Exclusive Interview with Zama CEO: How Did the $400 Million Valuation of FHE "Wave Maker" Come to Be in Four Years?

Interviewee: Rand Hindi, CEO of Zama

Interviewed and written by: Wendy, Foresight News

One week after selling his company, which he founded nearly seven years ago, 35-year-old Rand Hindi immediately embarked on another entrepreneurial journey because, "how many times in a lifetime do you get the chance to create a company based on a new scientific breakthrough with the right timing and the right partner?" ------ When this idea emerged, he dismissed the notion of spending half a year traveling the world and plunged into a field that had not yet garnered much attention at the time------ Fully Homomorphic Encryption (FHE).

Four years later, the valuation of this new company reached $400 million.

In early March this year, Zama, co-founded by Rand Hindi, announced a $73 million Series A funding round led by Multicoin and Protocol Labs, aiming to address privacy issues in the blockchain and AI sectors using FHE technology.

Since then, Zama, headquartered in Paris, has not only become one of the most representative startups in the global FHE space but has also sparked a wave of venture capital activities related to FHE. Rand Hindi revealed to Foresight News, "In the next 6 to 9 months, there will be 9 projects announcing the use of Zama's technology," adding, "These projects have all secured funding, with a total valuation exceeding $1 billion."

Born in the romantic city of Paris, Rand Hindi has a very geeky side. He started programming at the age of 10 and subsequently earned a bachelor's degree in computer science and a Ph.D. in bioinformatics from University College London. The creation of Zama can be traced back to his experiences of being bullied in school during his teenage years.

In the 1990s, as the internet bubble was nearing its burst, 14-year-old Rand Hindi co-founded a very popular social networking site in France. At the same time, he was bullied by older kids at school. With no way out, Rand had a sudden idea: could this bully possibly be a user of the social media platform he created? He indeed discovered some secrets about the bully from the website's database and used this information to threaten him, ensuring that the bully would never dare to bother him again.

While this "an eye for an eye" approach helped resolve his immediate troubles, it also led Rand Hindi to reflect on data security and privacy issues. At that time, he believed that "privacy" would become an important topic in the internet space. After obtaining his Ph.D., Rand founded Snips, a company focused on solving privacy issues related to AI.

In 2019, after hearing that Rand Hindi was going to sell Snips, Pascal Paillier, who had known him for many years and held a Ph.D. in cryptography, approached Rand to express his intention to establish an FHE company together. In early 2020, Zama was officially founded. Pascal Paillier became the Chief Technology Officer (CTO) of the new company, while Rand Hindi took on the role of Chief Executive Officer (CEO).

A new business adventure began.

Foresight News: Zama was officially founded in 2020, at a time when even ZK (zero-knowledge proofs) had not yet gained mainstream attention from the investment community, and the application of FHE was generally considered to be even further behind. As an entrepreneur and also an investor, you understand the importance of timing in entrepreneurship. So looking back, why did you want to start an FHE company at that time?

Rand Hindi: I have always been very interested in privacy issues.

In fact, since I founded my first company------a social network in the 90s------I have been focused on privacy. Because when you create a service that holds a large amount of data, you must consider how to protect that data. So for me, privacy is inseparable from everything we do online. The company I sold before (Snips) was an AI company trying to solve privacy issues for machine learning and artificial intelligence.

It was also because of running that company that I discovered homomorphic encryption technology in 2015. When I learned about this technology, I felt that everyone should use it, and it should be the default setting for everything. But at that time, FHE had not really taken off, and what could be done with it was very limited.

Even so, I (at that time) knew that I would want to establish a company focused on FHE in the future. So, one week after selling my AI company (Snips), I co-founded Zama with my co-founder Pascal to address the issues related to the application of FHE. It has been four years now.

Foresight News: When Pascal approached you, what breakthroughs had occurred in the FHE field? How does it compare to when you first learned about this technology in 2015?

Rand Hindi: In 2017, some new technologies were invented in the FHE field, opening the door to a new area of research. A significant breakthrough for Zama is that we were able to perform any type of FHE operation without the limitations of previous technologies for the first time.

When we started founding Zama, FHE technology faced three main issues: the first issue was that it was very difficult to use. You needed to be like a Ph.D. in cryptography to use this technology well; the second issue was that it was very slow. For example, multiplying two numbers could take several minutes; the third issue was that the use cases were very limited. Because existing FHE technology only allowed you to get approximate results on encrypted data. For example, calculating 2 times 2 would yield a result of not 4, but 3.8 or 4.2. This is fine for many use cases that do not care about approximation errors, but for medical use cases or blockchain use cases, (the accuracy is insufficient). If I want to give you a million dollars based on certain conditions, you want to ensure that those conditions are fully met, not just approximately met.

