FHE as a bet on quantum resilience

The Quantum Threat

For many years, the evolution of quantum systems has lurked as a severe threat that could unravel security standards everywhere. Headlines have raised concerns about the technology’s potential to expose sensitive data across finance, healthcare, AI, and media, as sufficiently advanced quantum computers in the hands of bad actors could destabilize society — laying bare sensitive data, compromising digital signatures, and tampering with security at large. Breaking the backbone of today’s internet infrastructure would have dangerous real world implications, as so much of our world has gone online in the last few decades. With increasing awareness of this threat, the idea of Quantum resilience–being prepared to face a post-quantum world — has steadily gained traction. And yet, even though quantum may seem relatively new, a potential answer already exists.

In many ways crypto already holds the keys to quantum-resilience, though the technology’s value as a security solution is often overshadowed by more speculative use-cases. While tokens, the incentive layer powering crypto, have become widely recognized and frequently discussed, the deeper mathematical principles underlying cryptography have become somewhat divorced from public understanding. Tokenomics, and decentralized finance have dominated the conversation, eclipsing knowledge of the foundational technologies that ensure the security and integrity of the very systems they power. Nevertheless, if you’ve ever used WhatsApp, Google Drive or gone to any website that starts with https, you’ve benefited from cryptography in your daily life.

Simply put, cryptography governs how data is encrypted, transmitted and verified across networks. Today, certain symmetric cryptography schemes within AES, SHA-2 and SHA-3 have high post-quantum confidence, if sufficiently large key sizes are used. Bitcoin for example leverages SHA-256 which is a SHA-2 variant in its proof-of-work network. Ethereum leverages Keccak-256 which is closely related to SHA-3 and also considered quantum resilient to the same degree as its relative. These algorithms are the basis of secure communication and data integrity across countless applications. Beyond symmetric cryptography, emerging post-quantum cryptography is focused on developing algorithms that are specifically designed to withstand attacks by large-scale quantum computers, offering an even higher degree of confidence in security. Post-quantum cryptography is rapidly undergoing development as many believe advances in Quantum computing will soon challenge current security standards.

Why Invest in Quantum Resilience Now?

Although “quantum” has become a bit of a buzzword, it isn’t new; it’s simply hit an inflection point. At its core quantum computing leverages the principles of quantum mechanics–the science of how particles behave at the smallest scales–to perform calculations faster than regular computers. In 1980 Paul Benioff proposed the quantum Turing machine, a computer based on quantum mechanics. Later in 1984, Quantum cryptography was proposed as a method for secure communication leveraging quantum keys. Then in 1994, Peter Shor designed an algorithm to efficiently factorize large composite numbers–a near impossible task for classical computers, particularly as numbers increase. The algorithm exponentially sped-up factoring relative to classical algorithms by employing modular exponentiation and quantum fourier transformation (“QFT”) after a classical number check for factors. The algorithm effectively challenged classical encryption schemes like RSA for the first time. Shortly after, in 1996, Lov K. Grover created Grover’s Algorithm, which provided quadratic acceleration in unstructured searches. This would for example allow users to select a specific input from a black box search of an unstructured database, also a significant improvement over classical methods.

Beginning in 1997, the National Institute of Standards and Technology (“NIST”) launched the development of a new encryption standard to combat a post-quantum world, which was later named Advanced Encryption Standard (AES) and was adopted in 2001.Today, breaking AES-128 (Advanced Encryption Standard with a key size of 128 bits) would require an outsized amount of time using Grover’s algorithm, which is currently infeasible with existing quantum technology. Symmetric algorithms like AES-128 and above are thus currently considered quantum-resistant, with large enough key sizes. Put simply, this means that all of your Whatsapp and Signal messages, which use AES-128 for encrypting content are safe — for now. However, as quantum computers continue to achieve increasing breakthroughs AES-128 will ultimately prove insufficient.

