MOUNTAIN THEORY VS UPWIND
New here? The short version: Mountain Theory checks each action an AI is about to take, against rules you write in plain English, before the action runs. Everything below compares that with what Upwind does.
Upwind is a runtime-first cloud security platform that added AI protection on top of its existing cloud context. The question is whether cloud runtime context is the same thing as an execution gate.
| Upwind | Mountain Theory | |
|---|---|---|
| What it controls | Cloud infrastructure and AI behavior | The action an AI agent is about to take |
| Where it sits | Cloud runtime, tracing | Inline at execution, between the decision and the action |
| How policy is set | Cloud posture and detection rules | Plain English, no code |
| Deployment reach | Code-to-cloud via eBPF | Model and framework agnostic, including custom and on-prem agents |
| Best fit when | You want cloud and AI posture from one vendor | An AI acting wrongly has physical or regulatory consequences |
Why you might pick Upwind
Upwind has a battle-tested CNAPP already deployed in complex enterprise cloud environments, and has been growing quickly. Their AI security rides on existing runtime cloud context, so buyers get cloud and AI posture from one vendor on one contract, with deep code-to-cloud visibility via eBPF.
Why you might pick Mountain Theory
Upwind secures cloud infrastructure and traces AI behavior for investigation and remediation, which is posture and detection. Mountain Theory is the inline gate that blocks the unauthorised action before it executes. Upwind tells you what an agent did across your cloud. Mountain Theory stops what it is about to do.
The honest verdict
Upwind gives you excellent forensics: which agent touched which resource, traced across your cloud, after the fact. When the question is why did our agent drop a production table last night, Upwind answers it thoroughly. Mountain Theory answers a different question, which is whether it drops the table at all. Detection tells you what happened; a gate decides whether it does.
What we can actually show
Proof
Four things with a published run behind them.
Claims in this category are easy to make and hard to check, so here is ours on the record. The same 10 actions were run in the same order under three configurations. Ungoverned, 10 of 10 executed. Under NVIDIA OpenShell alone, all 5 sandbox-boundary crossings were denied at the kernel, and all 3 in-bounds bad decisions still went through, including a secrets read that printed credentials to the screen. Under OpenShell plus Mountain Theory, those same 3 actions returned HOLD, HOLD and BLOCK, and the secrets read was stopped before it executed, so the credentials never printed. Terminal recordings of all three runs are published, including the two actions Mountain Theory has no policy for.
Separately, when a third-party provider updated the foundation model driving an autonomous agent, the agent began attempting multi-step actions it had never tried before. Nothing on our side changed. Every attempt was stopped on 30 and 31 July 2026, the days the behavior first appeared. No new rule, no signature, no patch.
Both products in this comparison operate at the execution layer. What execution-layer security is, and how it differs from prompt filtering and from access control.
Watch the three-configuration run against NVIDIA OpenShell
See novel agent behavior stopped the day it appeared
Ask Upwind, and every other vendor you are evaluating, for the same four things: the exact action set, the ungoverned control condition, the outcome per action including the ones the product did not stop, and the recording. A certification, an integration list or a customer logo answers a different question.
Compare all 61 AI security vendors