Asymptote Labs
Runtime intelligence and security infrastructure for enterprise AI agents.
Website: https://asymptotelabs.ai/
Cover Block
Public sources
| Field | Value |
|---|---|
| Name | Asymptote Labs |
| Tagline | Runtime intelligence and security infrastructure for enterprise AI agents |
| Headquarters | New York, United States |
| Founded | 2025 |
| Stage | Seed |
| Business model | Open Source / Commercial |
| Industry | Security |
| Technology | AI / Machine Learning |
| Geography | North America |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (2) |
| Funding label | Seed |
| Total disclosed | ~$5,500,000 [The Next Web, October 2026] |
Links
Public sources
- Website: https://asymptotelabs.ai/
- GitHub: https://github.com/jqdsouza
Executive Summary
PUBLIC Asymptote Labs is building runtime intelligence and security infrastructure for enterprise AI agents, and it has investor attention now because the company surfaced with a $5.5 million seed round in October 2026 as enterprises begin looking for controls around agent behavior in production [The Next Web, October 2026] [Crunchbase] [tower-research.com]. The public record is still early, but the core pitch is legible: Asymptote wants to become a system of record for agent activity, starting with telemetry and security visibility rather than broad workflow orchestration [docs.asymptotelabs.ai] [asymptotelabs.ai].
The founding story appears to come out of a familiar security-adjacent arc. Justin D'Souza and Shukan Shah previously worked together at Level and Doppel, and company materials describe a market where AI agents are becoming an operating layer before governance infrastructure has caught up [Unite.AI, October 2026] [asymptotelabs.ai] [docs.asymptotelabs.ai].
The product stack combines an open-source foundation, Beacon, with commercial enterprise offerings and private deployment options, which gives Asymptote a plausible open-source-to-enterprise conversion path if adoption forms around developer instrumentation and security operations workflows [docs.asymptotelabs.ai] [asymptotelabs.ai]. Its technical distinction, based on public materials, is that agent applications are instrumented through an Observe SDK that emits Beacon-compatible OpenTelemetry spans for model calls, tool calls, and agent steps, with exports to either a customer-managed collector or hosted Observe [docs.asymptotelabs.ai].
Founder-market fit is one of the cleaner parts of the file. Unite.AI and company materials identify D'Souza as CEO and Shah as CTO, with prior experience spanning machine learning, payments infrastructure, and AI engineering, including time at Doppel, a security company that raised a $70 million Series C in November 2025 after a $35 million Series B earlier that year [Unite.AI, October 2026] [Doppel Raises $70M Series C to Meet Rising Demand for AI-Driven Social Engineering Defense, November 19, 2025] [Doppel Raises $35 Million Series B, May 2, 2025].
On capitalization, the company is at seed stage with a disclosed $5.5 million round led by Tower Research Ventures and participation from South Park Commons, Lytical Ventures, and 1984 Ventures [The Next Web, October 2026] [finsmes.com, October 2026] [Crunchbase]. The business model is also unusually clear for a young infrastructure company: open-source Beacon at the base, then managed enterprise software and private deployments higher up the stack [asymptotelabs.ai].
The main questions for the next 12 to 18 months are not about category relevance so much as proof of commercial conversion. Investors should watch for named production users, evidence that security teams adopt Beacon beyond experimentation, and signs that open-source distribution can translate into enterprise contracts before larger platform and security vendors define the control plane for agent runtime data [The Next Web, October 2026] [Unite.AI, October 2026] [tower-research.com].
Company-stated, unverified -- This section mixes independently reported funding facts from The Next Web, FinSMEs, Crunchbase, and Tower Research Ventures with material product and positioning claims that remain primarily company-sourced.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | Open Source / Commercial |
| Industry / Vertical | Security |
| Technology Type | AI / Machine Learning |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | Seed, total disclosed ~$5,500,000 |
How the Company Got Here
PUBLIC
Asymptote Labs presents itself as infrastructure for a problem that is arriving before most enterprises have standard controls for it: AI agents are beginning to act like an operational layer, while the system of record for what those agents do is still thin [asymptotelabs.ai] [docs.asymptotelabs.ai]. Public company materials describe Asymptote as a New York based startup founded in 2025, operating as Asymptote Labs, Inc., and focused on runtime data, visibility, and governance for enterprise AI agents [Crunchbase] [asymptotelabs.ai] [asymptotelabs.ai, Unknown].
