurf/Codeium each add more than $200 million.

Even before counting broader developer infrastructure such as LangChain, Port or StackBlitz, the principal coding and software-building companies in our dataset represent well above $9 billion.

These companies are attacking different parts of the problem. Cognition wants agents to execute software-engineering work. Poolside has invested heavily in models for code. Lovable and Replit increasingly allow people with limited programming expertise to build complete applications. Factory focuses on autonomous software-development workflows inside enterprises.

Coding stands apart because investors are placing several very large bets on different versions of the same basic idea: software itself can increasingly be produced by software.

Startup	What it does	Total raised
Cognition	Autonomous software engineering	$3.9B+
Poolside	AI models for software development	$1.6B
Magic	Autonomous coding systems	$1.0B
Lovable	AI full-stack application builder	$953M
Replit	AI application development	$852M
Factory	Software engineering agents	$420M
Augment	Enterprise coding assistant	$252M
Codeium / Windsurf	AI-native coding IDE	$243M
Port	Agentic software engineering platform	$160M
StackBlitz	Browser-based AI application development	$113M

How much money is going into legal AI startups?

Legal Agentic AI has attracted roughly $4 billion across the companies in our ranking, making it one of the best-funded vertical AI markets.

Harvey leads with more than $1.8 billion. Legora follows at roughly $866 million. EvenUp has raised $385 million, while Eve, Luminance, Eudia, Supio, Spellbook and Robin AI have all raised meaningful amounts.

There is real depth below the two leaders. Solve Intelligence and DeepIP specialize in patent work. Supio and EvenUp focus heavily on plaintiff-side litigation. Spellbook and Robin AI work around contracts. Harvey and Legora aim much more broadly across legal workflows.

Harvey's latest financing shows how far the category has moved. The company raised $550 million at a $15.5 billion valuation and says its platform is used by 80% of Am Law 100 firms. Earlier in the year, it had raised $200 million at an $11 billion valuation.

Legora has followed a similar path, raising $550 million in a Series D at a $5.55 billion valuation.

Legal AI is well past the point where one successful company can be dismissed as an exception. Investors are financing a whole group of legal-agent businesses at serious scale.

Is customer service still one of the biggest Agentic AI markets?

Yes. Customer-service agents remain one of the largest Agentic AI funding categories, with several companies already in the hundreds of millions of dollars.

Sierra leads the category at roughly $1.6 billion of cumulative funding. Parloa has raised more than $560 million and Decagon about $481 million. Cresta has raised $276 million, while Kore.ai, PolyAI, Ada, Forethought, Bland AI, Yellow.ai and Maven AGI add hundreds of millions more.

There is a straightforward reason companies keep attracting capital here. Customer service already consists of large teams handling repeatable interactions that companies can measure through resolution rates, handling time and labor cost.

The product itself has changed too. The current generation increasingly promises to complete tasks rather than merely answer questions.

Sierra says companies now use its agents across interactions such as insurance claims, order returns and mortgage refinancing. Decagon describes its product as an AI concierge and says more than 100 new enterprise customers joined during its previous fiscal year.

Customer-service AI has moved a long way from the old chatbot market. Investors are funding systems expected to handle complete customer workflows, which can support much larger contracts.

Are investors mainly funding enterprise AI agents instead of consumer agents?

Yes. The Agentic AI funding leaderboard is overwhelmingly an enterprise leaderboard right now.

Look at the companies near the top. Cognition sells software-engineering automation. Harvey and Legora serve law firms and legal departments. Sierra, Parloa and Decagon sell customer-service agents. EliseAI works with property operators. Rogo targets financial professionals. Exaforce sells into security teams.

Even companies with large self-service audiences are moving deeper into business use. Lovable says nearly two-thirds of the Fortune 500 now have employees using its product, while Replit has continued pushing its application-building platform into professional workflows.

Consumer-focused autonomous agents are much harder to find near the top of the ranking.

The funding pattern is fairly easy to understand. Companies already have budgets for software engineering, lawyers, customer support, security, finance and recruiting. An Agentic AI vendor can enter one of those budgets and argue that its product saves hours, increases throughput or handles work that previously required employees.

Consumer agents still need to prove that individual users will pay enough, and stay long enough, to support similarly large businesses.

How much funding is going into infrastructure for AI agents?

Agent infrastructure has become a real funding category of its own, with companies such as Parallel, LangChain, Nimble and Arcade raising significant amounts to build the layers agents need underneath.

Parallel sits at roughly $386 million in our dataset and focuses on web infrastructure for AI systems. LangChain has raised about $160 million, including a $125 million Series B for its agent-engineering platform.

Nimble reached $75 million after raising a $47 million Series B to provide agents with live web data. Arcade has raised $72 million after a $60 million Series A aimed at the secure "action layer" that lets agents interact with external tools.

