Harvest Now, Decrypt Later: The Silent Threat to Internet Security
Why archiving encrypted data today becomes a ticking post-quantum — timebomband what organizations should do now.
Historically, it has generally been accepted that RSA is unbreakable. However, for some time now, it has been increasingly heard that quantum computers will be able to easily break it using Shor’s algorithm. RSA remains one of the most commonly used public-key algorithms in the TLS/HTTPS ecosystem, and its use continues to be widespread across web servers, client libraries, and legacy systems. Despite growth in elliptic-curve based algorithms such as ECDSA, many existing certificates issued by public Certificate Authorities are still RSA-based, and RSA support is virtually universal across browsers, servers, and cryptographic libraries. This continued prevalence is due in part to long-standing compatibility with large portions of the internet and the inertia inherent in global cryptographic deployments. Although elliptic-curve and post-quantum algorithms are increasingly supported and recommended, especially for forward-looking security, RSA remains a foundational element of current HTTPS infrastructure.. I am not suggesting that enterprises are living in stagnation and are unaware of the threat, as cryptographic standardization organizations (NIST and others) recommend the gradual phasing out of RSA and ECC by 2030–2035 in the context of migrating to cryptography resistant to quantum computers. Industry reports speak of major investments in “post-quantum cryptography” (PQC) and assessments that many companies are already prototyping or evaluating PQC in 2025, which is a sign that RSA/ECC will be replaced in the future. However, the key problem is the current ubiquity of RSA, and this is exactly where “Harvest Now, Decrypt Later” comes in.
I present to you a forecast of what may happen in the future, current problems, and events I consider most probable. Some may think: “So what if someone records my encrypted data that they can decrypt in 10 years? It will be outdated by then.” Yes, but what about genomic and medical data? Your DNA will not change. What about corporate trade secrets hidden for decades (e.g., the Coca-Cola formula or engine blueprints)? What about evidence of crimes or compromising photos of VIPs? If you are protecting data today that must remain secret for 20 years (e.g., nuclear power plant blueprints), and a quantum computer emerges in 15 years, you are already 5 years late with security implementation today. The deeper we go, the more potential damage we see. And the worst part? What has already been recorded by potential superpowers/attackers cannot be taken back; hundreds of thousands, maybe even millions of encrypted records may lie collected on separate drives for years, waiting for a sufficiently powerful quantum computer to emerge.
Let’s start with a fact. Most of our communication today (HTTPS, VPN, banking) relies on public-key cryptography (e.g., RSA).
How does this work deterministically?
- You take two huge prime numbers.
- You multiply them by each other.
- The result is a “padlock.” To open it, one must guess what those two initial prime numbers were (integer factorization).
The logic of today’s security:
For a classical computer, reversing this multiplication (guessing the numbers) would take thousands of years. Therefore, we considered it “secure.” Here enters the hacker or intelligence agency. Let’s ask a simple economic question: Is data storage cheap?
Yes. Hard drives are incredibly cheap. Storing petabytes of encrypted network traffic costs cents for a superpower. So, if a hacker cannot read the data today, but the cost of intercepting and storing it is near zero, the logical move is to hoard everything. It is like recording an encrypted phone conversation. You don’t understand it now, but you have the tape.
Here enters quantum mechanics. You don’t need to understand superposition to understand this one fact: There exists an Shor’s algorithm that turns finding those “prime numbers” from Step 1 from a task taking 1,000 years into a task taking few minutes.
It is not a question of if a computer capable of doing this will be built, but when. Physics allows it, and engineering is slowly catching up. The moment a quantum computer is switched on (so-called Q-Day), all “tapes” (datasets) accumulated over the last 10–20 years become an open book.
For a better explanation, I will use Mosca’s Theorem:
I define three time variables:
- X (Shelf Life): How long must your data remain secret? (e.g., 20 years for medical data).
- Y (Migration Time): How long will it take you to replace your entire infrastructure with one resistant to quantum computers (Post-Quantum Cryptography — PQC)? This is not just a software update; it is a replacement of hardware, protocols, certificates.
- Z (Collapse Time): Time until “Q-Day” (when a stable quantum computer is built).
Logical Inequality:
If
X+Y>ZX+Y>Zthen you have already been hacked. What can be deduced from this? Data you collect and encrypt today will be decrypted halfway through its validity lifecycle. HNDL is not a risk “for the future.” It is a risk management error committed now.
The Calendar of Inevitability: Y2Q and NSA Directives
When will this happen? Experts have coined the term Y2Q (Years to Quantum), analogous to the Y2K problem, only the stakes are higher.
We do not have to guess. We look at the regulators. The NSA (US intelligence agency) released the CNSA 2.0 document. This is not a request. It is an order for the national security sector and vendors.
NSA Schedule (The Timeline of Pain):
- 2025–2026: Beginning of mandatory publication of migration plans to post-quantum algorithms (PQC).
- 2030: Ban on using classic RSA/ECC for new systems. Software and firmware must be exclusively post-quantum.
- 2033: Complete phase-out of old cryptography. If you use RSA in government systems after this date, you are breaking the law.
Business Deduction:
Since the NSA gives itself until 2033, it means they assess the risk of a stable quantum computer appearing around that date as high. Business usually follows these standards with a 2–3 year delay, but in the case of HNDL (Harvest Now, Decrypt Later), we must act sooner.