(In response to the above three issues), Zama has achieved a series of significant breakthroughs: First, we solved the (approximate result) problem by inventing new cryptographic techniques, which allow you to guarantee that the results of encrypted computations are exactly the same as when the data is unencrypted; secondly, we combined new mathematics and engineering techniques to make FHE computations 100 times faster; finally, we created tools for developers so they can easily use FHE without needing to learn cryptography.

Foresight News: A 100-fold increase in computation speed is already impressive, but a 100-fold speedup may still not be good enough or fast enough, right?

Rand Hindi: You're right. If you want to do things like ChatGPT or LLMs (large language models) in FHE, it is still too slow. But the good news is that we can now do LLMs in FHE, which was impossible before, (it's just still slow now), generating a token takes a day, which is not a good user experience, and we need a 1000 to 10000 times speedup to do LLMs. Of course, we will (eventually) achieve this because many companies, including Intel, are researching special chips to accelerate FHE. This is similar to ASIC chips in Bitcoin mining, or ZK chips, or AI-specific GPUs; we will also have special chips specifically for accelerating FHE. When these special chips start entering the market------I think maybe by the second generation of products------we will achieve a 10000-fold speedup.

So, speed is no longer an issue with FHE itself; it is a hardware acceleration issue.

So, what can we do with FHE before these special chips appear? It turns out that blockchain is a great use case. Because compared to cloud (computing), blockchain is very slow and quite expensive, and blockchain applications and smart contracts are very simple. In contrast, applications like LLMs and ChatGPT are the most computationally expensive. So if you can (use FHE) to do ChatGPT, then you can (use it) to do everything else.

Thus, blockchain is an excellent starting point for FHE use cases. As FHE performance gradually improves, we can obtain more and more different use cases.

Foresight News: Similar to some cutting-edge technologies like FHE and ZK, they all face the same problem in the early stages of development, which is the lack of an existing market. So what is Zama's market creation strategy? How do you create a market at this stage?

Rand Hindi: (In addition to the above advantages) the second benefit of blockchain is that the performance requirements match what FHE can do today. FHE can currently handle about five transactions in one second, and in the coming months, we may be able to achieve around 20 to 30 transactions per second on GPUs. This TPS (transactions per second) is sufficient for most blockchain applications.

What is interesting about blockchain is that it is a decentralized public infrastructure. In traditional cloud computing, whether you believe it or not, companies can tell you not to worry about confidentiality; they won't be hacked, and they won't leak data; only the company itself can see the data. But in blockchain, you can't do that; every piece of data in every transaction must be public to everyone. So in today's blockchain, there can be no kind of sensitive data or confidential applications. However, with FHE, you can use the blockchain in the same way; you still put data on the chain, but the data will be encrypted. Smart contracts will run on encrypted data. So suddenly, you can have confidentiality on a public, permissionless, decentralized infrastructure.

Most use cases in this regard are finance-related. If you talk to any large financial company that wants to securitize assets, they do not want to disclose their positions to everyone; they need confidentiality. AMMs (automated market makers), if you want to avoid MEV (maximum extractable value, which refers to the potential profit that miners or validators can gain by manipulating the order of transactions during block production), it's best to keep people from knowing what you want to trade.

(Another big use case is) gaming. If you want to play poker on the blockchain, then showing your cards is pointless. (And) AI, decentralized AI. If you want to train a neural network on a decentralized GPU network, you wouldn't want people to see your dataset. As a company, if I spend millions or billions of dollars acquiring data, I wouldn't just put it on the blockchain for everyone to steal. Without confidentiality, decentralized AI cannot work.

So for me, the question is, what applications have we not built due to the lack of confidentiality? I would say most applications.

Imagine if I told you that all the data in every mobile application on your phone could be viewed publicly by anyone. Then you wouldn't use most of those applications. That is the state of a blockchain without FHE (that could arise). With FHE, you can do anything.

Foresight News: As you said, ideally, every project or application in the blockchain space should have a need for protecting data security and privacy. But when you pitch Zama's services or solutions to some projects or companies, if they are not planning to use it for the time being or want to wait and see, what reasons do they give?