The public milestone record is short but clear. The company launched its web and documentation presence with a positioning centered on "runtime intelligence" and "AI Detection & Response" for agent systems, then expanded that framing into a documented product progression that spans open-source Agent Beacon, a managed enterprise offering, and private deployment options [asymptotelabs.ai] [docs.asymptotelabs.ai] [asymptotelabs.ai]. Crunchbase lists the business at seed stage, and the clearest external financing milestone is a $5.5 million seed round announced in October 2026, which aligns with the company's emergence into broader public view [Crunchbase] [asymptotelabs.ai] [The Next Web, October 2026].
Lightly corroborated -- Core company facts are corroborated by Crunchbase and company materials, but legal entity detail and chronology remain primarily company-sourced.
Product and Technology
MIXED
The public record points to a product built around visibility first, then control. Asymptote describes itself as a runtime data platform for AI agents and, in its documentation, as an "AI Detection & Response layer for every agent harness" that gives security teams runtime visibility and governance over agent activity at scale [asymptotelabs.ai] [docs.asymptotelabs.ai]. The company’s October 2026 funding coverage is directionally consistent with that framing: The Next Web reported that Asymptote raised $5.5 million to help enterprises track and control what AI agents do, while Unite.AI described the platform as software to monitor and control enterprise AI agents [The Next Web, October 2026] [Unite.AI, October 2026].
The underlying architecture is described in more concrete terms on Asymptote’s documentation site. The company says agent applications, services, workers, and serverless functions are instrumented with its Observe SDK, which emits Beacon-compatible OpenTelemetry spans covering model calls, tool calls, and agent steps; those spans can be exported either to a customer-managed collector or to hosted Observe [docs.asymptotelabs.ai]. The same documentation states there is no endpoint agent in that path, which matters because it narrows what the product is today: based on public materials, this is an instrumentation and telemetry layer rather than a device-resident security agent [docs.asymptotelabs.ai].
Asymptote also presents the product as a tiered offering rather than a single package. Its overview page says the system is built as a progression, and its pricing page lists open-source Agent Beacon, managed enterprise Asymptote, and private deployment options for AI agent endpoint security [docs.asymptotelabs.ai] [asymptotelabs.ai]. Beacon itself is described in company materials as founder-built open source, with support for collecting agent activity across Claude Code, Cursor, Codex, OpenCode, and more than 20 other harnesses, creating a shared history across local, CI, and cloud environments [docs.asymptotelabs.ai]. The public hiring footprint adds a modest implementation signal, though any stack read is still tentative: current roles include founding security engineering, security research, developer relations, and engineering positions in New York, all on-site (inferred from job postings) [asymptotelabs.ai].
Company-stated, unverified -- This section relies heavily on company documentation and website claims, with partial framing support from The Next Web and Unite.AI.
Where the Demand Sits
PUBLIC
The market matters now because enterprise use of AI agents appears to be advancing faster than the control plane around them, which creates a narrow but timely opening for security and observability vendors that can become a system of record for agent activity [The Next Web, October 2026] [asymptotelabs.ai, Unknown].
Public evidence on category size is still thin, so the cleaner approach is to frame Asymptote against adjacent markets rather than force a precision TAM that the source set does not support. The company itself describes AI agents as a new operating layer inside companies and argues that most organizations still cannot see what those agents are doing [asymptotelabs.ai, Unknown]. Third-party coverage is directionally consistent: The Next Web reports that Asymptote is built to track and control enterprise AI agents, while Unite.AI describes Beacon and the broader platform as tools for monitoring and controlling agent behavior inside enterprises [The Next Web, October 2026] [Unite.AI, October 2026]. That does not establish a quantified core market, but it does establish the problem statement.
Demand drivers in the public record are practical rather than speculative. Asymptote's documentation says the platform is designed to instrument agent applications, services, workers, and serverless functions, and to emit Beacon-compatible OpenTelemetry spans for model calls, tool calls, and agent steps [docs.asymptotelabs.ai, Unknown]. Beacon is also described as creating a shared history of how agents work across local, CI, and cloud environments and supporting more than 20 agent harnesses, including Claude Code, Cursor, Codex, and OpenCode, according to the Beacon documentation cited in the research set [Welcome to Beacon - Beacon]. If that multi-environment, multi-harness sprawl is becoming normal in enterprise deployments, then the near-term budget line likely sits at the intersection of security operations, observability, and platform engineering rather than in a standalone "agent software" budget.