In 2019, Google reported a significant breakthrough in quantum: a 54-qubit quantum processor capable of computing faster than classical supercomputers, and in February 2025, Microsoft introduced Majorana 1, the first quantum chip powered by a new topological core architecture which they believe could help quantum computers realize industrial-scale problems in years rather than decades. In the same year, the Nobel Prize in Physics was awarded to three US scientists who showed that quantum mechanical properties could be made concrete on a macroscopic scale, theoretically paving the way for the development of the next generation of quantum technology. Quantum has certainly captured mindshare among corporations and academia alike in recent months, but how close are we to quantum computing in practicality?

Simply having more qubits isn’t enough to make a quantum computer powerful enough to break traditional encryption schemes at the requisite speed. Other factors including reliability, hardware and architecture are also key for acceleration in the space. For example, if qubits are noisy and unreliable, adding more doesn’t necessarily accelerate productivity. If qubits are unable to communicate with each other this also necessitates more complex operations to engender interactions, introducing further opportunities for error and computation break downs. Without error correction to protect qubits, running deep algorithms also becomes untenable, as computers struggle to perform numerous sequential operations. In reality it would take thousands of qubits with high reliability and deep interaction capabilities with error correction mechanisms to successfully run Shor’s algorithm, meaning that the gap between achieving quantum computing remains relatively wide today.

Nevertheless we believe that now is the right time to invest in quantum computing solutions because:

(1) Quantum technology is improving rapidly. For example, a 2019 study showed that 2048 bit RSA integers could be factored in 8 hours by a quantum computer with 20 million noisy qubits. The same author updated their study in 2025 showing that less than 1 million physical qubits were required under similar baseline assumptions but with improved error-correction & arithmetic.

(2) Advances in AI are starting to significantly accelerate timelines across various scientific disciplines and we believe that Quantum stands to benefit as well.

(3) National Security Memorandum 10 (NSM-10) established the year 2035 as the primary target date for completing the migration to PQC across federal systems, making quantum a particularly interesting category and potential recipient of focus from defense agencies. Following the transition guidelines, AES-112 & 128 would all be disallowed after 2035

Source: NIST

No Waiting for Quantum Breaks: FHE is Ready

While businesses and institutions around the world brace for the quantum threat, crypto is already ahead of the curve. At PL Capital we believe that among the most promising advancements in the space today is Fully Homomorphic Encryption (“FHE”), a cryptographic scheme that is already considered quantum resilient in most cases. FHE’s quantum-resilience stems from its reliance on lattice-based cryptography, which is generally built on top of Learning With Errors (“LWE”) or Ring-LWE problems. These schemes involve solving complex equations with a carefully introduced layer of randomness or “noise.” FHE’s resilience boils down to just how difficult it is to crack these problems. At present there are no known quantum algorithms that break LWE/Ring-LWE, making FHE theoretically secure against quantum computers.

FHE’s combination of practical functionality and strong security makes it a key emerging player in the quantum computing era. Another core feature of FHE is that it allows computation on encrypted data. In traditional systems encrypted data must be decrypted within a trusted environment like a server, database or secure enclave, before any operations can be performed with the data. However, the point at which data is decrypted is when it is usually most vulnerable to attack. Many high-profile data breaches have occurred exactly at this stage. FHE removes significant vulnerability by allowing computation directly across ciphertexts and spitting out encrypted results. Data remains encrypted at all stages: while stored, while in transit, and while being processed. By removing the need to decrypt data during processing, FHE significantly reduces attack surface and exponentially increases security.

