Quantum Unfiltered #12 — Five Documents Just Completed the US Federal PQC Mandate
In three weeks the United States went from four years of "prepare and plan" to a dated, enforceable migration mandate across civilian, defense, and contractor systems. Here is the whole map, plus a co
In this edition: When I led the last edition with the cryptography executive order, it was the only document on the table. Since then the rest of the architecture has landed, and a memorandum signed before the order turns out to be the piece that holds the whole thing together. I walk through the three-layer structure, built in a deliberate order: NSPM-12 as the governance keystone, signed ten days before the executive order with its role in the framework clear only now; the executive order and the Department of War strategy as the mandates; and OMB M-26-15 as the implementation playbook that arrived in two days when the order allowed ninety. I cover why the same 2030 and 2031 dates mean three different things depending on which document you read, why SLH-DSA is approved for civilian systems but banned from National Security Systems, where the three words “unless otherwise noted” put a 2033 deadline back on the table, and why the FAR contractor rule is the part of all this most likely to reach your organization whether or not you sell to the government. The companion quantum innovation order gets its due, the shared certification bottleneck gets named, and the Flapdoodle covers the quantum substitutes the Pentagon just declared out of bounds. Also: the official Federal Register numbers for the EOs are in, and are not what the whole media published the last few days.
What the Executive Order Started, Four Documents Finished
For four years the federal post-quantum framework told agencies to get ready and never told them when to be done. The pieces were real. NIST published three PQC standards in August 2024, the Quantum Computing Cybersecurity Preparedness Act created a statutory obligation, NSM-10 set a 2035 risk-mitigation goal, and OMB M-23-02 imposed planning deadlines for inventories and reporting. What was missing was a dated, enforceable line telling agencies when migration itself had to finish. June 2026 filled that gap, and it did so with five documents that interlock rather than overlap.
Read together, they form a three-layer architecture assembled in a deliberate order: governance first, mandates second, implementation third. The governance layer is NSPM-12, signed June 12, which overhauled cybersecurity governance for National Security Systems for the first time in 35 years. It re-established the Committee on National Security Systems with the power to issue binding directives to every agency that operates an NSS, and it named the NSA Director as National Manager. That memorandum has nothing to say about post-quantum cryptography directly. It is the layer that makes CNSA 2.0 enforceable government-wide instead of inside the Pentagon alone.
The mandate layer came ten days later. EO 14412 set hard deadlines for civilian high-value assets and high-impact systems, with NSS carved out by name. The Department of War PQC Strategy, released June 23, covered the systems the order deliberately does not touch. EO 14413 established the offense-side complement. Then the implementation layer: OMB M-26-15, the civilian execution playbook, landed June 24.
Here is the one sentence your board needs: every federal high-value asset and high-impact system, civilian and defense alike, now carries a dated PQC migration deadline with a named accountable official behind it, and through federal procurement that obligation is about to reach a large slice of the private sector. The carve-out for National Security Systems hands those systems to a separate authority chain, the one NSPM-12 created, where the CNSS and NSA enforce CNSA 2.0.
I put the whole framework in one frame in a complete reference guide; this edition is the curated tour.
The Civilian Half: EO 14412 and the Playbook That Arrived in Two Days
Executive Order 14412 sets the top-level dates for civilian federal systems. High-value assets and high-impact systems, meaning anything rated “high” under FIPS 199, must transition to ML-KEM (FIPS 203) for key establishment by December 31, 2030, and to ML-DSA (FIPS 204) for digital signatures by December 31, 2031. The one-year split is the right call, and it is the sequencing I have argued for through the Applied Quantum PQC Migration Framework for years. Key establishment goes first because Harvest Now, Decrypt Later is an active threat against data in transit today, and because swapping a key-encapsulation algorithm into a key exchange is operationally lighter than rebuilding a certificate hierarchy. Signatures get the extra year because signature migration touches PKI, cross-certification, and the entire certificate lifecycle. The order describes the harvesting threat in all but name, stating that adversaries can collect encrypted data now and decrypt it once quantum capabilities mature.
OMB M-26-15, signed by OMB Director Russell Vought, turns those dates into an operational plan. Section 4(b) of the order gave OMB ninety days to issue this guidance. OMB delivered it in two. Combined with the DoW strategy, which was cleared for publication in April and held for a June 23 release, the pattern is obvious: the orders, the defense strategy, and the OMB memo were drafted as one package, and the signing was the starting gun for releasing documents that were already written. Anyone waiting for the implementation details before acting has lost that excuse.
