Who are the top investors in agentic AI?
In our agentic AI market deck, you will find everything you need to understand the market
SUMMARY
Sequoia Capital is the top investor in agentic AI today, with Andreessen Horowitz, Benchmark and Accel forming the strongest group behind it.
The ranking is unusually fluid because agentic AI is still young. Most meaningful investments have not produced realized exits yet, so early entry, company quality, follow-on conviction and commercial progress tell us more than historical returns.
Deal count is particularly misleading in this market. Agentic AI now covers customer-service agents, legal automation, coding, cybersecurity, workflow orchestration, browser infrastructure, memory, identity and payments, meaning a large portfolio can hide very different levels of conviction.
The strongest investor records tend to start before a company becomes an obvious growth-stage winner. Sequoia entering Harvey in 2023, Benchmark backing LangChain at seed and a16z investing in Decagon before its public emergence are more informative for this ranking than later nine-figure checks.
Sequoia stands out because its portfolio is both broad and proven relative to the age of the category. Harvey, Sierra, XBOW, Factory and Sail Research give it credible exposure across professional work, customer interactions, cybersecurity, coding and long-running agent infrastructure.
a16z is the closest challenger, but its strength is slightly different. Its best positions are concentrated in applications that replace or automate chunks of enterprise work, including Decagon, HappyRobot, Lio, 11x and increasingly autonomous software development through Replit.
Benchmark has fewer relevant bets than most of the firms above it in raw AI rankings, yet that is almost the point. LangChain, 11x and Gumloop give it one of the highest concentrations of strong early agent investments in the market.
Accel has changed fastest. What began as a particularly strong Decagon position has expanded into orchestration, browser infrastructure, agent payments, workplace agents and security through companies including n8n, Kernel, Sapiom, Viktor and Tolmo.
Agent security is becoming its own investment battleground. Bessemer's recent activity around Onyx Security, Neo, Act and Dome suggests that controlling what autonomous software can access and do may become nearly as important as building the agents themselves.
Specialization can matter more than overall rank for founders. Felicis looks particularly strong in agent infrastructure, while General Catalyst has built an interesting cluster around agent identity, communication, permissions and training.
Growth-stage investors such as Coatue, Greenoaks, Tiger Global, GV and GIC remain important because the largest agent companies now require very large checks. They rank lower here because access to proven winners tells us less about early investment judgment than identifying those companies before the market had formed.
The ranking will probably change quickly. For now, Sequoia has the best combination of early winners, portfolio breadth and repeat conviction, while Accel and Bessemer are the two large firms whose recent activity makes the next revision most interesting.
This market map, featured in our agentic AI market deck, highlights top companies and startups in the agentic AI market
Why is ranking agentic AI investors so messy right now?
Ranking agentic AI investors is messy right now because the category has expanded faster than the investing track record can mature.
Our 12-month funding cut found 40 disclosed pure-play agentic AI rounds worth about $3.37 billion. A subsequent funding sweep captured another 12 qualifying rounds worth roughly $676 million. Money is arriving quickly, and it is heavily concentrated: in the 12-month sample, the ten largest rounds accounted for 77% of all disclosed capital.
At the same time, “agentic AI” now describes several markets that barely resemble one another. Sierra builds customer-facing agents. Harvey builds agents for legal work. LangChain provides infrastructure for developers building agents. Runlayer helps companies connect those agents securely to enterprise systems. Sapiom is building payment infrastructure for autonomous software. Putting all of those companies into one bucket makes a simple deal-count ranking fairly useless.
The technology is also moving from experiments into production. LangChain surveyed more than 1,300 people working with AI agents in 2026 and found that 57% already had agents in production. Among those teams, 89% had implemented observability. Adoption is now real enough to judge investor choices, while the category is still young enough that realized venture returns tell us very little.
So we have to judge the investors through the decisions they have actually made: when they entered, which companies they picked, whether they kept investing, and how often those companies later became important.
What counts as an agentic AI investment here?
For this ranking, an agentic AI company has to let software plan and carry out meaningful work, or provide infrastructure that agents need to operate reliably.
That includes application companies such as Sierra, Harvey and 11x, where software can complete parts of a job. It also includes infrastructure such as LangChain for building and observing agents, Letta for memory, Browser Use for web interaction and Runlayer for secure connections to enterprise tools.
