Killed By Claude Report
← HomeCurrent verdict
Assessment
Anthropic is moving toward Lunar's world, but it is not fully replacing Lunar's core job today.
The real overlap is around:
- MCP connectors,
- enterprise admin controls,
- agent orchestration across tools,
- and growing concern about AI agent autonomy and security.
But Lunar sells a cross-vendor, enterprise-wide governance layer. Anthropic mostly ships capabilities for using Claude more deeply inside enterprises, not an independent control gateway governing all AI apps and agents across the stack.
So Lunar is exposed, but not dead. If Anthropic expands from product admin controls into a serious neutral security/governance plane for all agent traffic, the heat goes up fast.
Biggest historical hit
The clearest hit is Introducing Sonnet 4.6.
That release explicitly brings MCP connectors into Claude surfaces and makes those connections reusable across products like Claude.ai and Excel. That matters because Lunar's wedge is governing MCP-connected tools and AI interactions across enterprise systems.
When Anthropic bakes tool connectivity directly into Claude's enterprise product, some buyers will ask the annoying question: do we need a separate gateway, or can we just manage this in Claude?
That is not a full replacement of Lunar's governance layer, but it is the most concrete encroachment in the evidence pack.
What still protects them
Lunar still has a real moat if it executes.
So yes, there is overlap. But Lunar is protected by being the adult supervision layer rather than the model itself.
Signals
Why this is in the blast radius
This is the most direct overlap.
Anthropic says Claude now supports MCP connectors across products, including Excel, and reuses those connections automatically. Lunar explicitly positions around governing MCP-connected tools and AI interactions across enterprise systems.
Anthropic is not offering Lunar's full cross-vendor gateway here, but it is absolutely eating into the same workflow surface: connecting AI to enterprise tools and making that usable at scale.
Lunar sells centralized oversight, governance, and monitoring for enterprise AI usage.
This Anthropic enterprise update adds RBAC, spend controls, usage analytics, and telemetry for Claude deployments. That does not replace a neutral gateway controlling third-party AI apps and internal agents, but it does cover part of the admin/governance story that security and IT teams care about.
In other words: not full substitution, but enough to make procurement conversations messier.
Lunar is built around controlling what agents can do across live enterprise tools and systems.
Vercept pushes Claude deeper into multi-step actions inside live applications. The more Claude can directly operate software like a human, the more enterprises need governance over those actions.
This does not replace Lunar today. But it expands the exact problem surface Lunar is trying to own, while giving Anthropic a plausible path to ship more native guardrails and permissions over time.
This is not a product launch, but it is strategically relevant.
Anthropic is studying how much autonomy people grant agents, where they are deployed, and what risks they pose. Lunar's entire pitch is basically: enterprises are deploying agents faster than they can control them.
The overlap is thematic and signals future product direction around agent governance and safety. It is not a substitute by itself, but it points the gun in Lunar's direction.
This release matters more indirectly than directly.
It emphasizes stronger cybersecurity abilities, new safeguards, and better developer control on the Claude Platform. That reinforces Anthropic's credibility with enterprise security buyers.
But Lunar is not a cyber-defender model vendor. It is an AI governance gateway. Better models and platform controls from Anthropic do not by themselves replace an enterprise-wide enforcement layer spanning internal agents, third-party apps, and MCP servers.