EN
Back to the archive

The archive · Developer & Business Tools · Product decision · 2020–2024

Zama's encrypted-compute bet: $73M Series A for usable homomorphic encryption

Paris's Zama bets fully homomorphic encryption is finally practical: compute on encrypted data, no decryption. $73M Series A, $50M+ signed in six months.

Zama

The betFully homomorphic encryption can get fast enough that encrypted computation becomes a product: analytics and contracts running on data nobody ever decrypts.Scaling

What the business is

Open-source FHE libraries — Concrete, TFHE-rs and fhEVM — that let developers compute on encrypted data for AI privacy and confidential blockchain smart contracts.

Starting capital$8M pre-seed plus seed, then a $73M Series A co-led by Multicoin Capital and Protocol Labs on March 7, 2024 at a valuation approaching $400M — $81M total, the largest round ever for a homomorphic-encryption company (TechCrunch).

How it started

Rand Hindi and Pascal Paillier built Zama in Paris around homomorphic encryption; the underlying work began in 2016, and a 2019 breakthrough sped up the calculations roughly 100x, making the product bet credible.

What happened

Zama released four open-source libraries adopted by 3,000 developers, commercialized first with blockchain customers willing to pay per transaction or per contract, and signed more than $50M in contract value within six months of starting to sell; the Series A, announced March 7, 2024, valued the company at nearly $400M.

How it ended up

Still independent and scaling at the raise: a 75-person team, with Solana co-founder Anatoly Yakovenko and Ethereum/Polkadot co-founder Gavin Wood among the round's participants alongside Estonian deep-tech investor Metaplanet.

Background

Zama was built in Paris by Rand Hindi (ex-Snips) and Pascal Paillier around fully homomorphic encryption (FHE), the decades-old idea that you can compute on encrypted data without ever decrypting it. Work began in 2016, and a 2019 breakthrough sped up the underlying math roughly 100x, turning the idea into a plausible product.

Zama's wedge was to open-source the tooling — the Concrete, TFHE-rs and fhEVM libraries — so developers could try encrypted compute without a research team. By March 2024, 3,000 developers were using its four libraries, and the company had signed more than $50M in contract value within six months of commercializing, mostly with blockchain customers paying for confidential smart contracts.

On March 7, 2024 Zama announced a $73M Series A co-led by Multicoin Capital and Protocol Labs at a valuation approaching $400M, bringing total funding to $81M — the largest round ever raised by a homomorphic-encryption company, according to TechCrunch. The round valued a bet that FHE would secure both blockchain privacy and data exchange around AI training and inference.

What has to be true

  • Zama shipped FHE as developer tools rather than research papers, compressing the gap between 'the math works' and 'a developer can build with it'.
  • Blockchain buyers were the right first market: they are structurally exposed to transparency and paid per transaction or per contract, so FHE's overhead became a price, not a blocker.
  • The open-source libraries became its sales channel — 3,000 developers and a 24.8k-star fhEVM repo — letting demand arrive instead of being hunted.
  • The 2019 ~100x speedup mattered strategically: it moved FHE from 'impossible' to 'expensive' — venture money arrived exactly when that framing changed.

What can be applied

A moonshot tech becomes a business when you give away the hard part as open-source tools and charge for the application — first paying customers are the ones legally forced to need it.

Aftermath

After the March 2024 raise Zama kept shipping: by September 2026 the fhEVM smart-contract library had grown to 24,819 GitHub stars and the org's public repos showed a push into accelerated encrypted compute (kms, coprocessor and hardware-register projects), signaling a move from pure software FHE toward hardware acceleration to cut remaining overhead.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases