# The Crypto Clock Is Already Running — And Hardware Makers Just Started Building the Lifeboats
Somewhere in a server farm, data is being vacuumed up and stored for later. Not because the attacker can read it today. Because they're betting they will be able to in ten years, when a quantum computer finally breaks the RSA key protecting it. That attack — harvest now, decrypt later — isn't a hypothetical. Intelligence agencies have been warning about it for years, and the encrypted traffic being scooped up right now includes health records, financial contracts, diplomatic cables, and source code.
Which is why what hardware makers are doing in 2025 and 2026 matters far more than most people realize. It's not preparation for a future threat. It's a race against data that has already left the building.
## NIST Lit the Fuse; Silicon Valley Is Playing Catch-Up
The National Institute of Standards and Technology finalized its first post-quantum cryptographic standards in August 2024 — ML-KEM (formerly CRYSTALS-Kyber) for key encapsulation, ML-DSA (CRYSTALS-Dilithium) for digital signatures, and SLH-DSA (SPHINCS+) as a hash-based fallback. That was the starting gun.
Software moved first, as it always does. Google added ML-KEM to Chrome. Apple retrofitted iMessage with PQC hybrid key exchange. Signal shipped PQXDH in 2023, ahead of the NIST finalization. These are the easy wins — update a library, push a release, done.
Hardware is different. When encryption lives in silicon — in a Trusted Platform Module, a Hardware Security Module, a secure enclave on a mobile SoC, or firmware burned into an industrial controller — updating it means replacing the chip or issuing a signed firmware patch to devices that may not have a functional update mechanism at all.
## The Layers That Actually Matter
What's actually happening at the hardware level is worth unpacking, because the press releases blur it:
Processor-level acceleration. Intel's 4th-gen Xeon (Sapphire Rapids) and newer AMD EPYC chips have begun shipping with cryptographic co-processors that can handle the compute-intensive lattice math behind ML-KEM and ML-DSA without torching CPU cycles. This is a prerequisite for PQC at scale — the new algorithms are significantly heavier than RSA or ECC.
Secure element upgrades. Infineon and NXP, which supply secure microcontrollers to everything from smart cards to automotive ECUs to IoT endpoints, have both announced PQC-ready product lines. Infineon's SLB 9672 TPM added ML-KEM support in firmware. This matters because TPMs are the root of trust for Windows BitLocker, remote attestation, and enterprise endpoint security.
HSM vendors. Thales and Entrust have shipped PQC-capable HSMs — hardware that manages cryptographic keys for banks, certificate authorities, and government agencies. These are the chokepoints for the PKI that underpins HTTPS across the web.
Mobile silicon. Qualcomm's Snapdragon 8 Elite includes a dedicated quantum-resistant crypto engine. Apple's Silicon team has been quiet on specifics, but the iMessage implementation signals that PQC is already present in at least the software stack, with hardware acceleration almost certainly in the pipeline for future A-series chips.
## The Part No One Is Talking About Enough
The devices getting upgraded are the visible ones — flagship phones, enterprise servers, new HSMs. The devices not getting upgraded are the dangerous ones.
Consider the installed base of industrial control systems in power grids, water treatment plants, and manufacturing floors. Many of these run on hardware from 2008 to 2015, with fixed-function firmware, no over-the-air update capability, and expected service lives of 20-30 years. They use RSA-2048 or ECC-256 for authentication and key exchange. When a cryptographically relevant quantum computer arrives — and the NSA's official position is to plan for one by 2030 — none of these devices will be patchable on a reasonable timeline.
The same problem applies to embedded medical devices, smart meters deployed by utilities, and a substantial chunk of the global automotive fleet. Hardware vendors shipping PQC-capable chips today are solving for new deployments. The attack surface that actually keeps security teams up at night is the installed base that won't see a PQC chip until it's physically replaced.
## What Good Looks Like Right Now
For organizations trying to get ahead of this, the migration isn't a single project — it's a sequence:
---
## HackWire Analysis
The hardware industry's PQC push is real, but the framing around it is dangerously optimistic. Coverage tends to focus on the progress — NIST standards finalized, chips shipping, software updated — and underweights the structural mismatch between where PQC hardware is being deployed and where the actual risk lives.
The harvest-now-decrypt-later problem reframes the urgency entirely. We're not talking about a future threat to future data. Sophisticated nation-state actors — primarily China's MSS and affiliated groups, per NSA and CISA advisories — have been known to collect and store encrypted traffic for years. This is documented behavior, not speculation. Encrypted VPN tunnels, diplomatic communications, and financial settlement data captured in 2019 are sitting in storage waiting for the quantum unlock. The cryptographic protection on that data was permanent when it was captured.
This makes the "we have 5-10 years" framing actively misleading. For data that was sensitive when it was transmitted, the window is already closed. The migration debate for historical data is over — it lost. What hardware makers are building now protects future data and future communications. That's valuable. But organizations that think they have a leisurely timeline to plan their PQC migration haven't grappled with what's already been harvested.
There's a secondary story here that almost no one is covering: the certificate authority ecosystem. The public key infrastructure that validates every HTTPS connection, every code-signing certificate, and every email authentication scheme runs on RSA and ECC. Migrating root CAs to PQC requires coordinated action across browser vendors, OS makers, CA operators, and every relying party globally. The hardware progress is real; the PKI migration is a separate, harder problem that hardware alone cannot solve.
The vendors shipping PQC chips deserve credit for moving. But the question worth asking is whether the market incentives align with the actual risk — and the answer, looking at the installed base of unmigratable critical infrastructure, is clearly no.
— HackWire Editorial
---
## Related Coverage