We also include specialized agents when autonomous execution is central to the product. XBOW autonomously performs penetration testing. Lio handles parts of enterprise procurement. Viktor works inside Slack and Microsoft Teams as an AI coworker. Tolmo is developing security agents that reason over code, infrastructure and production data.
A generic SaaS company does not enter our ranking just because it has launched an “agent” feature. We also leave the large foundation-model laboratories out unless an investment relates directly to an agent company. Otherwise, investors with gigantic general AI portfolios would automatically dominate a ranking that is supposed to tell us who has actually been good at investing in agents.
As this chart shows, and as featured in our agentic AI market deck, search interest in AI agents has been rising rapidly
Does writing the biggest check make someone a top agentic AI investor?
No. The strongest agentic AI investing records usually belong to firms that entered before the company became an obvious growth-stage winner.
Sierra is a good example. Sequoia and Benchmark had already put a combined $110 million into the company around its launch in 2024. By May 2026, Sierra was raising another $950 million in a round led by Tiger Global and GV at a post-money valuation above $15 billion. Sierra said at that point that more than 40% of the Fortune 50 were customers and that its agents were handling billions of interactions.
The later investment is huge, but the earlier decision tells us more about investor judgment. Sequoia and Benchmark committed when enterprise AI agents still had limited commercial proof. Tiger and GV invested after Sierra had become one of the clearest winners in enterprise AI.
Harvey shows the same progression. Sequoia led its $21 million Series A in 2023. Harvey later raised $200 million at an $11 billion valuation, with GIC and Sequoia co-leading the round. By then, customers were running more than 25,000 custom agents on Harvey.
For our ranking, we therefore give more weight to early entry, repeated conviction and the eventual quality of the company than to the absolute amount invested.
If you want more recent data on this point, please see our latest agentic AI market report.
Which VCs have the strongest agentic AI portfolios today?
Sequoia currently has the strongest overall agentic AI portfolio, with a16z and Benchmark forming the next tier and Accel closing the gap quickly.
The separation becomes clearer when we look at the portfolios rather than total AI deal counts. Sequoia has strong companies across customer agents, professional work, cybersecurity, coding, recruiting and agent infrastructure. a16z has repeatedly invested in autonomous enterprise applications. Benchmark has made fewer bets, yet several became important companies. Accel has built a much broader agent portfolio over the past year.
Bessemer has also moved up materially. Its recent activity in agent security, vertical AI and voice agents now looks too broad to treat as a side theme. Index remains unusually strong in enterprise agents, while Felicis has one of the cleanest portfolios in infrastructure specifically designed for autonomous software.
| Current tier | Investor | What we see in the portfolio |
|---|---|---|
| 1 | Sequoia Capital | Best mix of early winners, breadth and continued follow-on investing |
| 2 | Andreessen Horowitz | Strong early application bets plus aggressive expansion across enterprise agents |
| 3 | Benchmark | Smaller portfolio with an unusually high concentration of strong companies |
| 4 | Accel | Rapidly expanding across applications, orchestration, browser infrastructure, payments and security |
| 5 | Bessemer Venture Partners | Recent acceleration in agent security plus a deeper existing vertical-AI portfolio |
| 6 | Index Ventures | Deep conviction in large enterprise-agent platforms |
| 7 | Felicis | Particularly strong in the technical infrastructure agents need |
| 8 | General Catalyst | Growing position in agent identity, communication, training and control infrastructure |
This chart, included in our agentic AI market deck, illustrates yearly VC funding for agentic AI startups
Why is Sequoia the top agentic AI investor right now?
Sequoia is our clear number one in agentic AI today because its strongest investments cover several different ways autonomous software could win.
Harvey is probably the best example of Sequoia getting both the market and the company right early. Sequoia led Harvey's $21 million Series A in 2023, when generative AI in legal services was still an experiment. Harvey is now valued at $11 billion, customers run more than 25,000 custom agents on the platform, and its latest customer announcements include firmwide deployments at Jackson Lewis, Nelson Mullins and Davis Wright Tremaine.