Rand Hindi: We started commercializing our technology nine months ago, and so far, over 100 companies have expressed interest in using it. Our community has 4000 developers. So, we are the most successful and widely used FHE (solution provider).

This is because, without FHE, these companies cannot even build what they want to build. Companies like Fhenix, Inco, and Shiba are using our technology to build confidential L1 and L2. But there are also some companies that want to build applications that require confidentiality. For example, AMMs that hide transactions, or asset tokenization that wants to hide balances and amounts in transactions.

So we have two types of customers: one type is customers building infrastructure using our technology, and the other type is customers building applications on top of that infrastructure using our technology.

Foresight News: As you mentioned, Zama is a widely adopted solution provider, but there are indeed some competitors in the market. Why do many projects choose Zama?

Rand Hindi: I think the simple answer is that we are the only effective product. There are other technologies, but technology is not a product; just the technology itself cannot be practically used by developers.

So, Zama made some very strategic bets from day one. We believe the two most important things for developers are: 1) they don't want to worry about anything cryptographic, so they shouldn't have to learn anything about cryptography; 2) they should be assured that anything they do on encrypted data will be exactly the same as when the data is unencrypted. As a developer, you can create an application without FHE and then just add an FHE layer on top of it.

(Using) other technologies requires you to delve deeply like an expert, then adjust it, modify it, and I think the main reason people adopt our (solution) is that it is simple, effective, and it does what we promised it would do.

Foresight News: FHE is one of the solutions to privacy issues, but there are also other technologies, such as the ZK we mentioned, and MPC (multi-party computation). They are related yet different. Can you briefly tell us what roles these three different technologies will play in the broader crypto space?

Rand Hindi: ZK is a great way to prove that you have performed a computation without revealing the data on which you based your computation. For example, I can prove that I am over 18 years old without revealing my actual age. The problem is that ZK does not allow you to compute on encrypted data. So if you have multiple users and you want to do something with the data of multiple people, the ZK prover must see everyone's data. Therefore, it is difficult to have composability and multi-user privacy with ZK.

But ZK is very suitable for things like scalability. So if you just want to prove the correctness of something, ZK technology is great. But for privacy, FHE is better. Because FHE allows you to compute on encrypted data. So you can have the same level of composability and multi-user interaction as regular applications, but the data is encrypted.

However, the problem with FHE is that if you encrypt data with the same key, then someone must have the decryption key to display the results. So how do you protect the decryption key? That's where MPC comes in. If you want to protect encrypted privacy, MPC is a great technology because what you can do is split the key into multiple parts and distribute them among different people. This way, no one has the key, and people need to reach a consensus before decrypting something.

So, FHE can be used for private computation, while MPC can be used to protect private keys. Technically, Zama does use MPC, but we do not use it as a substitute for FHE. We use MPC in a very specific part of the product, namely to protect private keys.

Some people also try to use MPC for secure computation, but the problem is that MPC is hard to achieve composability because MPC is very slow, unless you already know what you want to compute and can pre-process it. But in blockchain, if my contract calls your contract or another contract, you can't really do this kind of pre-processing. You have to compute in real-time. For this use case, MPC is very unfriendly.

Foresight News: You mentioned "keys." In a previous interview with Epicenter, you said that FHE itself is not the problem. Now, the issue is actually key management. Can you elaborate on this? Because this is not only a problem that FHE may bring to blockchain but also an opportunity, right? If some vendors can solve this problem.

Rand Hindi: Yes, absolutely.

I want to illustrate this with our fhEVM product. In fhEVM, as a user of FHE smart contracts, I must use the network's key to encrypt my data------because everyone is using a public key, you need to encrypt data under the same key to have composability between users------then I send this data to the blockchain, and the blockchain performs the computation. But sometimes I need to be able to partially decrypt, for example, I need to decrypt the balance of my encrypted ERC20 token. So how do you allow users to decrypt only specific parts of the blockchain state without revealing the key and allowing them to decrypt everything? Then there needs to be a private key protected somewhere, and that's where MPC comes in.

So, next to this FHE blockchain, there is actually an MPC key management service (KMS). It distributes the private key of the blockchain across multiple nodes and allows people to decrypt based on what the smart contract says. As long as this key management system is secure, the only thing you can see is what the smart contract lets you see.

So that's why I say FHE is not the problem because FHE is unbreakable, even by quantum computers. The problem is how to protect this private key, and the answer is MPC.