The adjacent markets are easier to name than the core market. On the product side, Asymptote sits next to observability tooling because it relies on telemetry collection and OpenTelemetry-compatible spans, next to cloud and application security because it emphasizes detection, investigation, and response, and next to developer infrastructure because its open-source Beacon layer can serve as an adoption wedge [docs.asymptotelabs.ai, Unknown] [asymptotelabs.ai, Unknown]. On the budget side, that means the company could be competing for dollars that might otherwise go to SIEM, cloud workload security, developer security, or internal platform tooling, depending on who owns agent deployments inside the customer organization [The Next Web, October 2026] [tower-research.com].
| Market framing | Public evidence | Implication for Asymptote |
|---|---|---|
| Enterprise AI agent control plane | Asymptote and press coverage describe a need to track, monitor, and control agent activity [The Next Web, October 2026] [Unite.AI, October 2026] | Supports a category thesis, but not a quantified standalone TAM |
| Observability infrastructure | Asymptote documentation centers on telemetry, runtime visibility, and OpenTelemetry spans [docs.asymptotelabs.ai, Unknown] | Suggests buyer overlap with platform and observability teams |
| Security operations and detection | The company positions itself as an AI Detection & Response layer and emphasizes governance at scale [docs.asymptotelabs.ai, Unknown] | Suggests budget overlap with security teams and incident response tooling |
| Open source to enterprise conversion | Beacon is positioned as an open-source foundation, with managed and private deployment options on the pricing page [asymptotelabs.ai, Unknown] | Indicates a familiar OSS-to-commercial path if adoption is real |
The evidence points to a market that is real in problem terms but still early in budget formalization. That is often where new infrastructure companies can win, but it also means category boundaries remain unstable and purchase decisions may be routed through adjacent tools before a dedicated line item exists.
Macro and regulatory forces are present mostly as pressure to improve accountability rather than as a clearly cited compliance catalyst in this source set. Security teams are already the stated target user, and the product language stresses visibility, governance, and traceability of agent actions at scale [docs.asymptotelabs.ai, Unknown] [asymptotelabs.ai, Unknown]. For investors, the practical read is that adoption may be helped by internal governance mandates around AI use even before external regulation translates into a distinct software budget. That is supportive, but still one step removed from proof of durable spend.
Company-stated, unverified -- The section's core market framing is grounded in company materials and a small set of October 2026 news reports, with no independent third-party TAM study or confirmed market sizing data in the source set.
Competitive Landscape
MIXED
Asymptote is positioning itself against a messy set of alternatives, not a single direct rival: open-source agent observability tools, enterprise security controls that could expand into agent runtime monitoring, and adjacent life-sciences and automation incumbents that appear in database competitor lists but do not map cleanly to the company’s stated product surface [Asymptote, Unknown] [docs.asymptotelabs.ai, Unknown] [Crunchbase, Unknown].
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Asymptote Labs | Runtime data platform and "AI Detection & Response layer" for enterprise AI agents | Seed, $5.5M disclosed in October 2026 | Open-source Beacon foundation plus commercial and private deployment options for agent telemetry, detection, investigation, and response | [The Next Web, October 2026] [Asymptote, Unknown] [docs.asymptotelabs.ai, Unknown] |
| Charles River Laboratories International | Listed as a competitor in structured-company database coverage | Public company | Differentiator versus Asymptote is not established in the cited materials | [Crunchbase, Unknown] |
| Cytena | Listed as a competitor in structured-company database coverage | Stage not established in the cited materials | Differentiator versus Asymptote is not established in the cited materials | [Crunchbase, Unknown] |
| Lonza Group | Listed as a competitor in structured-company database coverage | Public company | Differentiator versus Asymptote is not established in the cited materials | [Crunchbase, Unknown] |
| Miltenyi Biotec | Listed as a competitor in structured-company database coverage | Private company | Differentiator versus Asymptote is not established in the cited materials | [Crunchbase, Unknown] |
The table itself shows the core problem with the public record. On the evidence available, the more useful competitive map is segmental: direct substitutes are likely to be other observability and security layers for agent workflows, adjacent substitutes are general-purpose security platforms and internal tooling, and incumbents are the existing enterprise systems that security teams already use to monitor software activity, even if the cited materials do not name them directly [The Next Web, October 2026] [docs.asymptotelabs.ai, Unknown].