XBOW gives Sequoia a completely different type of winner. Sequoia led XBOW's $20 million seed round in 2024. The company later raised a $120 million Series C at a valuation above $1 billion as its autonomous penetration-testing product moved into large enterprises.
Sequoia has kept finding new layers of the market as well. It led the seed financing of Sail Research, which is building infrastructure specifically for long-running agents. Sail subsequently announced $80 million across its seed and Series A financings at a $450 million valuation. Factory gives Sequoia exposure to autonomous software engineering, while Auctor and several newer investments extend the portfolio into specialized enterprise work.
Sequoia was also there before Sierra became a $15 billion-plus company. Put those investments together and we get a portfolio spanning legal work, customer interactions, cybersecurity, coding and long-horizon agent infrastructure. No other investor currently matches that range with the same number of companies that have already broken out.
If you want more recent data on this point, please see our latest agentic AI market report.
Is a16z really Sequoia's closest rival in agentic AI?
Yes. a16z is the closest rival right now because it has repeatedly found agent businesses across very different enterprise workflows.
Decagon is the strongest piece of evidence. a16z led the company's $5 million seed round before Decagon emerged publicly, while Accel led the $30 million Series A announced alongside it. Decagon went on to become one of the largest independent customer-service agent companies.
HappyRobot strengthens the case. a16z led the company's $15.6 million Series A when it was building AI workers for logistics operations. In its latest financing, HappyRobot raised another $150 million at a $1.2 billion valuation, and a16z invested again. The company says its agents now handle calls, emails and operational workflows across complex supply chains.
Lio shows that a16z is still making fresh early bets. The firm led Lio's $30 million Series A in 2026 to build autonomous procurement workflows for large companies. Its portfolio also includes 11x in digital sales workers and Replit, whose product has moved aggressively toward autonomous software creation.
We still give Sequoia the edge because more of its strongest agent companies were captured early across a wider set of markets. Two or three newer a16z investments becoming category leaders could easily tighten the gap.
This chart, included in our agentic AI market deck, shows how Cognition is positioned in agentic AI
Why does Benchmark rank so high with fewer AI-agent bets?
Benchmark ranks third because its relatively small agentic AI portfolio contains an unusually high number of companies that became important.
LangChain is the clearest example. Benchmark led its $10 million seed round in 2023, before agent engineering had become a major software category. LangChain later raised at a $1.25 billion valuation and has become one of the best-known development stacks for building, evaluating and observing AI agents.
Benchmark also led 11x's $24 million Series A as the company pursued autonomous digital workers for sales. More recently, Benchmark led Gumloop's $50 million Series B. Gumloop says companies including Shopify, Ramp, Gusto, Samsara, Instacart and Opendoor use its platform to let employees build agents for multi-step business workflows.
The impressive part is the concentration. Benchmark can point to important companies in agent infrastructure, autonomous enterprise workers and no-code agent creation without needing dozens of investments to get there.
We rank Benchmark below a16z because its portfolio is narrower. If the question were simply which investor made the highest concentration of good early agent picks, Benchmark could plausibly rank first.
If you want more recent data on this point, please see our latest agentic AI market report.
Is Accel the fastest-rising agentic AI investor?
Accel is probably the fastest-rising large VC in agentic AI right now.
Its Decagon history already gave it one excellent position. Accel led Decagon's $30 million Series A after partner Ivan Zhou had invested personally before the company was incorporated, then kept investing through later rounds.
What has happened since is more interesting. Within roughly a year, Accel announced investments touching almost every major layer of the agent stack. n8n handles agentic workflow orchestration. Kernel gives agents browser infrastructure. Sapiom is building financial rails for autonomous software. Viktor acts as an AI coworker inside workplace chat. Tolmo is building security agents with production context.
Those five companies are solving five different problems. The common thread is autonomous software moving from demonstrations into real company systems.
Accel is also an investor in Cursor, where autonomous coding has become increasingly central to the product. Its portfolio has therefore grown in two directions at once: agents that perform jobs and infrastructure that makes those agents useful.
A year ago, Accel looked like a strong investor with a standout Decagon position. Today, it looks like a firm deliberately trying to own several layers of the agent economy. That is why we now put it fourth.
This chart, included in our agentic AI market deck, illustrates yearly funding for agentic AI startups
Has Bessemer become a top agentic AI investor lately?
