Seed Round • Making AI Insurable

GLACIS

Proof Infrastructure for AI Systems

Cryptographic attestation that AI guardrails actually executed

Cloudflare Launchpad AI2 Incubator

[email protected] • glacis.io

The Thesis

Every Business Transformation Creates an Audit Industry

Financial Complexity → Big 4 & SOX

IT Complexity → SOC 2 & ISO 27001

Cyber Risk → Pen Testing & Security Audits

AI Risk → GLACIS

"AI is becoming the execution layer for material decisions."

Clinical recommendations. Loan approvals. Legal research. Autonomous driving. Material risks require independent verification.

The Standard of Care Already Exists

NIST AI RMF and ISO 42001 define acceptable community practice. Your legal team needs to point to it. GLACIS provides the proof you're meeting it.

The Market

$800B+ Enterprise AI Spend by 2030

Platform TAM

TAM $24-40B

AI Governance & Compliance (3-5% of AI spend)

SAM $4.5B

Regulated Industries (Healthcare + FinServ + Legal)

Beachhead $1.4B

Healthcare AI Governance by 2030 (35% CAGR)

Why Healthcare First

Hardest Compliance = Best Proof Point

$10.9M avg breach cost. Strictest regulation (HIPAA, FDA). Clearest liability. Win here, and every other vertical follows.

Adjacent: AI Liability Insurance

$1.2B

→

$15-25B

Today → 2030

The compliance tax is coming. Just like 3-5% of IT spend goes to security/compliance, the same pattern will emerge for AI.

Platform Vision

Healthcare Beachhead → Infrastructure Platform

Year 1-2

Healthcare

Clinical AI, ambient scribes, diagnostic support, prior auth

$1.4B SAM

→

Year 3-4

Financial Services

Trading algorithms, credit decisions, fraud detection, robo-advisors

$1.8B SAM

→

Year 4-5

Legal & AV

Legal research AI, contract review, autonomous vehicles

$1.3B SAM

→

Year 5+

All Enterprise AI

The "SSL Certificate for AI" — every regulated AI decision

$24-40B TAM

"Why start with healthcare?"

Hardest compliance environment = strongest proof point. HIPAA and FDA create the most rigorous test case. If we can prove compliance here, every other vertical is easier.

Regulatory Forcing Functions

Jun 2026: Colorado AI Act ($20K/violation)
Aug 2026: EU AI Act (full enforcement)
Jan 2027: California ADMT (healthcare)

The Problem

AI Governance Today: Trust Me, Bro

"We have guardrails."

Every AI vendor says this. Zero can prove it. Healthcare AI companies like Hippocratic AI, Nabla, and Abridge all claim "grounded, verifiable insights" — but have no independent proof the guardrails actually executed.

Champion Decay

400-question security surveys. Months of back-and-forth. By the time IT clears it, the physician who wanted it has lost interest.

Zero Visibility

Operators submit a ticket for software review and have zero visibility into where it stands. No one advocates. Deals stall.

$10.9M

Avg Healthcare Breach

18mo

Avg Procurement Cycle

The Gap

Logs ≠ Proof

Mutable logs can be edited. Screenshots can be faked. Post-hoc monitoring only tells you what was supposed to happen — not what did happen.

The real problem: It's not missing guardrails. It's missing proof that guardrails executed — independent, tamper-evident, cryptographic proof.

The Solution

A Digital Flight Recorder for AI Systems

Customer VPC / Trust Boundary

Application

GKE / ECS / Lambda

⟷

GLACIS

PII Redaction

Policy Check

Receipt Sign

⟷

AI Models

OpenAI / Anthropic / Vertex

↓

Witness Network

Attestation anchoring & audit export (Zero PHI egress)

<50ms

Latency

Zero

PHI Egress

∞

Providers

Competitive Moat

The Switzerland Position

"Can AWS Bedrock Guardrails attest that my team never leaked PHI to OpenAI?"

No — hyperscalers only see their own traffic. Multi-cloud governance requires independent attestation.

