The analysis of equity rounds done by New Market Pitch shows that there was a notable growth in the investments made in the area of the AI agent enterprise business segment. It jumped from about $1.5B through 31 transactions in 2024 to about $2.9B through 50 transactions in 2025. There is enough evidence to show that the segment is poised to grow even faster in 2026.

Such a funding trend can be attributed to the shift from the phase where agentic AI is proving its capabilities (proof-of-concept) to one in which such AI is actually being used (production-grade systems). In the first five months of 2026, agentic AI companies raised close to $1.1 billion in 29 deals as opposed to the $538 million raised by only nine deals in the first five months of 2025. That puts early 2026 at roughly twice the capital and triple the deal volume of the prior year’s comparable period.

Vertical AI captures most funding

Capital is being deployed unevenly. Vertical AI agents, companies that develop autonomous systems for cybersecurity, healthcare operations, and compliance, made up 48.3% of the number of transactions and 54.6% of capital raised in 2026 year-to-date, according to New Market Pitch.

Increasing amounts of attention are being dedicated to agent execution infrastructure. These include runtimes, sandboxes, identity layers, observability systems, and security testing, which accounted for 20.7% of all investments in Q1 2026. The agent development platform industry is also rebounding from its lackluster performance in 2025 and has reached about $124M in total across five deals by late May.

Orthogonal has secured $4.3 million in seed financing from Pantera Capital in order to create infrastructure for AI agents to locate services, coordinate task execution, and perform transactions across the internet, according to a June 25 press release. The company aims to help prepare for a future in which software agents increasingly take on tasks such as performing research, making reservations, or managing data flows on behalf of users, thus leading to an increasing need for an infrastructure allowing them to discover services and perform transactions automatically.

As one of the co-founders of Orthogonal put it succinctly: “There will be more agents than people online,” which means that machine-to-machine interactions will become predominant on the web instead of human browsing.

RunPod, a cloud platform for AI developers, has completed a $100 million growth round led by Summit Partners. The company will use this new investment to provide infrastructure for training, fine-tuning, and deploying AI models, servicing over a million developers.

Capital is skewed

Although the headline numbers are impressive, they mask the reality of the disparity between large deals and small deals. As of year-to-date (YTD) 2026, the three largest YTD 2026 deals accounted for 44% of total capital deployed, while the top ten accounted for approximately 78%, according to New Market Pitch. Conversely, the bottom half of all YTD 2026 deals made up around 11.5%.

In addition to these high dollar amounts, the sizes of funding rounds also paint a picture of this disparity. The average funding round for 2026 is approximately $36 million, while the median funding round for 2026 is around $19 million, indicating that there is a very limited number of large round sizes skewing the average upward.

March 2026 alone contributed roughly $650 million, more than 60% of the year’s total through May. This includes significant rounds in the security, customer service, and agent control plane markets.

Early-stage funding is not dead

There was a significant shift in the movement of capital to later-stage companies during 2025. About 77% of the dollars invested in all companies were invested in a company’s Series B and later. This pattern appears to be partially reversing for early 2026. Roughly $378 million, or 36% of the total capital invested in the first five months of 2026, was for seed and Series A rounds of financing, says New Market Pitch.

Mirendil, an AI lab formed by former researchers from Anthropic, OpenAI, Google DeepMind, and xAI, has raised $200 million in seed funding at a $1 billion valuation per Cryptopolitan. Andreessen Horowitz and Kleiner Perkins co-led the round, with NVIDIA also being a participant. Mirendil has not launched any commercial products, nor has it communicated any meaningful technical details.

According to McKinsey, agentic commerce (transactions that are either influenced or enacted by AI agents) will be a $3 trillion to $5 trillion global industry by 2030. The amount of money going into the infrastructure to support the expected scale of AI agents is likely why investors are funding the infrastructure now. If AI is going to take on responsibilities for discovery, decision-making, and payment activities at such a large scale, then infrastructure will need to be built to support these business activities.

The smartest crypto minds already read our newsletter. Want in? Join them.