Yes. Bessemer has moved into our top five lately, largely because its agent-security portfolio has become much stronger while its vertical-AI bets keep adding up.
Onyx Security is the clearest recent example. Bessemer led the company's $113 million Series B after Onyx said revenue had quadrupled during the four months following its public launch. Onyx also says its platform already secures more than 1.1 million agents and has inspected more than 66 million AI sessions.
That investment sits inside a wider pattern. Bessemer has backed Neo around endpoint security for AI agents, Act in proactive cloud access security and Dome Systems around operational control for the agentic enterprise. The firm also invested earlier in Vapi, which provides infrastructure for voice agents.
Its application portfolio is broader again. Bessemer is an investor in Legora for legal workflows, Relevance AI for building AI workforces and several vertical automation companies. The firm has also backed enterprise-agent platform Wonderful.
The recent pace changed our ranking. Bessemer used to look like a broad cloud investor with some relevant AI holdings. Now there is a coherent bet on what happens when autonomous software gets access to enterprise systems: companies will need agents, controls, identity, security and highly specialized applications. That is strong enough for fifth place today.
If you want more recent data on this point, please see our latest agentic AI market report.
Is Index still one of the best agentic AI investors?
Yes. Index remains one of the strongest agentic AI investors, with an especially valuable position in enterprise agents deployed across multiple countries.
Wonderful carries much of that weight. Index led Wonderful's $34 million seed and then its $100 million Series A only a few months later. Wonderful subsequently raised another $150 million at a $2 billion valuation, with Index investing again. Index had therefore backed the company three times within roughly a year.
The operational growth behind those financings is unusual. Wonderful says its agents are deployed across 30 countries in telecoms, financial services, manufacturing and healthcare. The company has reported containment rates above 80% in some deployments and reductions in handling time of as much as 60%.
Index has also moved into infrastructure through NewCore, which treats AI agents as enterprise identities that need their own permissions and controls.
We place Index sixth because the portfolio is still concentrated. Its Wonderful position may ultimately prove more valuable than dozens of smaller bets, but Sequoia, a16z, Accel and the newly expanded Bessemer portfolio currently give us more independent evidence across the market.
This chart, included in our agentic AI market deck, compares the main business model options for autonomous AI agent platforms
Is Felicis the best specialist investor in agent infrastructure?
Felicis is currently the strongest specialist we found for the plumbing around AI agents.
The portfolio almost reads like a list of technical problems that appeared once agents started doing longer and more complicated tasks. Letta gives agents persistent memory. Felicis led its $10 million seed when the company emerged from Berkeley's MemGPT work. Browser Use gives agents a way to interact with websites and had already passed 46,000 GitHub stars around the time Felicis led its $17 million seed.
Runlayer handles secure connections between enterprise systems and agents. Felicis invested at seed and later led its Series A. Automat takes business instructions and turns them into agents that operate software like a person would; Felicis led its $15.5 million Series A.
The portfolio now extends into computer-using agents such as Simular and more specialized autonomous applications. Felicis has also published detailed work around context engineering, agent security and the division of work between humans and autonomous systems.
We rank Felicis seventh overall because most of these companies are still earlier than the flagship holdings of Sequoia or Benchmark. For a founder building agent memory, browser interaction, runtime security or another new infrastructure layer, Felicis would rank much higher than seventh on the practical shortlist.
Does Y Combinator belong in the same ranking as venture funds?
Y Combinator belongs in the agentic AI conversation, and its real advantage is the sheer number of agent companies it sees before traditional venture rounds begin.
YC has been involved in Gumloop, Juicebox, AgentMail, HappyRobot and a long tail of younger agent startups. Its batches give it exposure to new agent ideas while the markets are still being invented, including communication infrastructure, vertical workers, recruiting agents and workflow automation.
HappyRobot shows how that early access can compound. YC entered long before the company's latest $150 million round, and it invested again alongside a16z and other existing backers as HappyRobot crossed a $1 billion valuation.
The difficulty comes when we compare investment models. Benchmark may make only a handful of concentrated bets. YC systematically invests across large startup batches. Counting winners without accounting for the number of companies backed would make YC look artificially dominant.