FHE was first proposed in 2008, but it wasn’t until 2022 that PL Capital led the Series A in Zama*. Throughout this post we will continue to denote PL Capital investments with an asterisk at their initial mention. When we invested in Zama, we felt that privacy needs were reaching an inflection point and recognized the potential for a breakout moment in the space. At the time Zama’s team was already sufficiently advanced in their development of FHE technology and we were particularly excited by the company because of its functionality. Zama’s protocol serves as a layer on top of existing chains, effectively upgrading the security of networks without bridging, and allowing for smart contract execution with confidentiality. Since CEO Rand Hindi and his team started building, Zama has come to dominate the FHE landscape, partnering with companies across both web2 and web3. Zama’s ecosystem has given birth to several general purpose L2s focused on scaling FHE use-cases with privacy-preserving smart contracts. This includes protocols such as Fhenix, INCO, Mind Network, Shiba & Airchains.

Source: PL Capital

Going up the stack, we believe that finance is one of the strongest potential sectors for FHE and have talked with or looked at several companies already building at this unique intersection of cryptography and cryptocurrency. Within the Zama ecosystem, Fluton is building the first confidential intent-centric cross-chain liquidity protocol, enabling users to perform on-chain and cross-chain transactions with fully encrypted data, while companies like Zaiffer enable stablecoin users to add a layer of privacy to any EVM-compatible stables and actually use them on DeFi apps as well. Outside of the Zama ecosystem, companies like Arcium for instance, leverage a mixture of MPC, FHE and ZKPs to enable private swaps, confidential tokens and private resolution across prediction and other markets on Solana. Singularity uses a similar mixture of tools to allow for on-chain settlement of darkpool trading.

Zama tested out one of the major potential financial use-cases for FHE in real-time when they launched their public token auction in January, selling ~10% of their token supply in a sealed-bid Dutch auction and using their own protocol to keep these bids confidential. Confidentiality is crucial for Dutch auctions, due to their unique structure. At the end of the bidding period, the auction smart contract executed from the highest bid price to the lowest and allocated tokens sequentially to each participant, filling bids from high to low based on remaining supply. The price at which the token supply was depleted became the clearing price. Bidders above the clearing price can typically claim a refund for the difference and bidders below this price miss out on allocation.

Identity is another space where FHE could have potential applications, particularly given the threat that quantum could have in this area. Currently there are projects like Privasea and Galactica Network already active in the space. Privasea is building FheID, a proof-of-humanity technology built with FHE that shields its users digital presence from bots and deepfakes. Galactica by contrast is an L2 focused on making it easier for developers to build apps that are natively compliant and privacy-preserving as it relates to identity/reputation. The protocol’s identity virtual machine is designed to compute over encrypted private data — not just what’s public on-chain. They also leverage a mixture of both ZK and FHE.

Within AI more FHE companies are emerging as builders begin to understand its benefits as it relates to AI inference. With Trusted Execution Environments (“TEEs”) data is decrypted inside hardware that parties are required to trust, leaving users open to attack if the vendor is compromised. This introduces single points of failure and censorship risk. So while TEEs do offer a high level of protection, they also introduce dependencies and other vulnerabilities. ZK can’t be used for ML because it lacks the ability to perform the complex computation required–like matrix multiplications, non-linear activations, and handling large tensors. ZK proofs allow a prover to assure a verifier that a computation was correct after the fact, but it is not the actual compute engine. Finally, multi-party computation (“MPC”) which is another solution often touted as an alternative to FHE, has significant limitations as it requires multiple online parties to jointly compute. FHE by contrast doesn’t require synchronization and trust is not steeped in assumptions around potential collusion. Companies like Lattica AI are building inference platforms that run AI models on encrypted queries using FHE, while Beacon Protocol is building private data infrastructure for smarter AI leveraging FHE technology.

In many ways, FHE could be a critical solution for any industry with high compliance requirements or for any company or network seeking to collaborate with a high degree of trustlessness, while still preserving privacy and security. By eliminating the need to expose raw data during computation, FHE dramatically reduces risk and expands the realm of what is possible with cross-functional and multi-party workflows. As data sharing demands compound alongside regulation, FHE offers a solution to collaboration without compromising security or compliance. However, the technology is not wholly without limitations.