LlamaIndex, Sema4.ai and several smaller companies operate in adjacent parts of the stack.

The amounts are smaller than the money going into application leaders such as Cognition or Harvey, but the infrastructure category becomes more important as companies move agents into real production environments.

A useful agent needs more than a model. It needs current information, business-system access, authentication, permissions, monitoring and some way to check whether actions actually succeeded.

Those requirements are creating businesses that barely existed during the first chatbot wave.

Startup	Agent infrastructure focus	Total raised
Parallel	Web infrastructure for agents	$386M
LangChain	Agent development and observability	$160M
Nimble	Real-time web data for agents	$75M
Arcade	Secure agent action infrastructure	$72M
Sema4.ai	Enterprise agent development	$56M
LlamaIndex	Data framework for agents	$28M
Tavily	Web search for agents	$25M

Is cybersecurity becoming a major Agentic AI funding category?

Yes. Agentic cybersecurity is still smaller than coding or legal AI, but funding is now large enough to treat it as a serious category.

Exaforce is the strongest example. The company raised a $125 million Series B after a $75 million Series A, bringing total funding to $200 million in about a year.

The company says its system has already processed millions of security investigations. The idea is to let specialized AI agents handle work such as detection, triage, investigation and response that security analysts previously performed manually.

Dropzone AI has raised about $57 million, Prophet Security $41 million, Tuskira $29 million and Radiant Security $25 million. Simbian is another smaller entrant.

Security operations fit agents unusually well because analysts already work through a constant stream of alerts and investigations. Each alert means gathering context from different systems, checking what happened and deciding what should happen next.

If agents can do that reliably, companies can attach the technology to a very clear labor and response-time problem. Investors are now financing that possibility with much larger checks than they were only a couple of years ago.

How much funding are healthcare AI agents raising?

Healthcare agents have raised hundreds of millions of dollars, although the category remains well behind software engineering, legal AI and customer service.

Hippocratic AI leads the healthcare group in our ranking at roughly $403 million. Suki has raised about $168 million, Infinitus $103 million and Hyro $95 million. Sully.ai is smaller at a little above $20 million.

These companies are solving quite different problems.

Hippocratic AI develops patient-facing healthcare agents. Suki works around clinical assistance. Infinitus automates healthcare administrative communication. Hyro focuses on voice and conversational workflows.

That fragmentation helps explain why healthcare does not yet have a single company funded at Sierra or Harvey scale.

Healthcare also gives agents less room for error. A coding system can produce code that an engineer reviews. A patient-facing agent works around medical information, privacy rules and situations where an incorrect response can have far more serious consequences.

The funding is substantial, but investors have so far committed the very largest sums to workflows where deployment can move faster.

How much money is going into AI sales and recruiting agents?

Sales, recruiting and go-to-market Agentic AI startups have attracted well over $1 billion, although Mercor accounts for a large part of the total.

Mercor has raised roughly $1.1 billion in our dataset, but its business has expanded beyond its original recruiting marketplace. The company now talks much more about supplying human expertise to AI labs, building evaluations and helping enterprises deploy AI.

Among more traditional recruiting and sales-agent businesses, Paradox raised $253 million before Workday acquired it. 11x has raised $76 million, Unify $59 million, Regie.ai $51 million, Artisan $46 million and Landbase $43 million.

Farther down, 1mind, ConverzAI, Breakout, AiSDR and OptimHire show how crowded the category has become.

Sales and recruiting are obvious targets for agents because the workflow contains plenty of repeatable work: finding prospects or candidates, researching them, sending messages, booking meetings and updating software.

The funding totals suggest investors expect those jobs to change substantially. They have been much less willing, however, to put Cognition- or Harvey-level capital behind one dominant independent sales-agent company.

Who are the biggest investors in Agentic AI startups?

Accel, Andreessen Horowitz, General Catalyst, Sequoia, Khosla Ventures, Lightspeed and Bessemer appear again and again across the best-funded Agentic AI companies.

Accel backs companies including Cognition, Lovable, H, Port and several others. Andreessen Horowitz appears around Cognition, EliseAI, Decagon, Eve, Infinitus and 11x. General Catalyst has invested in companies including Mercor, Adept and Hippocratic AI.

Sequoia has exposure across coding, finance and agent infrastructure. Khosla Ventures appears in Factory, PolyAI, Exaforce, Regie.ai and Spellbook. Bessemer has invested across property, legal, software engineering and general Agentic AI.

The more interesting pattern is that many investors are making several overlapping bets rather than choosing one supposed Agentic AI winner.