The memo establishes a five-phase timeline running from 2026 to 2035. Phases 1 and 2 cover strategy, discovery, and pilots through 2028. Phase 3 completes key-establishment migration for the priority categories by 2030, Phase 4 completes signatures by 2031, and Phase 5 sweeps up the remaining civilian systems by 2035. That last phase is the first official federal document to reconcile the order’s hard 2030 and 2031 dates with the broader 2035 horizon. For any organization that has been unsure whether 2030 or 2035 was its planning target, the answer is now precise: it depends on how your systems are classified. Hold federal high-value data or run a high-impact system, and your date is 2030. Everything else gets 2035 at the latest.
Two provisions are written for the engineering team as much as for the planners. First, M-26-15 treats crypto-agility as an implementation requirement with named architectural patterns rather than a principle to aspire to. A system hardcoded to a specific algorithm is out of compliance even when that algorithm is correct today. Second, the memo calls hybrid cryptography “an intricate and resource-intensive stopgap,” cooler language than most European regulators use, which signals that OMB sees full PQC as the target and treats dual-stack operation as a transitional state agencies should plan to exit. Agency migration plans are due to OMB and the Office of the National Cyber Director around October 22. The plan that lands on a federal CISO’s desk this October has to include phased timelines, automated inventory, resource estimates, and governance roles, and none of that gets assembled in 120 days by an agency that has not already started.
One civilian-side wrinkle that matters for contractors. M-26-15 approves SLH-DSA (FIPS 205, the hash-based scheme formerly called SPHINCS+) as a conservative signature fallback that avoids lattice assumptions. CNSA 2.0 deliberately excludes it. A civilian agency can choose it; a National Security System cannot. That divergence is small on paper and a real tracking problem for any organization that straddles both worlds.
The Defense Half: the DoW Strategy and the 2033 Hiding in Three Words
The Department of War PQC Strategy is the defense-side companion to the executive order, and that pairing is the story. For the first time, US post-quantum migration carries dated deadlines on both sides of the line separating ordinary federal systems from National Security Systems. DoW Chief Information Officer Kirsten Davies announced the release and tied it to the previous day’s executive actions, though the document itself carries an April 16 clearance stamp. It was finished early and held to land alongside the orders.
The strategy sets two headline dates. By December 31, 2030, all DoW systems must support PQC or be phased out. By December 31, 2031, all DoW systems must use PQC, “unless otherwise noted.” It organizes the work into five lines of effort and two acquisition tracks: a High Assurance track for NSA-certified devices that depend on the agency’s Key Management Infrastructure, and a Commercial Solutions track built on NIST-standardized algorithms in commodity IT. It builds on a November 2025 CIO memorandum that already required component-level inventories, named migration leads, and pre-deployment approval for PQC technologies.
Now the three words. “Unless otherwise noted” is the most consequential clause in the document, because it is what puts 2033 back into a plan that otherwise reads as 2030 and 2031. The strategy commits NSS to CNSA 2.0, whose exclusive-use dates run to 2033 for operating systems, web and cloud services, large PKI, and constrained devices. Those are exactly the categories the strategy spends most of its length on: embedded encryption units in weapon platforms, enterprise defense PKI, the edge devices in unmanned and space systems. The hardest systems in the portfolio answer to a 2033 backstop baked in by reference, even though the headline date reads 2031.
This is worth flagging because the DoW press release claimed the department would “accelerate ahead of the timelines” set by the executive order. That claim is not quite justified by the actual text. The order does not set timelines for the DoW’s core National Security Systems in the first place; it scopes its dates to civilian high-value and high-impact systems and routes NSS through a separate channel under CNSA 2.0.
On the systems where a fair comparison exists, the DoW’s milestones are softer, not harder. The order requires civilian high-value systems to complete key-establishment migration by 2030. The strategy requires only that systems support PQC by 2030, with use deferred to 2031. A capability gate is a weaker commitment than a finished migration. The department moved fast on tempo, and that speed is real and to Davies’s credit. It did not commit to earlier end dates than the rest of government.
Where the strategy is strongest is its refusal to let confidentiality stand in for the whole job. A solution that migrates only encryption, with no PQC authentication, will not count as PQC at all. This writes the Trust Now, Forge Later thesis into federal policy: the signature and authentication threat is systematically underweighted next to encryption, and for a department whose catastrophic scenarios include forged firmware signatures on weapon systems, putting authentication on equal footing is the correct judgment. The strategy also defines “done” as deprecation across the entire data pathway, with the vulnerable algorithms removed rather than left running alongside the new ones, and it rejects QKD and other false substitutes outright. Even outside the Pentagon this lands: a two-order, one-strategy, one-sensing-program rollout in a single week is a national-security posture shift, and the effective federal deadline moved to 2030 whether you wear a uniform or not.
