Strategic Brief 🔒 Post-Quantum Security | Securing Finance, Healthcare, and Utilities from Quantum Collapse with AWS-LC

The crucial systems societies rely depend on math to prove what’s real. Cryptography does this quietly in the background, making sure data isn’t forged or altered. As long as that math holds, society hums along unnoticed. When it breaks, markets freeze, hospitals stall, and cities can go dark.

This fragility is visible every day in ransomware headlines. Hospitals have been forced to cancel surgeries, pipelines have shut down, and companies have gone bankrupt after attackers encrypted records and erased recovery options. In the UK, more than 70% of businesses reported a ransomware attack last year, and one 158-year-old logistics firm folded entirely after a $6.3 million ransom demand it could not meet. When the math falters, systems collapse.

Quantum computing could provide a considerably more devastating escalation. These machines don’t just run faster — they run differently. By using qubits that represent many possibilities at once, they can explore paths in parallel that classical computers must grind through sequentially.

For everyday tasks that difference doesn’t matter. But for the cryptographic foundations of digital security, it’s existential. Problems that would take billions of years on today’s hardware could be unraveled in days or hours on a sufficiently powerful quantum computer.

Instead of breaking into a system to plant ransomware, quantum machines threaten the cryptographic foundations themselves. RSA and elliptic curve cryptography — the algorithms that secure financial transfers, health data, and government communications — could become forgeable. Breaking RSA-2048 may require around four thousand stable qubits, elliptic curve just twenty-three hundred. Today’s largest machines hover around one hundred qubits, but the U.S. Government Accountability Office warns that capable systems could arrive within 10–20 years. Cisco estimates a one-in-three chance within fifteen. Some researchers caution it could be sooner — possibly before 2030.

Ransomware carnage proves the risk is already present. Quantum computing could raise the threat from organizational collapse to systemic societal failure unless action is taken now.

In payments, magnetic stripes once stood as the barrier between a purchase and a counterfeit. Criminals skimmed billions by copying the unencrypted strip, and only when the industry shifted to EMV chips — tiny computers that perform a cryptographic handshake in milliseconds — did confidence return. Counterfeit fraud in the U.S. dropped by nearly 90% after adoption (Federal Reserve study ).

In banking, forged telex messages in the 1960s were enough to move millions across borders until SWIFT introduced cryptographically signed messages. That shift restored trust in international transfers, and today SWIFT secures more than $7 trillion every day (SWIFT history ).

Energy systems tell an even harsher story. Stuxnet, revealed in 2010, slipped past defenses by wielding stolen digital certificates (Wired analysis ). A few years later, Ukraine’s power grid dimmed under attacks exploiting weak authentication (CISA alert ).

And in healthcare, the costs have escalated to breathtaking levels. In 2024 alone, more than 275 million U.S. medical records were breached, with an average cost of over $10 million per incident (HIPAA Journal ).

Each of these examples shows the same pattern: systems hold until the cryptography or mechanisms behind them fail, and then we react. We've shown an ability to address issues, but when societal fabric is at stake we need ways to mitigate risks before the first breach occurs.

The U.S. Government Accountability Office (GAO) — a nonpartisan watchdog agency that advises Congress on risks to federal programs and national security — estimates that machines powerful enough to break today’s encryption may be only 10–20 years away. Cisco, a leading technology company, is even more blunt: they put the odds of disruption at one in three within the next fifteen years. Cryptographers have run the numbers and found that breaking RSA-2048 would require about 4,098 qubits and ECC-256 around 2,330 qubits — while today’s largest quantum computer, Google’s Willow, has just 105 qubits (The Verge ).

Rather than wait for that gap to close, Amazon Web Services Libcrypto (AWS-LC) was created with a mind towards prevention. AWS-LC is designed to help bridge the transition to post-quantum security before the tipping point arrives. And importantly, this isn’t an AWS only initiative. It’s a global collaboration, shaped through organizations including NIST (National Institute of Standards and Technology), IETF (Internet Engineering Task Force), ETSI (European Telecommunications Standards Institute), and a wide network of open-source communities.

At the center of this global effort are the standards bodies that guide how the world secures its information. The U.S. National Institute of Standards and Technology has been running an open competition to identify algorithms strong enough to resist quantum attacks, and AWS researchers have contributed to two of the leading candidates: ML-KEM and ML-DSA.

