PQ Adoption: Crypto RPC vs. Open-Web Top-1M
Both numbers come from our own prober. The RPC rate is measured by probing every public endpoint in the RPCScan catalogue; the Top-1M rate is measured by probing the latest Tranco Top-1M list from a scheduled scan on the same box. No vendor telemetry, no proprietary logs. Repeatable from the source code and the public Tranco list.
Where the gap comes from
Cloudflare turns on post-quantum (PQ) key exchange by default for every site it serves, so nearly every URL on its network supports PQ without its operator doing anything. The bars split each population into URLs on Cloudflare’s network and URLs on any other network, and show what share of each group supports PQ key exchange. Measured 2026-10-27.
If the same share of the open web sat on Cloudflare as crypto does, the share of open-web URLs supporting PQ would be 77.8%, against crypto’s 77.3%.
Adoption trend
PQ-hybrid share over time for each population. Both series land daily, and a date is published once both measurements for it are in. Each dot is one measurement; the Y axis is percent of TLS-reachable endpoints that negotiate X25519MLKEM768(or any other ML-KEM NamedGroup).
PQ support on networks other than Cloudflare
The same trend with every URL on Cloudflare’s network left out. Those URLs support PQ by default, so they barely move. What remains are URLs whose operators had to turn PQ on themselves.
By hosting network
The ten networks hosting the most of either population on 2026-10-27, each with the share of each population’s URLs it hosts and the share of those URLs that support PQ key exchange. Crypto counts each provider domain once, so a provider with a thousand RPC hostnames counts one.
| Network | Crypto RPC | Open web | ||
|---|---|---|---|---|
| Share of domains | Support PQ | Share of URLs | Support PQ | |
| Cloudflare, Inc. | 46.3% | 97% | 39.2% | 98% |
| Amazon.com, Inc. | 14.9% | 17% | 11.0% | 52% |
| Google LLC | 5.0% | 46% | 2.0% | 57% |
| Hetzner Online GmbH | 3.8% | 28% | 2.6% | 26% |
| DigitalOcean, LLC | 3.6% | 20% | 1.0% | 11% |
| OVH SAS | 2.3% | 3% | 1.9% | 30% |
| Alibaba (US) Technology Co., Ltd. | 1.7% | 13% | 1.1% | 12% |
| Hostinger International Limited | 1.4% | 32% | 1.1% | 86% |
| Contabo GmbH | 2.2% | 6% | 0.2% | 13% |
| Microsoft Corporation | 0.6% | 22% | 1.7% | 25% |
How we measure
- Tranco list. Downloaded from tranco-list.eu, a KU Leuven / imec research project. One million domains, ranked by combined popularity.
- Probe. One TLS 1.3 handshake per domain to
:443, offering onlyX25519MLKEM768as the key exchange. Success means the server supports PQ-hybrid; failure plus a follow-up classical handshake splits “server up, no PQ” from “unreachable.” - Network owner. Each probe records the IP it reached. A public IP-to-network table (iptoasn.com, refreshed daily) maps it to a network, and CAIDA’s AS-to-Organization dataset maps the network to the company that runs it, so Amazon’s two networks count as one. “Behind Cloudflare” means a network whose organisation name starts with Cloudflare. Both populations use the same table on the same day.
- Aggregate. The daily aggregate is uploaded to this site via a HMAC-signed POST and appended to a JSONL file.
All runs
| Date | List size | TLS-reachable | PQ-hybrid | PQ rate | Elapsed | Tranco list |
|---|---|---|---|---|---|---|
| 2026-10-27 | 1,000,000 | 864,009 | 521,640 | 60.4% | 6103s | 2026-10-22 |
| 2026-10-26 | 1,000,000 | 864,105 | 517,169 | 59.9% | 6147s | 2026-10-22 |
| 2026-10-25 | 1,000,000 | 864,032 | 516,775 | 59.8% | 6106s | 2026-10-22 |
| 2026-10-24 | 1,000,000 | 863,887 | 516,194 | 59.8% | 6201s | 2026-10-22 |
| 2026-10-23 | 1,000,000 | 863,910 | 515,201 | 59.6% | 6218s | 2026-10-22 |
| 2026-10-22 | 1,000,000 | 861,976 | 513,499 | 59.6% | 6318s | 2026-10-22 |
| 2026-10-21 | 1,000,000 | 862,827 | 509,169 | 59.0% | 6014s | 2026-10-14 |
| 2026-10-20 | 1,000,000 | 862,826 | 508,294 | 58.9% | 5932s | 2026-10-14 |
| 2026-10-19 | 1,000,000 | 863,041 | 508,283 | 58.9% | 6149s | 2026-10-14 |
| 2026-10-18 | 1,000,000 | 863,091 | 508,097 | 58.9% | 5868s | 2026-10-14 |
| 2026-10-17 | 1,000,000 | 863,073 | 507,782 | 58.8% | 5950s | 2026-10-14 |
| 2026-10-16 | 1,000,000 | 863,254 | 507,900 | 58.8% | 6011s | 2026-10-14 |