The Same Two Years, Sliced Three Different Ways
Here is the part almost every write-up of this rollout has missed, and it is where migration programs go to fail. The executive order and the DoW strategy both land on 2030 and 2031, so the convergence looks like alignment. The two documents are slicing the same deadline along entirely different axes, and a third axis sits underneath them.
The executive order slices by cryptographic function: key establishment by 2030, signatures by 2031. The DoW strategy slices by migration stage: support by 2030, exclusive use by 2031. CNSA 2.0, which governs the National Security Systems at the center of the defense plan, slices by product category: software and firmware signing and traditional networking reach exclusive use by 2030, while web and cloud, operating systems, large PKI, and constrained devices run out to 2033.
Picture a CISO at a defense prime. As a federal contractor, the firm answers to the order’s FAR rule, which flattens everything to one date: comply with NIST’s PQC standards by 2030. The civilian agencies it sells to slice that deadline by function. Its NSS-adjacent work slices it by stage. The NSS hardware underneath slices it by product category, with the categories running to 2033. The same two years, decomposed three ways, plus a flat procurement date, and none of the four congruent with the others. The headline convergence is real. So is the divergence underneath it, and the systems hardest to migrate are precisely the ones where the schedules pull apart. If you span civilian and defense systems, map every system against all of these lattices, because the earliest binding deadline that applies to you is your real deadline, and “2030 and 2031” hides more than it reveals.
The Contractor Cascade Is the Part That Reaches You
If you take one operational point from this edition, take this one. The procurement provision is where the federal mandate reaches furthest into the private sector, and it reaches organizations that may never deal with a federal agency directly.
The order directs the FAR Council to publish a proposed rule within 180 days, putting it around December, requiring covered contractors to comply with NIST’s PQC standards by December 31, 2030. Once finalized, that rule extends the federal mandate into the private sector through the largest lever the government has over technology: the point of sale. A prime contractor that needs PQC-validated products by 2030 imposes that requirement on its subcontractors, who impose it on their suppliers, and the obligation propagates down every tier until it reaches companies far from any federal contract. This is the pattern that made FedRAMP the de facto cloud security baseline and is making CMMC the baseline for defense supply-chain security. The DoW strategy compounds it by directing CMMC updates to include PQC requirements and by sitting on top of the November 2025 memo’s inventory and approval regime.
For a defense prime, that means the FAR rule for civilian work, CMMC-PQC and the November memo for defense work, and CNSA 2.0 for anything NSS-adjacent, all at once. The one piece of good news is the Cryptographic Bill of Materials mandate, which directs CISA and NIST to define minimum CBOM elements within 270 days, around March 2027. That gives contractors a common format to demand structured cryptographic data from their own suppliers. I have argued for years that you cannot migrate what you cannot inventory, and three overlapping compliance regimes cannot be tracked in a spreadsheet. The bottom line: a rule nominally aimed at “covered federal contractors” will, in practice, set PQC expectations across a large share of the US technology economy by the end of this decade, and the companies that cannot demonstrate compliance by 2030 risk losing access to supply chains that serve the largest technology buyer on earth.
The Bottleneck Both Tracks Share
Every part of this plan funnels through a validation chokepoint, and the coordinated rollout makes the problem worse rather than better, because the chokepoint exists on both tracks at once. On the High Assurance side, NSA certification gates every encryption unit, and Type 1 certification has historically been the binding constraint on military crypto modernization. On the commercial side, the Cryptographic Module Validation Program gates every product that needs FIPS 140-3 validation, and that process currently runs around 18 months or longer.
The order tells NIST to accelerate the validation program. The DoW strategy calls for streamlined NSA certification. Both documents acknowledge their respective chokepoint. Neither has widened it. An order to accelerate a certification queue is a statement of intent, and intent has never validated a cryptographic module. Here is where I would put my own money on a slipped date: if any of the 2030 deadlines gives way, it gives way here, in the gap between the volume of products that need validating and the throughput of the programs that validate them.
The Other Order: Building the Threat While Defending Against It
EO 14412 has a twin. Executive Order 14413, “Ushering in the Next Frontier of Quantum Innovation,” is the offense-side complement, and the two are designed to work as a pair. One races to defend the nation’s data; the other races to build the machines that will eventually threaten it.