Asymptote’s edge today appears to rest less on scale and more on product shape. The company has an open-source foundation in Beacon, says it supports telemetry across Claude Code, Cursor, Codex, OpenCode, and more than 20 other harnesses, and offers managed enterprise and private deployment variants on top of that base [Welcome to Beacon - Beacon] [Asymptote Labs, Pricing] [docs.asymptotelabs.ai, Unknown]. If that cross-harness compatibility is real in production use, it gives the company a credible wedge with security teams that do not want fragmented visibility across agent environments; the durability of that edge is still uncertain because the same integration layer could be copied by larger security vendors or absorbed into the agent platforms themselves [Welcome to Beacon - Beacon] [Asymptote, Unknown].
The exposure is easier to identify than the moat. Asymptote is early, company-disclosed customer evidence remains thin, and even supportive publisher coverage notes that named customers were not disclosed publicly [The Next Web, October 2026] [Unite.AI, October 2026]. That leaves it vulnerable both to adjacent incumbents that already own the security budget and to platform vendors that control the agent harnesses Beacon is built to observe; among the named database competitors, Lonza Group has the clearest scale advantage as a large incumbent organization, but the cited materials do not establish that it competes in the same functional category, which limits how much weight that comparison can bear [Crunchbase, Unknown].
Over the next 18 months, the most plausible competitive outcome is a land grab around becoming the system of record for agent activity before that layer consolidates. Asymptote is the winner if enterprise buyers decide they need a dedicated, cross-harness control plane and continue to adopt open-source entry points before expanding into managed deployments; Beacon’s broad telemetry claim is the key fact to test in that scenario [Welcome to Beacon - Beacon] [Asymptote Labs, Pricing]. Asymptote is the loser if the market decides agent monitoring belongs inside existing security stacks or first-party harness vendors, because the company does not yet appear to own a channel, customer base, or regulatory position that would force adoption on its own terms [The Next Web, October 2026] [tower-research.com].
Single unverified source -- Core company positioning is corroborated by company materials and publisher coverage, but the named competitor set used here comes primarily from Crunchbase taxonomy rather than direct competitive reporting.
Opportunity
PUBLIC
If Asymptote executes, the prize is not a niche security tool but a control plane for how large companies observe, govern, and secure AI agents as those agents become operational software inside the enterprise [Asymptote, Unknown] [The Next Web, October 2026].
The largest plausible outcome is that Asymptote becomes the default runtime intelligence layer for enterprise AI agents, first through security and observability, then through broader operational infrastructure. That is reachable, not merely aspirational, because the company is already framing the problem in system-of-record terms, offering both open-source Beacon and managed enterprise deployment options, and targeting a buyer with budget and urgency, namely enterprise security teams that need visibility into heterogeneous agent activity [docs.asymptotelabs.ai, Unknown] [Asymptote Labs, Unknown] [Asymptote, Unknown] [The Next Web, October 2026]. The October 2026 seed round, $5.5 million led by Tower Research Ventures with participation from South Park Commons, Lytical Ventures, and 1984 Ventures, gives some external validation that investors see this as infrastructure rather than a point feature [The Next Web, October 2026] [finsmes.com, October 2026].
The evidence also suggests the company is making an architecture choice that can travel across environments rather than sit only at the endpoint. Asymptote's documentation says agent applications, services, workers, and serverless functions are instrumented through an Observe SDK that emits Beacon-compatible OpenTelemetry spans for model calls, tool calls, and agent steps, exporting either to a customer-managed collector or hosted Observe [Asymptote, Unknown]. If enterprise buyers standardize on that telemetry path, the company has a route into the runtime layer where policy, detection, investigation, and governance can compound over time [docs.asymptotelabs.ai, Unknown] [Asymptote, Unknown].