PQC Security Standardized by NIST:
In August 2024, NIST published the first Federal Information Processing Standards (FIPS) defining post-quantum cryptography algorithms intended to replace current methods (such as RSA) vulnerable to attacks by future quantum computers. Standard FIPS 203 specifies a module-lattice-based key-encapsulation mechanism (ML-KEM) based on the CRYSTALS-Kyber algorithm for secure key establishment; FIPS 204 introduces a module-lattice-based digital signature algorithm (ML-DSA) based on CRYSTALS-Dilithium; and FIPS 205 standardizes a stateless hash-based digital signature scheme (SPHINCS+) as an alternative to traditional digital signatures. All three standards were officially approved on August 13, 2024, and designed to withstand both classical and quantum cryptographic attacks. Furthermore, in March 2025, NIST selected the HQC algorithm as an additional quantum-resistant key mechanism to be developed as the next PQC standard.
FACTS
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
MY SPECULATIONS:
Attacks and Limitations (My Forecasts):
Get X-3306’s stories in your inbox
Join Medium for free to get updates from this writer.
The AI-Handshake Downgrade: Network protocols are designed to be “backward compatible.” When your computer connects to a server, they negotiate: “Do we use Kyber? No? Then maybe RSA?”.
Hackers know they don’t have to break Kyber if they force the server to use RSA.
Logical chain of events:
Currently, we think of “Man-in-the-Middle” attacks as static scripts. A new, not yet public threat which I consider very probable observing the rapid progress of AI, is AI operating in real-time on edge routers.
AI will not try to break encryption. AI will learn to manipulate “Hello” packets (the first milliseconds of connection) in a way undetectable to current firewalls (which do not yet understand the full nuances of PQC), to force both parties to downgrade to weaker encryption (Downgrade Attack), even if both sides support quantum standards. It will do this by introducing microscopic delays or bit corruptions that the negotiation algorithm will interpret as “incompatibility.”
Hardware Physics: “The Helium-3 Cold War” (The Isotope Crisis):
Although alternative technologies exist (photons, ions), the current arms race of leaders (IBM/Google) relies on superconductors dependent on rare isotopes, which creates a geopolitical bottleneck. Quantum computers (superconducting) require the Helium-3 isotope to operate dilution refrigerators to reach temperatures close to 0 Kelvin. Helium-3 is incredibly rare (mainly a byproduct of tritium nuclear weapon production).
Scaling quantum computers to a useful level (thousands of logical qubits) requires building massive cooling farms. I believe Earth’s Helium-3 resources are insufficient to build such infrastructure for the commercial sector. However, I must emphasize that currently, both the supply market and research into alternative cooling technologies are working to mitigate this limitation.
What might happen:
Brutal rationing of access to computing power will occur. Quantum computers will not make it to the “cloud for everyone” (like AWS today). They will be deemed a strategic/military resource.
Private companies will not have access to a “real” quantum computer capable of breaking ciphers, not because the technology doesn’t exist, but because there is no physical fuel (Helium-3) to cool these machines.
Proposed Safeguards:
1. Offensive Defense: Chaffing and Winnowing / Data Flooding
Until now, the conversation about security in the HNDL era was defensive:
“They record — we encrypt better.”
This is reaction logic, not strategy.
A more realistic approach sounds different:
If I cannot stop the interception, I must make the intercepted data economically and analytically useless.
Every attacker, regardless of whether they possess a classical cluster or a quantum computer, operates within finite resources: time, storage, and computing power. Quantum computers do not eliminate the cost of computation: they change its structure.
If you transmit 1 GB of high-value data and an adversary archives it, the risk is obvious.
However, if the same channel generates 100 GB of traffic which:
- is cryptographically indistinguishable,
- has identical statistical properties,
- uses the same encryption mechanisms,
then the cost of analysis, selection, and subsequent cracking grows exponentially on the attacker’s side.
Strategy:
This approach is known as Chaffing and Winnowing — mixing real data with cryptographically correct “chaff.”
In the pre-quantum era, this was often impractical due to bandwidth and infrastructure costs. In the HNDL era, this cost shifts: generating noise becomes cheaper than its subsequent analysis.
This is not a universal solution nor resistant to every form of analysis (e.g., metadata correlation), but in specific system classes, it can effectively:
- increase the cost of attack,
- delay data selection,
- decrease the value of intercepted archives.
Conclusion:
The future of security is not just stronger algorithms, but also data visibility management. Encryption protects content. Camouflage protects meaning.
Zero-Knowledge Proofs — ZKP:
Until now, the model was:
Client sends data -> Server receives and checks it. (Data is in motion = risk of interception).
The new paradigm (ZKP) is:
Client proves they know the data WITHOUT sending it.
How it could work:
Imagine you must prove you are of legal age to enter a site.
Traditionally: You send a photo of your ID (which is risky).
ZKP: You perform a complex calculation on your device based on your birth date and send only the result “TRUE.” The server mathematically verifies that you could not have obtained the result “TRUE” by lying, but never sees your birth date.
Why is this key for the Quantum Era?
If you do not transmit the secret over the network, HNDL has nothing to steal (“Harvest”). There is no encrypted packet with data, there is only a mathematical proof.
Even if a quantum computer breaks the transmission in 10 years, it will only find the information: “Yes, he was right,” not “This is his card number”.
Bibliography:
https://en.wikipedia.org/wiki/Shor%27s_algorithmhttps://utimaco.com/service/knowledge-base/post-quantum-cryptography/what-mosca-theorem
https://www.f5.com/labs/articles/the-2021-tls-telemetry-report
https://ieeexplore.ieee.org/document/365700
https://www.infosecurity-magazine.com/news/half-of-websites-still-using/
https://nvlpubs.nist.gov/nistpubs/ir/2024/NIST.IR.8547.ipd.pdf
https://en.wikipedia.org/wiki/Commercial_National_Security_Algorithm_Suite
