diff --git a/docs/05-comparatives/vs-verkada.md b/docs/05-comparatives/vs-verkada.md
new file mode 100644
index 0000000..766a705
--- /dev/null
+++ b/docs/05-comparatives/vs-verkada.md
@@ -0,0 +1,58 @@
+# vs Verkada / Rhombus
+
+## Snapshot
+
+| Dimension | Verkada / Rhombus | Blocao |
+|---|---|---|
+| Architecture | Camera → vendor cloud | Camera → on-site Cell |
+| Storage location | US/EU vendor data centers | Customer site (raw) + EU sovereign hub (metadata) |
+| Pricing | Per-camera per-month subscription | Hardware + optional support |
+| AI | Modern, cloud-side | Modern, edge-side (Frigate + custom) |
+| Sovereignty | CLOUD Act exposure | EU sovereign by design |
+| Lock-in | High (proprietary cameras + cloud) | Low (ONVIF cameras + open stack + GitOps) |
+| Open source | None | Frigate, Mosquitto, Whisper, etc. |
+| Multi-site fleet | Yes | Yes (hub) |
+| Evidence chain | Vendor claims, not verifiable | Manifest + signature, post-MVP fully transparent |
+
+## Where Verkada wins
+
+- **Polish**. Their UX is the slickest in the market.
+- **Ecosystem**. Door access, alarms, sensors integrated.
+- **Sales velocity**. Their reps close fast.
+- **First-time-user simplicity**. Plug in camera, scan QR, done.
+- **Brand recognition**. Many customers know the name already.
+
+## Where Blocao wins
+
+- **Sovereignty**. Verkada streams raw video to AWS. CLOUD Act applies regardless of customer location. For Argentine, EU and any sovereignty-conscious buyer this is a disqualifier.
+- **TCO over 3 years**. Verkada's per-camera-per-month model adds up. A 30-camera site over 3 years can cost more than the entire Blocao deployment hardware.
+- **No vendor lock-in for cameras**. Verkada cameras only work with Verkada cloud. Blocao works with any ONVIF camera, including ones the customer already owns.
+- **No vendor lock-in for data**. Verkada export is partial and lossy. Blocao's clips, embeddings, manifests are open formats.
+- **GitOps auditability**. "What changed in the config last Tuesday?" — Blocao: read the commit log. Verkada: file a support ticket, hope.
+- **Evidence chain (post-MVP)**. Verkada signs internally; not third-party-verifiable. Blocao's roadmap targets RFC 3161 + transparency log.
+
+## Where each is comparable
+
+- **AI quality**. Both run modern models on detection. Verkada's cloud has more compute per query but Blocao's edge models are competitive for surveillance use cases.
+- **Camera support**. Verkada locks to their hardware; Blocao supports any ONVIF. Each can argue their angle.
+- **Multi-site**. Both handle it. Verkada via cloud-native; Blocao via hub.
+
+## Sales conversation patterns
+
+When prospect mentions Verkada:
+
+1. **Ask: "Why are you considering them?"** Listen.
+2. If they say "ease of deployment": acknowledge — Verkada's simplicity is real. Counter with first-boot wizard demo.
+3. If they say "AI capabilities": demo Blocao's forensic query. Equal or better.
+4. If they say "we just want it to work": acknowledge — Verkada has a reliability story too. Counter with "and you keep your data".
+5. If sovereignty matters at all: lead with it. Show MQTT panel + GitOps panel. Hard to argue against demonstrable sovereignty.
+
+When prospect dismisses Verkada due to cost or sovereignty: don't pile on. Just confirm they've made a sound decision and proceed.
+
+## Honest assessment
+
+Verkada is not the enemy. They're a successful product that solves a real problem. They've validated that there's a market for "modern AI-powered video forensics for non-enterprise customers".
+
+Blocao competes on a different axis: **same value prop, sovereignty included, no vendor cloud**. We're not faster than Verkada at adding new features. We're more honest about where the data goes.
+
+For a sovereignty-conscious buyer or a TCO-conscious buyer at scale, Blocao wins. For a buyer who values polish and ecosystem above all, Verkada is reasonable. Don't fight that fight.