Zama: Making "The Holy Grail of Cryptography" Practical Enough to Become a Unicorn

Key Facts

• Zama is a Paris-based open-source cryptography company, founded by Dr. Rand Hindi and Dr. Pascal Paillier, that became the first unicorn built specifically around fully homomorphic encryption in June 2025, raising $57 million at a valuation exceeding $1 billion.
• Who founded Zama: Dr. Pascal Paillier, a cryptographer, and Dr. Rand Hindi, a data scientist and serial entrepreneur previously known for founding Snips, an AI voice assistant company acquired by Sonos.
• What is fully homomorphic encryption: A cryptographic technique that allows computations to be performed directly on encrypted data without ever decrypting it, so sensitive information stays hidden even while being actively processed.
• How much has Zama raised: Over $150 million total by mid-2025, including a $73 million round in 2024 and a $57 million (€49 million) Series B in June 2025 that pushed its valuation above $1 billion.
• What is the Zama Confidential Blockchain Protocol: A platform enabling confidential smart contracts on any blockchain using FHE, launched alongside the Series B, with initial support on Ethereum through Zama's FHEVM.

Cracking Cryptography's Long-Standing "Holy Grail"

Zama's entire business is built around commercializing a technology long regarded as one of cryptography's most difficult and consequential unsolved problems. The company's own educational materials describe fully homomorphic encryption, or FHE, directly: it is "often called the holy grail of cryptography," representing "the endgame for bringing confidentiality to blockchain" and, more broadly, to any computing system handling sensitive data. The core capability FHE unlocks is genuinely remarkable in its simplicity of description and difficulty of execution: "FHE enables data processing without decryption—companies provide services without accessing user data, while users experience unchanged functionality". Protocol.ai's independent explainer frames the contrast with conventional encryption directly: "Unlike conventional encryption methods that require data decryption during computation — a point at which data can be vulnerable — FHE permits computations directly on the encrypted data," meaning "you can process and analyze the information" while its "content remains hidden, providing an unparalleled level of data privacy".

That capability addresses one of the most persistent, structurally difficult problems in modern computing: the moment sensitive data needs to actually be used, processed, or analyzed by any system, it has traditionally needed to be decrypted first, creating an unavoidable window of vulnerability where the data sits exposed in plaintext, regardless of how strongly it was encrypted while merely stored or in transit. FHE closes that window entirely, letting computation happen on data that remains mathematically encrypted throughout the entire process.

The Founders: A Cryptographer and a Serial AI Entrepreneur

Zama's founding team pairs deep specialist cryptographic expertise with prior startup and AI commercialization experience. Dr. Pascal Paillier, a cryptographer, and Dr. Rand Hindi, described across multiple sources as a data scientist and entrepreneur, co-founded the company together. Hindi's background carries particular relevance to Zama's current AI-adjacent positioning: a LinkedIn analysis of the company notes that "in March, Zama founder Rand Hindi published the Zama roadmap," describing Zama as "the platform that the burgeoning FHE industry has rallied around and is leveraging," adding pointedly that "nearly every FHE project is using Zama" — a claim suggesting Zama's open-source cryptographic libraries function less as one competitor's product among many, and more as foundational, widely adopted infrastructure that the broader FHE development ecosystem has organized itself around, similar in structural position to how certain widely adopted open-source frameworks have become default industry standards in other technical domains.

Overcoming FHE's Historic Computational Bottleneck

The single greatest obstacle that has historically kept FHE a theoretical curiosity rather than a practically deployed technology is computational cost: performing calculations on encrypted rather than plaintext data is dramatically slower using conventional computing approaches, often by orders of magnitude, making genuinely widespread real-world FHE deployment impractical for decades after the underlying cryptographic theory was first developed. Zama's specific technical strategy for closing that gap, according to LinkedIn's analysis of the company's approach, involves "leveraging techniques like CKKS encryption, GPU acceleration, and custom silicon to overcome FHE's computational bottlenecks," with a stated goal "to make FHE a practical reality within the next decade, ushering in a new era of privacy-preserving computing".

That combination — specialized cryptographic schemes paired with both existing GPU acceleration and purpose-built custom silicon — mirrors a broader pattern across several AI infrastructure companies profiled elsewhere in this series, including d-Matrix's custom chip approach to inference acceleration: when an existing computational bottleneck is severe enough and the underlying market opportunity large enough, building specialized hardware specifically tuned to the mathematical structure of the target problem becomes a viable, sometimes necessary strategy rather than simply optimizing software running on general-purpose infrastructure.