Areas of Opportunity in FHE

Today, scalability is one of the biggest factors limiting FHE adoption. In AI for example training is usually infeasible with FHE because of the significant computational cost it incurs as every operation is done on high-dimensional lattices. Additionally, while FHE is more compatible with ML than zk because of its ability to compute, native ML math doesn’t exist, requiring workarounds that threaten model accuracy and may require manual redesign. Finally, each incremental FHE operation increases ciphertext noise, the accumulation of which could lead to decryption failures. This noise can be reset, but is also costly. These limitations have been the gating factor to FHE expansion and are why many systems use hybrids that leverage FHE + MPC + ZK/TEEs.

However, while many believe that pure FHE is a decade away, we believe it’s closer than most expect. Zama has invested in algorithmic implementations which have boosted FHE performance significantly. However, algorithmic acceleration can only take FHE so far. Exponential improvement in latency requires innovation from the hardware front as well. Recognizing this, at the end of May, Zama announced the first FPGA purpose-built for FHE acceleration. Their homomorphic processing unit (HPU) is 10x faster and cheaper to run than a CPU-based counterpart and represents a major step forward for FHE scalability. Companies like Optalysys, Niobium, Cornami, Fabric Crypto, Belfort are also focused on building hardware solutions that accelerate encrypted compute, enabling real-time processing. FHE has already hit an inflection point in scale that is making it a more attractive alternative and breakthroughs in hardware will only serve to further accelerate this.

Though we haven’t seen many use-cases in crypto around healthcare x FHE, we also think this is a promising focus area and is one of the dominant FHE use-cases outside of crypto today. Healthcare data is also governed by strict regulatory regimes like HIPAA in the US and GDPR in Europe. This is due to the sensitive and deeply personal nature of healthcare data. Providers and payors take on a high degree of legal, ethical and financial liability when they process or share this data leveraging traditional methods. Whenever data is processed there is a significant window of time in which patient medical histories, genetic information and billing records are at risk of exposure. Leveraging FHE, would allow healthcare organizations to analyze encrypted medical records without ever exposing the underlying data. FHE would lead to improved privacy for consumers, as well as fewer data breaches. It would also expand the bounds of how healthcare data could be leveraged, unlocking advanced analytics and predictive modeling without putting patient data at risk.

We also believe there are still significant opportunities to build around use-cases in Finance as well. Imagine for instance using FHE for credit decisions. Today, personal financial data is highly regulated, meaning that personnel at banks who interface with raw customer data to develop credit scoring models take on a degree of liability should anything happen to this data. However, with FHE, the user would be able to use the same data without decrypting it, improving privacy for consumers and leading to fewer data breaches. FHE would also enable a higher degree of collaboration between lenders and data providers without risking compliance.

Finally, bringing things back to Quantum. It’s important to note that not all FHE schemes are synonymous with quantum resilience as there could be rare cases that leverage classical assumptions like factoring or discrete algorithms. As quantum computing develops, the parameters required to ensure FHE and other typically lattice-based schemes against quantum attacks could widen significantly. The key takeaway here, however, is that in order for crypto to truly be resilient, there needs to be a major shift in cryptography standards. Given the decentralized nature of blockchains like Bitcoin & Ethereum, it may require significant coordination to push through the required protocol change, meaning that we should start to plan these changes sooner rather than later.

Disclosure

This post is for informational purposes only and is not intended as a recommendation to purchase or sell any commodity, digital asset or security. The information for this post has been obtained from public sources believed to be reliable. Protocol VC, LLC (“PL Capital Crypto”) makes no representation as to the accuracy or completeness of such information. Opinions, estimates, forecasts and projections in this post constitute the current judgment of the author and are subject to change without notice. It can be expected that some or all of such assumptions will not materialize or will vary significantly from actual results.

Past performance is no guarantee of future results. Investing in digital assets involves a high level of risk including the loss of principal.