Runtime proof

~ Post-hoc logs

✗ Policy docs

✓ Cryptographic

Cross-provider visibility

✗ Own cloud only

~ Post-hoc logs

✓ All providers

Cryptographic attestation

✗

✗ Screenshots

✓ Tamper-evident

Independent verification

✗ Self-attest

~ Centralized

✓ Third-party

Why hyperscalers can't replicate: Conflict of interest. AWS can't independently attest that AWS guardrails worked. Regulators and insurers require third-party verification.

Self-attestation fails the insurance test.

Independent verification becomes mandatory when liability is on the line.

The Business

The Insurance Flywheel

Regulators define the floor. Insurers set the price.

$1.2B

→

$15B+

AI Liability Insurance (Today → 2030)

1. Health systems install to unblock procurement
2. Usage scales with inference volume (Direct Revenue)
3. Data proves reduced risk profile to carriers
4. Insurers mandate usage for coverage (The Moat)

"Vanta scales with headcount. GLACIS scales with inference volume."

Unit Economics

Growth Tier 85%+ GM

$499 platform fee + $0.002/inference. Lightweight WASM at the edge — we capture value without compute weight.

Enterprise Tier

$50K-200K annually. Custom policy engines, dedicated witness nodes, carrier integrations, compliance exports (OSCAL, ISO 42001).

Endgame: GLACIS attestation becomes the FICO score for AI risk. Premium discounts for attested systems.

Traction

Early Momentum

2

Design Partners Confirmed

Converting to paid pilots Q1 2026

Healthcare AI vendors with active PHI leakage concerns and 18-month procurement cycles blocking deployments

What We've Built

✓

Working platform (~40ms latency, crypto attestation)

✓

Core IP filed (Fenwick & West)

✓

OSSP open standard published

✓

Delaware C-corp, data room ready

Backing

AI2 Incubator ($600K) • Cloudflare Launchpad • Sourdough Ventures

Healthcare Pipeline

✓

2 Design Partners: Healthcare AI vendors — converting to paid Q1

→

HLTH 2025: 15+ qualified conversations with ambient scribe vendors

→

EHR Vendors: Active PHI leakage concerns driving urgency

→

Insurance carriers (Lloyd's, Hartford)

Customer Archetype

Healthcare AI companies claiming "grounded, verifiable insights" need independent proof those claims are true. GLACIS provides the attestation layer they can't build themselves.

Team

Built by People Who Lived the Problem

Joe Braidwood CEO

Scaled SwiftKey to 1B+ devices

SwiftKey: Founding exec. $250M exit to Microsoft.

Cambridge Law • Healthcare AI experience

Dr. Jennifer Shannon Co-Founder & Chief Medical Officer

FDA Authorization for AI

Cognoa: Medical Director, first FDA-authorized AI diagnostic.

CHAI Coalition: Drafting standard of care for AI governance.

Caer Sanders Principal Engineer

Rust/WASM systems at billion-scale

WPI Robotics Faculty • Wayfair: Staff engineer, <10ms ML infra

Patents: Edge-deployable cryptographic consensus.

Advisors

Selvan Senthivel • GE Healthcare Chief Technologist

Nakis Urfi, JD, MPH • Cantex CCO (37 facilities)

David Márton • Harvard AI Research

Brett Murray • Nvidia, Walmart, Nokia C-Suite

John Ryley • Former Head, Sky News

Why this team: We've scaled consumer AI to a billion devices, navigated FDA medical device authorization, and seen firsthand how health systems evaluate AI vendors. GLACIS exists because we lived the compliance gap.

Network Access

CHAI Coalition connections to 50+ health systems. AI2 portfolio introductions. Cloudflare enterprise distribution.

The Ask

Seed Round

40%

Engineering

Core platform & integrations

20%

Coverage

Coverage & integrations

25%

GTM

Design partner conversion

15%

Legal/Ops

IP & carrier partnerships

Month 6

3+ Live Pilots

Month 12

1M+ Attestations

Month 15

First Carrier Integration

Month 18

$1.5M ARR

~12 enterprise customers

2

Confirmed Partners

$600K

Already Raised

18mo

Runway Target

In 18 months, you won't deploy AI in regulated industries without independent proof of control.

Seattle-based. AI2-backed. Building the compliance infrastructure the world needs.

[email protected] • glacis.io