Hong Kong

What's verified

The Hong Kong Monetary Authority (HKMA) launched the "Whitepaper on Quantum Preparedness of Hong Kong's Banking Sector" together with the first Quantum Preparedness Index (QPI), on 27 July 2026. The initial sector-wide QPI score is 2.3 on a 10-point scale (assessed across four dimensions: Awareness, Planning, Pilots, and Practical Preparedness), reflecting an early stage of preparedness. The HKMA aims to support the banking sector in achieving full sectoral readiness (a QPI score of 10) by 2030 through practical guidance, training, and industry engagement. Thirty-two percent of surveyed banks had not begun. One of the industry's own top-three asks, recorded in the whitepaper, was for "Clear Supervisory Expectations & Timelines" — a request for rules that do not yet exist.

The binding rule still says RSA-2048, and never mentions quantum

Hong Kong's actual mandatory cryptography standard is not the HKMA whitepaper. It is the OGCIO / Digital Policy Office "IT Security Guidelines" [G3], binding on government bureaux and departments, whose current version 10.2 dates from April 2025. It sets the floor in the language of obligation:

"For information classified as CONFIDENTIAL or above, the symmetric encryption key length shall be at least 128-bit for the AES encryption or equivalent. Subject to operational needs, the requirement can also be met by SM4. Whereas the asymmetric encryption key length shall be at least 2048-bit for the RSA encryption. Alternatively, the requirement can be met by Elliptic Curve Cryptography (ECC) encryption with a key length of at least 224-bit or equivalent, as well as SM2 subject to operational needs."

We searched the full document: it contains no occurrence of "quantum" or "post-quantum." RSA-2048 and ECC-224 — precisely the primitives every other jurisdiction on this tracker is planning to retire — remain Hong Kong's binding minimum as of its most recent revision, with no migration provision attached. The guidelines also permit the Chinese national algorithms SM2, SM3, and SM4 as alternatives.

That is the gap worth understanding about Hong Kong. The quantum work is real but advisory and confined to banking; the instrument that actually carries "shall" has not been touched by it.

What the whitepaper does and does not do

The HKMA whitepaper names no approved algorithm and sets no parameter set. ML-KEM, ML-DSA, SLH-DSA, XMSS, LMS, and HQC appear in it only in the appendix glossary, as expansions of acronyms, and the document contains no reference to FIPS at all. It specifies no deprecation and no key-size threshold. Its Figure 5 tabulates government PQC migration deadlines for Canada, the United States, Australia, the EU, Singapore, India, the United Kingdom, and Germany — Hong Kong is not a row in its own table. On hybrid deployment it is neutral rather than directive, noting that "hybrid encryption has additional overheads." Its obligations on authorized institutions are consistently phrased "should" or "may."

It does, however, take a real position on cryptographic inventory. The QPI's four dimensions break down into named sub-indices, one of which is "Cryptographic Discovery and Inventory," and one of the whitepaper's top recommendations is to "commission a cryptographic inventory, accepting that it will be imperfect at first and improve with iteration." It names the artifact directly: "a comprehensive cryptographic inventory, sometimes referred to as a CBOM, provides the foundation for assessing exposure, prioritising remediation activities, and tracking transition progress." The baseline is thin — most institutions surveyed still rely on spreadsheets or have no formal discovery process at all — but the recommendation itself is explicit, named, and measured, which is more than several binding instruments elsewhere on this tracker manage.

The clearest statement of the whitepaper's own posture is in its foreword: "the HKMA's approach is to encourage the industry to begin early." That is advisory language from a real supervisor to a defined, supervised sector, not a bystander's commentary — which is why this scores above zero on mandate strength despite carrying no enforcement mechanism.