Foresight News: I'm also curious about what changes FHE will bring to the broader crypto ecosystem. For example, in the ZK space, there are some projects building so-called privacy chains, like Aleo and Mina. What about FHE? Do you think we will see similar situations arise?

Rand Hindi: (In terms of solving privacy issues) FHE is growing much faster than ZK because FHE is very easy to use in blockchain. You are just writing a Solidity smart contract, and there is no change for developers, and nothing is different on the user side. Everything is on-chain, and you can still have composability.

Before (FHE), ZK was the only technology people could try. But with FHE, everyone knows it is a better technological solution. The question has never been whether people want to use ZK or FHE; it was that FHE was not usable before Zama. So there was no choice.

Now that we have usable FHE, people are starting to turn to FHE as a privacy layer for blockchain. In the next 6 to 9 months, there will be 9 projects announcing the use of our technology, including Inco, Fhenix, and Shiba. These companies/projects have already raised funds, with a total valuation exceeding $1 billion. So, in 9 months, FHE has grown from zero to a $1 billion ecosystem, and this does not even include the value of Zama itself, but just the valuations investors have for these new chains and protocols. This organic growth proves that people really want this technology.

Foresight News: Will we see more L1s, L2s, or sidechains based on FHE? That's what I really want to ask.

Rand Hindi: I think all three will exist. FHE is no different from any other blockchain protocol; it does not change anything about blockchain consensus and security; it just adds an extra layer of confidentiality on top of what you initially wanted to do.

Foresight News: Let's go back to the speed issue for a moment. New technologies like FHE and ZK consume a lot of computing power, resulting in high costs. You just mentioned that it will become much faster in the coming months. Can you give us more phased expectations in this regard?

Rand Hindi: I have roughly introduced some performance numbers earlier. Today, we can do about five ERC20 transfers per second in FHE, with the balances and amounts encrypted. If we make some engineering improvements on GPUs, we can achieve 20 or 30 TPS.

If you want to reach 100 or 1000 transactions per second, you need ASICs. ASICs are not just for improving performance and speed; they can also significantly reduce costs. We have built an FPGA prototype internally at Zama that can match the performance of (NVIDIA) H100 GPUs on FHE, but at only one-third the cost. This is very interesting. Shifting to ASICs could not only greatly speed things up but also reduce costs to half or even a tenth. Of course, like every new technology, the first iteration of networks or FHE chains may be relatively more expensive; you might have to pay 10 cents for an ERC20 transfer, but that number will only go down.

We are now at a critical point where FHE is already good enough for most blockchain applications, and it is moving towards being able to do anything.

So think about it: if we can achieve 1000 or 10000 transactions per second in FHE, then everything you do on the blockchain can achieve end-to-end encryption. FHE will become as natural as HTTPS for blockchain transactions.

Ten years ago, I thought only about 10% of internet traffic was encrypted; today that figure is over 95%. The same will happen in blockchain; initially, perhaps 5% or 10% of blockchain transactions will be encrypted, but eventually, 95% will be encrypted, which is not unreasonable because it won't cost you more and will allow you to do more things this way.

Foresight News: In the crypto world, we often regard 2022 as the year of ZK, when basically every VC focused on ZK after a research paper was published by Paradigm. So what about FHE? Do you think there will be a year of FHE, or is this year actually it?

Rand Hindi: From my perspective, it feels like it is happening. I think FHE is becoming a big narrative, and investors are investing in FHE.

I have a Telegram group with 50 VCs, and I share company information with them. Whenever a company collaborates with Zama, I send it to our investor group and ask about their investment interest.

It's not just startups using FHE; we are now also collaborating with some large companies that are starting to use FHE for existing protocols. This is very exciting.

Foresight News: So how does Zama want to position itself in this large ecosystem?

Rand Hindi: Do you remember "Intel Inside"? In the 90s and early 2000s, when you bought a computer, you would see a sticker on it that said "Intel Inside," right? When you bought a computer, you knew the chip was an Intel chip, even if you couldn't see it. We believe Zama is the "Intel Inside" of confidentiality. You know, our technology will be present in every application and every blockchain you use, but you won't see it because it won't be consumer-facing; it will exist at the protocol level.

So, our vision for the company is that at some point, just like we transitioned from unencrypted HTTP to using HTTPS to encrypt the data we send, we will use something like HTTPS at the protocol level to encrypt every service we use. When that happens, you won't even know Zama exists.

Just like if we do our job well, you won't know Zama exists. Maybe there will be a sticker on the website that says "Protected by Zama," but that's it.