Latest / Cybersecurity Business with Fexingo: Security Companies, Breaches, and Enterprise Defense / How Quantum Computing Threatens Encryption by 2030
Transcript
- Lucas: Here's a number that should keep every CISO up at night: a research team in Shanghai published a paper in March 2025 claiming they had, in simulation, factored a 2,048-bit RSA key using a quantum algorithm. They didn't break real internet traffic — yet. But the clock is ticking. Luna: Wait — factoring RSA-2048 is the benchmark for breaking public-key encryption. If that becomes practical, every HTTPS connection, every VPN tunnel, every signed software update is toast. How close are we to that actually happening? Lucas: The consensus among cryptographers is that a fault-tolerant quantum computer capable of breaking RSA-2048 is still 7 to 10 years away. But 'away' is the operative word. The transition to post-quantum cryptography takes years — think a decade or more for a global infrastructure shift. So the window to act is now. Luna: So it's a Y2K-style problem, except the deadline is fuzzy and the fix isn't a simple patch. What exactly needs to change? Lucas: The core issue is that our current public-key algorithms — RSA, Elliptic Curve Diffie-Hellman, ECDSA — rely on math problems that are hard for classical computers but trivial for quantum ones. Shor's algorithm can factor large primes exponentially faster. Post-quantum cryptography replaces those with lattice-based, hash-based, or code-based problems that quantum machines can't crack. Luna: Right. And the National Institute of Standards and Technology, NIST, has been running a competition since 2016 to standardize these new algorithms. Where does that stand now? Lucas: In August 2024, NIST finalized three standards: crystals kyber for key encapsulation, and crystals dilithium and SPHINCS+ for digital signatures. They also selected a fourth, FALCON, for scenarios that need smaller signatures. So the standards exist. The hard part is deployment. Luna: And deployment touches everything. I've read that a typical large bank has thousands of certificates, dozens of protocols, and legacy hardware that might never be upgradeable. How do you even start? Lucas: You start with an inventory. You need to know where every cryptographic key lives — TLS certificates, code signing, firmware updates, VPNs, even smartcards and IoT devices. Then you prioritize: anything that encrypts data with a long shelf life — like healthcare records or national security documents — is most urgent because of what's called 'harvest now, decrypt later'. Luna: Harvest now, decrypt later — that's the idea that adversaries can collect encrypted data today, store it, and decrypt it once they have a quantum computer. So anything you send today could be exposed in 2033. Lucas: Exactly. That's why the NSA, the UK's NCSC, and Germany's BSI have all published migration timelines. The NSA, for instance, wants all National Security Systems to be quantum-resistant by 2035. But if you're a bank moving money today, you need to think about it now. Luna: And it's not just about algorithms. The performance of these new algorithms is different. Kyber keys are larger than RSA keys, and signatures can be enormous. SPHINCS+ signatures, for example, are about 40 kilobytes — compared to 256 bytes for an ECDSA signature. That matters for constrained environments. Lucas: It does. So enterprises need to test for latency, throughput, and compatibility. Some companies, like IBM and Google, have already run pilot migrations. Google's Chrome team experimented with CECPQ2, a hybrid protocol combining classical and post-quantum key exchange. They found the overhead was acceptable — about a 10% increase in handshake time. Luna: That's encouraging. And there are startups focused entirely on this. SandboxAQ, which spun out of Alphabet, offers a platform that analyzes cryptographic inventories and helps plan migrations. They've raised over $500 million. Lucas: Right. And the market is waking up. In 2025, we saw the first enterprise-grade post-quantum VPNs and the first quantum-safe code-signing certificates from companies like DigiCert. But the scale of the problem is immense. Let me give you a concrete number: a typical Fortune 500 company might have 10,000 TLS certificates. Each one has to be reissued with a new algorithm, and each reissuance may require firmware or hardware updates. Luna: And some hardware can't be updated. Think about medical implants, industrial control systems, or satellites. They have chips that were designed without upgradeability in mind. Those become ticking time bombs. Lucas: That's why crypto-agility — the ability to swap algorithms without breaking systems — is now a design requirement. NIST has also been clear that hybrid modes are a good interim step: use classical and post-quantum together, so you're protected even if one is broken. Luna: So what's the biggest obstacle? Is it technical, financial, or organizational? Lucas: I'd say organizational. Most CISOs I've talked to know this is coming, but they're already overwhelmed with ransomware, zero-day exploits, and compliance. Quantum migration is a long-term, high-effort project that doesn't make headlines — until it's too late. The cost of not acting is asymmetric: a single catastrophic breach could dwarf the migration cost. Luna: And there's also the question of confidence in the new algorithms. What if someone finds a flaw in Kyber or Dilithium? That would be a massive setback. Lucas: It's a real risk. That's why NIST designed the process with multiple rounds and independent analysis. But we've already seen attacks that reduce the security margin of some candidates. In 2022, researchers showed a key recovery attack on a variant of the Rainbow signature scheme — that candidate was dropped. So the process works, but it's not foolproof. Luna: Given all this, what's your advice to a mid-size company listening right now? Concrete steps. Lucas: First, conduct a cryptographic discovery audit. Use a tool like the one from SandboxAQ or even open-source scripts that scan your network for certificates and algorithms. Second, identify your high-risk data — anything that needs confidentiality beyond 10 years. Third, start a pilot for a non-critical system: replace a single internal TLS certificate with a hybrid or post-quantum one. Fourth, build crypto-agility into your procurement specs. And fifth, train your engineers. This is a new skill set. Luna: That's a solid roadmap. I'd add: start today. Because the worst case isn't a quantum computer arriving tomorrow — it's waking up in 2032 with a decade's worth of unplanned migrations and a lot of stored data that's already exposed. Lucas: Exactly. And I'll leave you with this: the cost of transitioning the entire internet to post-quantum cryptography has been estimated at around $7 billion. That sounds huge, but compare it to the cost of a major data breach or the collapse of trust in digital infrastructure. The investment is small insurance against a catastrophic failure. Luna: Well said. So the message is clear: quantum may be coming, but the window to prepare is now. Lucas: And that's what we'll be watching. For Fexingo's Cybersecurity Business, I'm Lucas. Luna: And I'm Luna. Thanks for listening.