Alongside this, the Internet protocol groups that shape how systems communicate — including the Internet Engineering Task Force and the European Telecommunications Standards Institute — are extending TLS to support hybrid key exchanges, standardizing post-quantum digital signatures, and writing migration guides for critical infrastructure.

Equally important is the open-source community. AWS-LC itself, validated under the strict FIPS-140-3 program, already implements ML-KEM and ML-DSA. The Linux Foundation’s Post-Quantum Cryptography Alliance and AWS’s lightweight TLS library, s2n-tls, are producing reference implementations to ensure these algorithms are not just secure on paper, but practical and interoperable in real systems.

These efforts matter because security only works if it works everywhere. A web browser in Boston must be able to verify a healthcare record in Berlin. Encryption only protects if everyone moves together.

AWS is already deploying post-quantum protections. Key Management Service (KMS), Secrets Manager, and Certificate Manager now combine the familiar Elliptic Curve Diffie-Hellman (ECDH) with ML-KEM to defend against “harvest now, decrypt later” attacks.

At the foundation of these deployments is AWS-LC, the first open-source cryptographic module to include ML-KEM in its FIPS validation — a U.S. government certification process that tests cryptographic modules to make sure they meet strict security requirements.

Quantum-safe doesn’t mean slower. In fact, AWS-LC is already delivering measurable speed improvements. Python TLS handshakes run 2× faster with AWS-LC compared to OpenSSL 3.x (AWS blog ).

On AWS’s own hardware, the gains are even sharper: RSA signing is 94% faster on AWS Graviton chips, AES-GCM encryption is 100% faster on Intel Ice Lake processors, and AES-XTS, which protects stored data, is up to 360% faster (AWS blog ).

Even with these new protections, the overhead is minimal. A TLS 1.3 handshake with ML-KEM adds only 2.3 KB — about 1% delay on a typical Internet connection. Human users don’t notice the difference, but attackers harvesting encrypted traffic will (AWS blog ).

This isn’t about AWS alone. It’s about a global, open, and collaborative effort to make sure the systems we rely on — from markets to hospitals to power grids — remain secure when quantum arrives. AWS-LC is not only strengthening AWS’s own services, it is spreading across the broader ecosystem of open-source infrastructure that powers the Internet.

One of the earliest proofs of this came through curl, the ubiquitous data transfer tool. Every change to AWS-LC is tested against curl, and curl does the same in reverse, creating a tight loop of reliability. Daniel Stenberg, curl’s creator, put it plainly:

Adoption has only accelerated. HAProxy, the load balancer used across finance, healthcare, and large-scale web deployments, now bundles AWS-LC by default (HAProxy blog ).

In the Rust ecosystem, rustls — the standard TLS library — migrated its default crypto provider to aws-lc-rs, which already handles more than 200,000 daily downloads and over 32 million lifetime downloads. As of v0.23, rustls now defaults to AWS-LC’s post-quantum algorithms (Memory Safety blog , PQ Key Exchange ).

Python has also seen major gains. In CPython 3.10, AWS-LC delivers more than a 2× speedup for TLS handshakes compared to OpenSSL 3.x. AWS has rolled this into its internal distributions, and developers can experiment themselves using the example Dockerfile maintained by AWS engineers (GitHub repo ).

Looking ahead to Fall 2025, the AWS-LC roadmap includes a release AWS-LC-FIPS v4.0 and submission of ML-DSA — a lattice-based digital signature algorithm — for NIST validation. ML-DSA is NIST’s chosen standard for post-quantum signatures, which are what prove the identity of servers, verify software updates, and secure digital certificates.

By bringing ML-DSA into a FIPS-validated module, AWS-LC opens the door for adoption across federal agencies and regulated industries, where FIPS compliance is mandatory. And because signatures sit at the heart of HTTPS, this step directly impacts the ~98% of U.S. internet traffic now encrypted by HTTPS and the hundreds of millions of websites that rely on digital certificates. In short, ML-DSA extends post-quantum protection from key exchanges into the everyday proof-of-identity layer that keeps the Internet — and critical infrastructure — running (AWS Security Blog , NIST CMVP ).

For those who want to dig deeper:

Many of these references were provided during AWS Hero Summit 2025 by presenters including Will Childs-Klein, Senior Software Development Engineer, AWS Cryptography.

References provided at #AWSHeroSummit 2025 by team members including Will Childs-Klein , Alex Weibel, Senior Software Development Engineer, AWS Cryptography.