Asymptote Labs raises $5.5M Seed for AI agent control. Tower Research Ventures has a second way to examine its latest AI infrastructure investment: its own security organization is evaluating Beacon on the trading firm's endpoints. That gives Asymptote Labs an unusually concrete setting for its pitch, as the New York startup works to make AI-agent activity something an enterprise can inspect and control across different tools. Asymptote announced a $5.5M Seed round on October 20, 2026, led by Tower Research Ventures, with Lytical Ventures, 1984 Ventures and South Park Commons participating. The financing supports engineering, product development, hiring and go-to-market. Co-founders Justin D'Souza, CEO, and Shukan Shah, CTO, are building a business around the evidence agents leave while doing useful work. The purchase decision reaches beyond choosing a better coding assistant. Security teams need to follow actions across those assistants, while engineering teams want useful history to survive the end of a session. Asymptote's combination of open-source collection and commercial governance puts those two needs on the same underlying record. What Asymptote's Seed round supports. Tower Research Ventures is the venture arm of Tower Research Capital. Its investment and its security team's evaluation are distinct relationships: one finances Asymptote, while the other explores how Beacon works in an operating environment. The announcement does not establish Tower as a paying enterprise customer or disclose the outcome of that evaluation. The founders bring experience from the work of identifying threats at Doppel. Asymptote's current team page identifies D'Souza's background in machine-learning infrastructure and Shah's in security operations automation, alongside earlier payments work. Their familiarity with both detection and workflow matters to a product that has to interpret what happened and make the resulting information usable by another team. South Park Commons lists Asymptote as a New York company founded in 2025 and identifies both founders. Its involvement includes the Founder Fellowship described in Asymptote's announcement, giving this financing continuity with the company's formation. No valuation or detailed lifetime funding total has been disclosed in the sources reviewed. For buyers, the use of proceeds points toward a practical expansion of the supplier as well as its software. Engineering and product investment support a growing range of agent environments; go-to-market and hiring support the people who explain, deploy and maintain the offering. Asymptote's careers page provides a current route into that work, without disclosing a numerical hiring target. A common record across different agents. Agent Beacon is the open-source foundation. Its repository describes a shared event model for prompts, responses, commands, file activity, approvals and other session details across agent harnesses. For an engineering organization, that offers a way to preserve history as employees move among tools with different interfaces and conventions. A common record can change the handoff between a developer and a security analyst. Rather than first deciding which assistant holds the evidence, the analyst can begin with the work itself and follow its sequence. The normalized record provides a starting point for that investigation; an analyst still has to interpret the action and its consequences. Beacon's current repository also describes a process for turning session history into reusable knowledge. Lessons can be extracted, reviewed and made available to future agents through MCP or Agent Skills. That connects security evidence with a different engineering concern: retaining a useful debugging path or repository convention after the person and assistant who discovered it have moved on. Deployment choices deserve attention alongside those capabilities. The README explains that interactive setup preselects Beacon Cloud with a local opt-out, while other installation paths have different behavior. A buyer assessing the product should examine the actual destination, capture mode and retention choices for its deployment, because prompts and traces can contain material the organization also needs to govern. Where open source becomes an enterprise purchase. Asymptote's pricing and deployment comparison separates free Community collection from a custom-priced Enterprise service and Private Deployment. The paid offering adds managed ingestion, retention, search and analytics, alongside identity, policy, approvals and governance. Private deployment options address organizations that need tighter control over infrastructure and data location. That packaging gives the developer and the enterprise buyer different reasons to participate. An engineer may value the portable session record immediately, while an organization pays for the work of operating that record across a fleet. Keeping the collector open also lets a prospective customer inspect an important part of what it would introduce into its environment. The commercial argument depends on the relationship between observation and intervention. Asymptote advertises runtime detection and policy enforcement for agent actions, while its documentation describes the managed platform as a layer for visibility and governance. Teams evaluating that offer still have to decide which actions require approval, who owns the policy and how exceptions fit their development process. Those decisions place Asymptote beside existing security and observability systems, within the enterprise-agent security market also covered in DevCuration's Opsin reporting. The company offers forwarding into established tools, so a deployment can extend the systems a team already operates. For procurement, integration and responsibility become useful evaluation criteria: where an event arrives, which team investigates it and where a decision to block an action is enforced. The operating relationship behind the round. Asymptote's announcement reports production use by security teams without naming customers or disclosing revenue. Developer interest and an investor's internal evaluation describe different parts of adoption, and neither supplies a public contract value. Keeping those distinctions visible gives operators a clearer view of the supplier they are assessing. The broader opportunity follows from letting developers choose among capable agents while maintaining organizational continuity. A company benefits when a useful activity record can outlast a tool decision, support an investigation and carry approved learning into later work. Asymptote is trying to serve that continuity through an open foundation and a managed layer that different teams can use for different purposes. Tower's security organization now has a role in exploring that proposition on its own endpoints. D'Souza and Shah have new capital to build the coverage, controls and support around it, as each additional agent brings another sequence of actions for someone inside the enterprise to understand.

Asymptote Labs raises $5.5M to track and control what AI agents do. Its free Beacon tool logs activity from Claude Code, Codex, Cursor and more than 20 other agent tools. October 20, 2026 - 4:17 pm New York startup Asymptote Labs has raised a $5.5M seed round led by Tower Research Ventures. It makes Beacon, a free open-source tool that records what AI agents do, and sells a paid platform that can block risky agent actions. The money will go on engineering, hiring and sales. Asymptote Labs, a New York startup that monitors what AI agents do inside companies, has raised $5.5M in seed funding, it said on Tuesday. Tower Research Ventures, the venture arm of trading firm Tower Research Capital, led the round. Lytical Ventures, 1984 Ventures and South Park Commons also took part. The company did not disclose its valuation. Asymptote's main product is Beacon, a free, open-source tool that records what AI agents do. It works with coding agents such as Claude Code, Codex and Cursor, and more than 20 others. It puts their activity, including prompts, commands and file edits, into one standard format. That gives security teams one place to see what different agents did, and on whose behalf. Asymptote's paid platform builds on Beacon and can block risky actions, such as data leaks, before they happen, the company said. "AI agents are becoming a new operating layer inside every company, but the infrastructure to understand and control them hasn't caught up," said Justin D'Souza, Asymptote's co-founder and chief executive. D'Souza and co-founder Shukan Shah previously worked as engineers at Doppel, a startup that fights social engineering attacks. Beacon launched in May, and the company says its GitHub page grew its stars by more than 460% in September. Asymptote said security teams already use its tools in production, but did not name customers. Tower's own security team is testing Beacon on its computers, the company said. Investors are betting that companies need new tools as agents take on real work. Straiker raised $64M in June to test and monitor AI agents. This month, researchers said a single prompt let them take over agents on Amazon's AgentCore, though AWS disputed the findings. Get the most important tech news in your inbox each week.