I scanned 39 AI companies' DNS records — DomainIntel Research

← DomainIntel

I scanned 39 AI companies' DNS records — here's who's verified with Anthropic, who's deploying MCP keys, and whose email you can spoof

May 20, 2026 · John Leslie · All data independently verifiable with dig

17/39

Anthropic verified

MCPv1 deployers

23%

Weak email security

12

Microsoft MCP keys

DNS records are public. When a company verifies domain ownership with a service provider, that verification lives in their TXT records — visible to anyone who runs dig TXT example.com. These records form a map of vendor relationships that companies may not realize they're broadcasting.

I queried the DNS, SSL, WHOIS, and HTTP headers for 39 AI-adjacent domains — labs, safety orgs, tooling companies, prediction markets. Here's what's actually in there, verified with independent dig queries.

Anthropic domain verification: who's completed it

17 out of 39 domains contain an anthropic-domain-verification TXT record:

$ dig +short TXT meta.com | grep anthropic<br>"anthropic-domain-verification-vb762t=ezl2XXGyq8tXMTx0oArON1hm7"

$ dig +short TXT stripe.com | grep anthropic<br>"anthropic-domain-verification-zk7x9c=QfN52ECybLPUWh51R9pKF0QO3"

Meta<br>Microsoft<br>NVIDIA<br>Midjourney<br>Perplexity<br>Fireworks AI<br>Modal<br>Aleph Alpha<br>Hugging Face<br>Vercel<br>Cursor<br>Sourcegraph<br>Pinecone<br>Stripe<br>Open Philanthropy<br>Polymarket<br>Jasper AI

What this means: these companies completed Anthropic's domain verification process — likely part of Claude API enterprise onboarding. It doesn't prove an enterprise contract, but someone at these organizations generated a verification token and added it to their DNS.

Notable absences:

OpenAI<br>Google<br>Apple<br>Amazon<br>xAI

Cursor verification shows a similar pattern. 14 domains have cursor-domain-verification records, including Anthropic itself. Companies like Hugging Face, Modal, Fireworks, Pinecone, Vercel, Sourcegraph, and Stripe appear in both Anthropic's and Cursor's verification lists.

MCPv1 DNS records: cryptographic MCP server identity

This one surprised me. Multiple companies have v=MCPv1 TXT records containing public keys:

$ dig +short TXT microsoft.com | grep MCPv1 | wc -l<br>12

$ dig +short TXT microsoft.com | grep MCPv1 | head -3<br>"v=MCPv1; k=ecdsap384; p=A/Mf6IKdZzcHfBvpiVz9rkdPTIcCP5IbR..."<br>"v=MCPv1; k=ecdsap384; p=AqXeTHJ/1FCYeuvJ8dc1B+X3uHaa7m2W0..."<br>"v=MCPv1; k=ecdsap384; p=AoHTKEi2W8L2P8cf9CoDicIxYiuttTkwtI..."

$ dig +short TXT stripe.com | grep MCPv1<br>"v=MCPv1; k=ed25519; p=WMeka0C1fIH9HQLMtsSM9DD9cM6Bz6Wz34mHnK86UcM="

CompanyKeysAlgorithm

Microsoft12 ecdsap384<br>Stripe1ed25519<br>Perplexity1ed25519<br>Hugging Face1ed25519<br>Vercel1ed25519<br>Sourcegraph1ed25519

The record format (v=MCPv1; k=; p=) resembles DKIM's DNS key structure. This appears to be a DNS-based identity verification mechanism for MCP (Model Context Protocol) servers — allowing clients to verify that an MCP endpoint is actually operated by the domain owner.

I couldn't find a public specification for this record format. Microsoft's investment of 12 separate key pairs suggests multiple verified MCP services. The ecosystem is early — only 6 of 39 domains have these records.

Email security: who's vulnerable to spoofing

DMARC tells receiving mail servers what to do with messages that fail authentication. p=reject blocks spoofed messages. p=none lets them through.

No DMARC record at all

Alignment Forum<br>Manifold Markets

Alignment Forum also has no SPF record — zero email authentication.

DMARC monitoring-only (p=none — spoofed emails delivered)

MIRI<br>Hugging Face<br>Inflection AI<br>xAI<br>Aleph Alpha<br>Qdrant<br>Metaculus

MIRI also has no SPF record. xAI sends DMARC reports to Alibaba Cloud.

That's 9 out of 39 domains (23%) with weak or absent email authentication. For AI safety organizations handling sensitive research — MIRI and Alignment Forum — this is more concerning, given the sophistication of potential adversaries.

The strong side: Anthropic (p=reject), OpenAI (p=reject), Stripe (p=reject), and Cloudflare (p=reject) all have strict policies.

Smaller findings

xAI's unusual infrastructure

SSL certificate from a Chinese issuer (Guangdong Baota Security Technology). DMARC reports to Alibaba Cloud. Domain registered in 1994 (32 years before xAI was founded). Response time: 660ms.

Market share

Email: Google Workspace 74% (29/39), Microsoft 365 8% (3/39).

Hosting: Cloudflare 46%, Vercel 28%.

SSL: Let's Encrypt + Google Trust Services 72%.

Response times: 14ms (Claude.ai) to 5.8s (Meta).

Methodology

Data collected May 20, 2026. All sources are public — DNS records, SSL certificates (Certificate Transparency), WHOIS registration, HTTP headers. To verify any claim:

dig +short TXT # TXT records (verification, SPF)<br>dig +short TXT _dmarc. # DMARC policy<br>dig +short MX # Mail servers<br>echo | openssl s_client -connect :443 2>/dev/null | openssl x509 -noout -issuer -dates

Data was collected using DomainIntel, which combines DNS, SSL, WHOIS, and tech...