Strategic investors are doing the same thing. Nvidia has appeared in financings across software-development, enterprise AI and professional applications. Salesforce has backed multiple enterprise-agent companies. Databricks invested in Nimble. SAP invested in Parloa. Wipro and Morgan Stanley participated in Arcade's recent financing.

The investor map increasingly looks like a portfolio approach to Agentic AI: own pieces of several workflows and several infrastructure layers because nobody yet knows which one will capture the largest share of the market.

Are newer Agentic AI startups raising money faster than older AI companies?

Yes. Some of the newest Agentic AI companies have raised more money in two or three years than earlier conversational-AI companies raised over most of a decade.

Cognition was founded in 2024 and has already raised more than $3.9 billion. Lovable launched its product in late 2024 and has reached roughly $953 million. Wonderful has accumulated about $834 million despite being a very young company.

Compare that with companies such as Ada, Kore.ai, Yellow.ai, Amelia and Leena AI. Those businesses entered conversational AI much earlier, but their cumulative funding totals generally remain below $250 million.

The difference shows how dramatically investors have repriced the opportunity.

Earlier conversational-AI companies were mostly expected to automate pieces of communication. Today's leading Agentic AI startups are pitching systems that write software, complete legal work, investigate security incidents or execute full customer workflows.

That promise supports much larger potential markets, but it also creates much larger expectations.

The fundraising speed is one of the clearest signs of how unusual this cycle has become. Venture capital is reaching companies at late-stage scale long before their operating histories look like traditional late-stage software companies.

Does raising more money mean an Agentic AI startup is actually winning?

No. Funding tells us how strongly investors are backing a company and how much money it has available to compete, but it does not tell us who has the best product or business.

Some highly funded startups can spend aggressively for years before proving that the economics work. A lower-funded company can sometimes reach meaningful revenue much more efficiently.

Funding also stays in the historical total after the competitive situation changes. Moveworks raised roughly $308 million before ServiceNow acquired it in a $2.85 billion transaction. Paradox raised about $253 million before being acquired by Workday. Windsurf/Codeium, Cognigy, Forethought, LawGeex and other companies in the ranking have also been acquired.

Those companies have not disappeared technologically, but they no longer compete in exactly the same way as independent startups.

Cumulative funding works best as a measure of financial scale. It shows which startups attracted the most equity capital and therefore had the largest pools of outside money available to build their businesses.

Revenue growth, customer retention, gross margins, usage and capital efficiency would be needed to judge which companies are actually building the strongest businesses.

Which Agentic AI startups are raising money fastest right now?

Cognition, Harvey, Wonderful and several other companies are raising money at a pace that would have looked extraordinary for software startups only a few years ago.

Cognition raised $1 billion and then followed with more than $2 billion within months. Its valuation moved from about $26 billion to $48 billion over the same stretch.

Harvey raised $200 million at an $11 billion valuation and later added another $550 million at $15.5 billion. Wonderful went from a $150 million Series B at a $2 billion valuation to a $550 million Series C at a $5 billion valuation within roughly half a year.

Lovable raised $400 million in its Series C after reaching substantial scale extremely quickly. Sierra raised $950 million and said more than 40% of the Fortune 50 were already customers. Exaforce moved from a $75 million Series A to a $125 million Series B in about a year.

These funding sequences tell us more than an isolated large round. Investors are returning quickly and putting substantially more money into companies they believe are converting the Agentic AI boom into real usage.

The pace will be difficult to sustain indefinitely. For now, though, a small group of companies is moving through private-market financing stages extraordinarily quickly.

What does the Agentic AI funding ranking tell us about where this market is heading?

The clearest conclusion from the Agentic AI funding ranking is that investors are betting on agents becoming workers inside specific business functions, with software engineering leading the race.

Our dataset contains roughly $25.3 billion across 99 companies, yet about three quarters of that capital belongs to the top 20. Cognition alone has raised more than $3.9 billion, while six companies are already around or above $1 billion.

Coding has attracted the most money. Legal AI has become one of the deepest vertical markets. Customer service remains huge. Finance, cybersecurity and healthcare have produced smaller but increasingly well-funded groups. Underneath all of them, a separate infrastructure layer is emerging around agent data, orchestration, search, permissions and secure actions.

Enterprise use is also much more prominent than consumer use among the funding leaders. Investors are following existing pools of professional spending where companies can measure whether an agent saves time or completes work.

The biggest uncertainty is whether the amount of capital being committed now is ahead of what the eventual economics can support. Several companies are already being financed like mature category leaders while the underlying technology is changing incredibly fast.

The Agentic AI funding boom is very real, but the important part is where the money is going. Investors are concentrating extraordinary amounts of capital behind a small number of companies trying to automate expensive professional work. Today, software engineering is the clearest center of gravity, with legal work and customer operations forming the next large battlegrounds.