We therefore treat YC as the strongest agentic AI pipeline rather than as the best stock picker. It still belongs in the overall ranking because access to promising founders at inception is a genuine investing advantage.
This chart, featured in our agentic AI market deck, shows the share of revenue generated by each customer segment in the agentic AI market
Which agentic AI investors are strongest at seed, Series A and growth?
The best agentic AI investor depends heavily on when a company is raising money.
At seed, Felicis, Benchmark, a16z, Sequoia, General Catalyst and YC all have evidence of entering before much commercial proof existed. General Catalyst deserves more attention here than its overall ranking suggests. It led AgentMail's $6 million financing, backed Portia's infrastructure for controllable agents, led Kite's seed round around agent identity and payments, and most recently led Arga's $10 million seed for training environments that help enterprise agents learn how to use systems such as Salesforce and Workday.
Series A is where Accel becomes particularly strong. Decagon, Viktor and Tolmo show a willingness to lead while the market is still forming. Sequoia's Harvey investment and Benchmark's 11x position also came at this stage.
Growth investing produces a different group. Coatue, Greenoaks, Tiger Global, GV and GIC become much more visible once companies have already shown large-scale adoption. Those firms can still generate excellent returns, though their investment records tell us less about who recognized agentic AI earliest.
| Stage | Investors we would watch first | Why they stand out |
|---|---|---|
| Pre-seed / Seed | YC, Felicis, Benchmark, a16z, Sequoia, General Catalyst | Strong access before categories and business models are settled |
| Series A | Sequoia, Accel, Benchmark, a16z, Index | Repeated willingness to lead once early product evidence appears |
| Series B / C | Sequoia, Accel, a16z, Benchmark, Bessemer, Index | Strong follow-on capacity while companies are still establishing leadership |
| Growth | Sequoia, Coatue, Greenoaks, GIC, GV, Tiger Global | Access to large proven agent companies and ability to write very large checks |
Which investors are strongest in each part of agentic AI?
Different investors are already developing recognizable strengths inside agentic AI.
Customer-service agents have attracted some of the largest venture bets because the job is repetitive, measurable and expensive for enterprises. Coding has created another huge pool of capital because software agents can work inside an environment where actions and outputs are relatively easy to test. Security is now accelerating as enterprises discover that autonomous software needs permissions, monitoring and controls of its own.
Infrastructure is more fragmented. Memory, browsers, identity, payments, observability and secure tool access could each become meaningful software markets if the number of production agents keeps growing.
That fragmentation helps explain why Sequoia leads the overall ranking. It has credible exposure across several of these categories, while most other firms have developed stronger concentrations in one or two.
| Part of agentic AI | Investors that stand out | Examples of the underlying thesis |
|---|---|---|
| Customer and enterprise agents | Sequoia, a16z, Index, Accel, Benchmark | Autonomous customer service, operational work and digital employees |
| Agent builders and orchestration | Benchmark, Accel, Sequoia | Agent creation, workflow orchestration, evaluation and deployment |
| Agent security and governance | Bessemer, Felicis, Accel | Runtime control, secure tool access and autonomous security work |
| Vertical professional agents | Sequoia, a16z, Bessemer | Legal, procurement, recruiting and other specialized knowledge work |
| Browser and computer use | Felicis, Accel | Letting agents operate existing websites and software interfaces |
| Identity and agent transactions | General Catalyst, Accel, Index | Credentials, permissions, communication and payments for autonomous software |
This chart, included in our agentic AI market deck, shows how autonomous AI agent platform technology has evolved over time
Who are the top investors in agentic AI overall?
Sequoia Capital is the top agentic AI investor today, followed by a16z, Benchmark and Accel, while Bessemer has moved into the top five after a burst of strong agent-focused investing.
Sequoia wins on the combination we care about most: early entry, multiple companies that later became valuable, breadth across the market and continued investing as those companies scaled. a16z has the closest comparable application portfolio. Benchmark has fewer bets and an exceptional concentration of good ones. Accel's portfolio has expanded so quickly that it could move higher if its newer infrastructure investments break out.
Bessemer is our biggest upward revision. Its recent run across autonomous security, enterprise controls and vertical AI now looks like a real portfolio strategy. Index remains one of the strongest investors in enterprise agents, while Felicis stands out for getting unusually specific about the infrastructure autonomous systems require.