Becoming the First FHE Unicorn

Zama's valuation milestone in June 2025 carries particular significance not just for the company itself, but as a marker for an entire nascent technology category. Tech.eu's coverage of the June 2025 round is explicit about that framing, headlining its report "Zama becomes 1st fully homomorphic encryption unicorn". SiliconANGLE's parallel coverage confirms the specific figures: Zama, "an open-source cryptography startup focused on building fully homomorphic encryption or FHE to protect privacy in blockchain and artificial intelligence applications," raised $57 million in that Series B round, co-led by Blockchange Ventures and Pantera Capital, bringing total funding to over $150 million and pushing the company's valuation above $1 billion — a round that came, per that same reporting, "a little over a year after the company raised $73 million in 2024".

EU-Startups' coverage adds direct commentary from the company itself on the significance of that valuation milestone, quoting Zama describing how "reaching a $1 billion valuation represents a significant increase that reflects the market's confidence in our FHE technology and our team's" work — a notably candid acknowledgment that the valuation increase specifically reflects growing market confidence in a technology category, FHE, that had for decades remained largely theoretical and academic rather than commercially deployed at scale.

Launching the Confidential Blockchain Protocol Alongside the Unicorn Round

Zama's June 2025 funding announcement coincided deliberately with a major product launch: the Zama Confidential Blockchain Protocol, alongside a public testnet enabling developers to begin building confidential applications on top of the technology. EU-Startups' coverage specifies the initial technical scope: the protocol launched "initially on Ethereum through Zama's FHEVM, with support for other EVM chains and Solana to follow in 2026" — indicating a deliberate, staged rollout strategy starting with Ethereum's dominant smart contract ecosystem before expanding to additional blockchain networks over time.

The underlying use case that protocol unlocks is significant for blockchain technology specifically: public blockchains have historically faced an inherent tension between their core transparency — every transaction typically visible to anyone examining the chain — and the privacy needs of many legitimate financial and business applications that genuinely require confidentiality. Zama's technology description frames FHE's blockchain application directly: it "allows truly private smart contracts — where the code runs transparently but the underlying data remains confidential" — preserving blockchain's core auditability and trustlessness properties while still shielding the actual sensitive data content of individual transactions from public view, a combination that had previously seemed almost contradictory to achieve simultaneously.

Beyond Blockchain: The AI Privacy Application

While Zama's most visible commercial application to date has centered on blockchain confidentiality, the company's own materials are explicit that FHE's implications extend directly and significantly into AI as well. Company representative Jeremy Bradley, in an interview about Zama's mission, described the founding motivation in terms directly applicable to both domains: "The key problem we saw was that while data is increasingly valuable for AI and blockchain applications, it is almost always exposed in use. We wanted to change that by making privacy-preserving computation practical and accessible". For AI specifically, Bradley notes FHE "enables training and inference on private datasets" — meaning organizations with genuinely sensitive data, such as healthcare records or financial information, could in principle train or run AI models directly on that data without the data itself ever needing to be exposed in decrypted form to the AI system's operators, a capability that could meaningfully expand which sensitive datasets organizations are willing to apply AI to at all, given how much current AI deployment is constrained specifically by data privacy and regulatory concerns.

A Note on Conflicting Valuation Reporting

Readers should be aware that publicly available reporting on Zama's precise, current valuation contains some inconsistency across sources — a reflection of how difficult it can be to track exact valuations for private companies with complex, multi-instrument capital structures spanning both traditional equity and, in Zama's case, blockchain-native token mechanisms. While the company's June 2025 Series B round was widely reported at a valuation exceeding $1 billion, some independent tracking services have subsequently listed different, lower figures tied to specific later events, including a reported public token sale valuation around $550 million in January 2026 — a divergence that likely reflects the distinct and sometimes disconnected ways private equity valuations and public token sale pricing can be calculated and reported for a company operating at the intersection of traditional venture-backed cryptography and blockchain-native tokenomics.

Why Zama's Bet on FHE Matters Beyond Its Own Valuation

Zama's rise as the first dedicated FHE unicorn carries significance extending well beyond the company's own commercial success. If FHE genuinely becomes practical and widely deployed at the scale Zama's roadmap envisions, it would represent a fundamental, structural shift in how sensitive data can be used across both AI and blockchain applications — eliminating a privacy-versus-utility tradeoff that has constrained data-driven technology for decades, rather than merely making that tradeoff somewhat more favorable. Zama's specific combination of deep, credible cryptographic research leadership under Paillier and prior AI commercialization experience under Hindi appears to have positioned the company as the closest thing the nascent FHE industry currently has to a foundational, ecosystem-defining platform — a position that, if the underlying technology continues maturing at its current pace, could make Zama considerably more consequential to the future of data privacy than its current valuation alone might suggest.

FAQ

Source: CrackTheDeck Research.