The order’s centerpiece is the Quantum Computer for Application Development and Discovery Science initiative, a national effort to deliver a science-grade quantum computer to a Department of Energy facility, one that performs scientific work no classical computer can replicate. The language carefully sidesteps the trap of “quantum supremacy” benchmark claims. The Secretary of Energy has ninety days to publish technical specifications, which will be the first real signal of what the government considers achievable and which hardware approach it believes is closest. The order also directs three next-generation quantum sensor projects to be fielded by September 30, 2028, with the DoW announcing a sensing initiative of up to $200 million, and it treats the quantum supply chain as a sovereignty problem, echoing the argument I made in Quantum Sovereignty: the “just buy the box” escape hatch is closed, because the vendor still controls firmware, calibration, and support, and answers to its own government’s jurisdiction.
The honest caveat is funding. The order directs spending “subject to the availability of appropriations,” which means Congress still has to authorize it. The $625 million the White House cited is existing investment, not new money attached to the order. The policy direction on both offense and defense is now set, and the cross-reference between the two orders ties them together: the innovation order asks the DNI and the Secretary of War to track the national-security implications of advancing quantum computers, “such as the implications for the migration to post-quantum cryptography.” Whether execution matches ambition is a question for the procurement machinery and the appropriations process, not the policy.
Quantum Flapdoodle: The Substitutes the Pentagon Just Named Out of Bounds
This rollout did the debunking work for me, so let me amplify it. Both the DoW strategy and OMB M-26-15 close the door on the alternative-physics shortcuts that vendors have been selling as escapes from the migration work. The strategy refuses quantum key distribution, quantum networking, non-local quantum randomness, and pre-shared-key schemes that lack PQC asymmetric key establishment as legitimate routes to quantum resistance. The OMB memo lists only the NIST-standardized algorithms and adds a one-line warning at the bottom of its table: symmetric-key-based protocols should also be avoided. There are no alternative routes on offer.
If you are still being pitched a “mathematically proven unbreakable” box, or a proprietary algorithm that conveniently sidesteps the standards process, or QKD as a drop-in replacement for the cryptography the order just mandated, the federal government has now told you in two documents that those things do not count. The reasoning on symmetric key distribution is worth keeping straight, because it is subtle. The symmetric algorithm itself is not the concern, Grover’s algorithm and its debatable practical feasibility aside. The concern is symmetric key establishment and distribution riding on a channel that itself uses quantum-vulnerable cryptography, where the symmetric protection is only as strong as the key-distribution mechanism underneath it. Any vendor still pitching one of these excluded technologies into federal procurement is selling a future audit finding dressed up as a security product.
From the Applied Quantum Desk
A quiet note of validation, then back to work. The framework I publish freely, the Applied Quantum PQC Migration Framework, already answers most of what these documents make mandatory: the minimum-viable CBOM, the separation of key-establishment and signature tracks, crypto-agility architecture, and vendor governance. The federal package did not change the methodology. It removed the margin for delay.
If you are a CISO staring at an October plan deadline, a program manager mapping systems against three deadline lattices, or an investor doing diligence on a vendor’s PQC roadmap, Applied Quantum does this work: crypto-inventory, crypto-agility implementation, PQC migration, vendor due diligence, and procurement support. The book-length version of the methodology lives in Quantum Ready, and the self-assessment scorecard is a free diagnostic if you want to see where you stand before the plan is due.
What to Do Now
The actions have not changed since I first wrote them. The deadline has.
Build a cryptographic inventory you can audit. Use automated tooling now, software composition analysis and protocol scanners, and build toward a central CBOM even before the federal standard drops in early 2027.
Identify your binding deadline. High-value or high-impact system, 2030. DoW system, support by 2030 and use by 2031. NSS under CNSA 2.0, category-dependent up to 2033. Federal contractor, 2030. Non-priority civilian, 2035. If more than one applies, the earliest one governs.
Put crypto-agility in your procurement specs today. A system you buy in 2027 that cannot swap algorithms without re-architecture is a system you will be replacing before 2030.
Force your vendors onto PQC roadmaps with named delivery dates. Procurement-driven compliance reaches further and faster than regulation, and it is about to reach your supply chain.
Stop handicapping Q-Day. None of these documents estimate when a cryptographically relevant quantum computer will exist. They treat it as a planning assumption and set deadlines on migration complexity and the harvesting threat. As I have written for years, the ecosystem, not the physics, sets the clock now.
If this map saved you an afternoon of reading five policy documents, forward it to the colleague who still thinks they have until 2035. Corrections are welcome and I publish them, as the EO-number fix at the top of this edition shows. Hit reply, tell me what I missed, and tell me whether a single-theme special edition like this one is useful or whether you prefer the usual mix. I read everything.
— Marin