General Catalyst is the investor just outside that group that we would watch most closely. Its bets in Portia, Kite, AgentMail and Arga cover control, identity, communication and training, four problems that become more important as companies deploy more agents.
The ranking will move. Agentic AI is still early enough that many of today's seed investments will disappear, while a company that looks small now could define an entire layer of the market. Based on the evidence available today, though, Sequoia has the strongest claim to the top spot.
| Rank | Investor | Why we rank it here | Representative agentic AI exposure |
|---|---|---|---|
| 1 | Sequoia Capital | Strongest mix of early wins, breadth and follow-on conviction | XBOW, Factory, Sail Research, Auctor |
| 2 | Andreessen Horowitz | Excellent application portfolio with repeated early-stage conviction | 11x, Lio, Replit |
| 3 | Benchmark | Exceptional quality relative to the number of bets | LangChain, 11x, Gumloop |
| 4 | Accel | Fastest-expanding large-fund portfolio across the agent stack | n8n, Kernel, Viktor, Tolmo |
| 5 | Bessemer Venture Partners | Major recent move into agent security plus strong vertical AI exposure | Vapi, Legora, Neo, Relevance AI |
| 6 | Index Ventures | Deep and repeated conviction in global enterprise agents | Wonderful, NewCore |
| 7 | Felicis | Best specialist portfolio around core agent infrastructure | Letta, Automat, Simular |
| 8 | General Catalyst | Strong emerging position in identity, control and agent-native infrastructure | Portia, Kite, Arga |
| 9 | Y Combinator | Unmatched early-stage pipeline across many emerging agent categories | Juicebox, AgentMail and numerous seed-stage agent companies |
| 10 | Lightspeed Venture Partners | Broad enterprise-AI exposure with several relevant agent platforms | Glean, Stacks and other enterprise automation companies |
| 11 | Coatue | Strong access to later-stage companies once agent businesses reach scale | Major growth-stage agent investments |
| 12 | Kleiner Perkins | Selective exposure to important vertical agents and new agent infrastructure | Harvey, Sail Research |
If you want more recent data on this point, please see our latest agentic AI market report.
OUR METHODOLOGY
This ranking looks at who has built the strongest investing record in agentic AI rather than who has simply invested the most money in AI. We compare investors across entry timing, company selection, portfolio breadth and coherence, follow-on conviction, and subsequent evidence that their portfolio companies are becoming meaningful businesses.
We give more weight to early investments than to large growth-stage checks made after a company has already become an obvious winner. Deal count is also treated carefully because a broad portfolio does not automatically imply strong selection, while a smaller portfolio can be highly significant when several early bets develop into category leaders.
Our definition includes companies where autonomous execution is central to the product, along with infrastructure that agents need to operate reliably. That covers application companies such as Harvey, Sierra, 11x and XBOW, and infrastructure around orchestration, memory, browser interaction, security, identity and agent transactions. Generic software companies are not included merely because they have added an “agent” feature.
We reviewed recent financings, investor announcements, portfolio activity, customer adoption, valuations, deployment milestones and follow-on rounds. Recent evidence receives particular weight because agentic AI portfolios are changing unusually quickly and several investors have built substantial positions within the past year alone.
Key sources include LangChain's 2026 State of Agent Engineering survey, Sierra's financing and adoption update, Sequoia's announcement of its Harvey investment, Harvey's 2026 financing announcement, XBOW's Series C announcement, XBOW's earlier financing history, Sail Research's seed and Series A announcement, LangChain's Benchmark-led seed announcement, 11x's Series A announcement, Accel on its Decagon Series A, Andreessen Horowitz on HappyRobot, HappyRobot's Series C announcement, Andreessen Horowitz on Lio, Sapiom's Accel-led financing announcement, Accel's portfolio updates, Bessemer on Onyx Security, Index Ventures on Wonderful's seed and Series A, Index Ventures on Wonderful's Series B, and Felicis's investment notes on Letta, Browser Use and Runlayer.
In our agentic AI market deck, we identify pain points entrepreneurs should prioritize
Related blog posts
- Where can a new company still break in AI agents?