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Security system of record for agents | Asymptote becomes the standard runtime visibility and response layer used by security teams to monitor agent behavior across internal tools, coding agents, and automated workflows. | A category shift in enterprise AI adoption that pushes CISOs to require unified auditability and controls for agent actions. | The company already positions itself as an "AI Detection & Response layer" and says Beacon captures activity across Claude Code, Cursor, Codex, OpenCode, and more than 20 other harnesses, which fits a cross-environment control-plane strategy [docs.asymptotelabs.ai, Unknown] [Welcome to Beacon - Beacon]. |
| Open-source wedge to enterprise platform | Beacon adoption creates developer familiarity, then managed Asymptote and private deployment convert that usage into enterprise contracts. | Open-source traction that makes Beacon the default instrumentation base for teams deploying agents in production. | Pricing and documentation show a deliberate progression from open-source Agent Beacon to managed enterprise Asymptote and private deployment, which is a recognizable open-core path when the production buyer differs from the initial user [Asymptote Labs, Unknown] [Asymptote, Unknown]. |
| Embedded security layer for agentic coding | Asymptote wins inside AI-assisted software development and becomes part of the standard stack for coding agents and agent-enabled engineering workflows. | Continued growth of agentic coding, plus demand for security review and telemetry around model calls, tool calls, and agent steps. | Asymptote's own product language emphasizes the "security layer for agentic coding," and its architecture is already tailored to instrument model calls, tool calls, and agent steps, which matches that workflow directly [Asymptote, Unknown] [Asymptote, Unknown]. |
The compounding mechanism here is straightforward. More instrumentation across more agent harnesses should produce a broader runtime history, and a broader runtime history makes detection, investigation, and governance more useful, which in turn gives security teams a stronger reason to standardize on the same telemetry layer across local, CI, and cloud environments [Welcome to Beacon - Beacon] [Asymptote, Unknown]. That is not a classic consumer network effect, but it can become infrastructure lock-in if policies, workflows, and investigations are built on top of the telemetry stream.
There are early signs of that flywheel design in the product surface. Beacon is open source, the managed enterprise product exists alongside it, and private deployment is already part of the packaging, which suggests the company is trying to meet developers, security teams, and stricter enterprise procurement requirements in one progression rather than through disconnected products [Asymptote Labs, Unknown] [docs.asymptotelabs.ai, Unknown]. Tower Research Ventures said its investment interest began with Beacon's open-source traction and that Tower's security organization is evaluating the technology on its own endpoints, according to Unite.AI, which is limited evidence but directionally consistent with a land-then-standardize motion [Unite.AI, October 2026].
The upper bound is easiest to frame through category ownership rather than near-term revenue math. If AI agents become a durable enterprise software layer, the control point that records what those agents did, how they did it, and whether they stayed within policy could resemble prior security and observability winners in strategic value, even if the exact market size is still unsettled. On a scenario basis, not a forecast, if Asymptote were to become a widely adopted infrastructure layer for enterprise agent security and runtime governance, the outcome could support a multibillion-dollar enterprise value, because the company is positioning at the intersection of security, observability, and AI platform infrastructure, and investors are already financing that thesis at formation stage [The Next Web, October 2026] [finsmes.com, October 2026] [Asymptote, Unknown]. The missing piece in public evidence is market sizing from a named industry report, so the upside case rests on architectural position and buyer urgency more than on a published TAM figure.
Company-stated, unverified -- This section relies heavily on company materials and product documentation, with partial corroboration from The Next Web, FinSMEs, and Unite.AI for financing and market framing.
Sources
Public sources
[The Next Web, October 2026] Asymptote Labs raises $5.5M to track and control what AI agents do | https://thenextweb.com/news/asymptote-labs-5-5m-seed-ai-agent-security
[Crunchbase] ASYMPTOTE LABS - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/asymptote-labs-ltd
[tower-research.com] Tower Research Ventures Invests in Asymptote Labs | https://tower-research.com/asymptote-labs-the-system-of-record-for-agentic-work/
[docs.asymptotelabs.ai] Welcome to Asymptote - Asymptote | https://docs.asymptotelabs.ai/
[asymptotelabs.ai] Asymptote: The Runtime Intelligence Layer for AI Agents | https://asymptotelabs.ai/
[Unite.AI, October 2026] Asymptote Labs Raises $5.5M Seed to Monitor and Control Enterprise AI Agents | https://unite.ai/asymptote-labs-raises-5-5m-seed-to-monitor-and-control-enterprise-ai-agents/
[finsmes.com, October 2026] Asymptote Labs raises $5.5M in Seed Funding | https://www.finsmes.com/2026/10/asymptote-labs-raises-5-5m-in-seed-funding.html
Articles about Asymptote Labs
- Asymptote Labs Raises $5.5M to Put a Security Layer on AI Agents — The New York startup is betting that enterprise security teams will pay to monitor and control a new wave of autonomous software.