What Crypto Already Solved for Enterprise AI | Rami Akeela | Nera Systems
What Crypto Already Solved for Enterprise AI
There's a wall that enterprise AI is running into right now, and it's not a model capability problem. It's a data problem. The most powerful AI tools in the world need access to your most sensitive information to be useful, and the moment that data leaves your environment, you've lost control of it. For banks, hospitals, and accounting firms sitting on exactly the kind of data that AI could transform, that tradeoff has become a quiet, daily standoff.
The Conundrum at the Center of Enterprise AI
Rami Akeela has spent twenty years building privacy and cryptographic infrastructure for problems most people didn't know they had yet. Long before generative AI became a boardroom conversation, he was building zero-knowledge proof systems, technology that lets you prove something is true without ever exposing the underlying data behind it. Now, as founder of Nera Systems, he's applying that same premise to enterprise AI: the more valuable your data is, the less you've historically been able to use it. As Rami puts it, that doesn't have to be a paradox. There's a solution to it.
Locked Data, Unlocked
The conversation moves quickly from theory to scale. Once you understand that AI can extract insight without ever seeing the raw data behind it, an enormous amount of previously locked information becomes usable. Rami points to financial services and healthcare as the most obvious examples, but the use cases run far wider than that. Accounting firms running tax analysis on client data. Private equity firms conducting diligence. CPG companies working with retailer data they need to keep confidential while still extracting value from it. In every case, the constraint wasn't a lack of demand for AI. It was the absence of a way to use it without violating the trust the data depended on.
The Numbers Behind the Quiet Workaround
What makes this more than a theoretical concern is how widespread the workaround already is. Reports suggest a significant share of employees are using AI tools like ChatGPT without their employer's knowledge or approval, frequently with sensitive data attached. Rami describes hearing this directly from friends in healthcare, where the pressure to move fast collides daily with HIPAA obligations. The risk isn't hypothetical. It's already happening, quietly, inside organizations that have no visibility into it.
Twenty Years of Training for This Moment
What's striking about Rami's path is how little of it was planned around AI specifically. His earlier work spanned privacy-enhancing technologies, systems architecture for different applications and domains, and the discipline of taking complex research and turning it into something people could actually use. None of it was built with generative AI in mind. But looking back, it reads less like a series of separate bets and more like training for a problem that hadn't fully arrived yet.
The Grind Beneath the Cryptography
The conversation closes on something that has nothing to do with zero-knowledge proofs at all: what actually keeps a founder going through the parts of building a company that aren't the interesting parts. The fundraising, the pitches, the accounting, the long stretches where nothing visible seems to be happening. Rami's answer is grounded and honest. Not everything is a breakthrough moment, but the work underneath those moments is what makes them possible at all.
Listen to the Full Conversation
This episode is a clear-eyed look at a problem that's already inside every regulated industry, whether leadership knows it yet or not, and a conversation with someone who's spent two decades building toward exactly this moment without realizing it at the time.
Subscribe on Spotify: https://open.spotify.com/show/4F8uOLxiscYVWVGEfNxTnd
Watch on YouTube: https://www.youtube.com/@moneyneversleeps1814
Listen on Apple Podcasts: https://podcasts.apple.com/ie/podcast/moneyneversleeps/id1455819294
CHAPTERS
00:00 Cold open
00:42 What zero-knowledge proofs taught Rami about AI
02:05 Why it doesn't have to be a paradox
02:50 The biggest use cases: banking, healthcare, accounting, PE
06:40 Training for this problem his whole career
07:53 What's kept him building through the grind
10:13 Where to find Rami and Nera Systems
MoneyNeverSleeps — sharp riffs, big ideas, and real insights from smart people, in under 15 minutes. Hosted by early-stage investor Pete Townsend, GP at Norio Ventures, sitting at the intersection of crypto, fintech, AI and onchain finance.
Website: moneyneversleeps.ie
LinkedIn: https://www.linkedin.com/in/petetownsendnv/
X/Twitter: https://x.com/PeteTownsendNV
Instagram: https://www.instagram.com/petetownsend353/
Email: info@norioventures.com
[00:00:00] Rami Akeela: Not everything is a Kodak moment.
[00:00:01] Rami Akeela: But for that Kodak moment to happen, a lot of work needs to, be put into it.
[00:00:08] Rami Akeela: You start to feel joy, doing the mundane stuff and the boring stuff and the hectic stuff
[00:00:14] Rami Akeela: and I think that's what I'm experiencing,, building these startups. I feel like it, it has a purpose.
[00:00:21] Rami Akeela: No matter how tired, you keep going, even if you're running on fumes, doesn't matter.
[00:00:27] Rami Akeela: 'Cause you believe in this thing, you wanna see it, in the hands of more people. You wanna see it add more value, fix real problems. So you keep going. And it doesn't matter how tough it gets.
[00:00:37] Rami Akeela: This all means nothing when you see the fruit of your work.
[00:00:41] Pete Townsend: This is Money Never Sleeps. Sharp riffs, big ideas, and real insights from smart people.
[00:00:46] Pete Townsend: I'm Pete Townsend, GP at Norio Ventures. Let's go
[00:00:53] Pete Townsend: Rami Akeela has spent twenty years building privacy and cryptographic infrastructure that most people didn't know they needed yet. Now, he's the founder of Nera Systems, solving the conundrum that's about to define enterprise AI. The more valuable the data, the less you can use it. And a quick disclosure, I invested in Nera via Techstars in 2025, so I've got a front row seat on this one.
[00:01:18] Pete Townsend: Rami, welcome to the show
[00:01:20] Rami Akeela: Thank you for having me. Appreciate it
[00:01:22] Pete Townsend: You've been building zero-knowledge proof systems for years. What did that teach you that most AI founders simply don't know yet, Rami?
[00:01:32] Rami Akeela: in zero-knowledge proofs, the idea was, I could still, prove that something is right and not have to expose my most sensitive data. And, and of course, on top of that, so many things were built.
[00:01:44] Rami Akeela: Some, some of the coolest applications were built,
[00:01:47] Rami Akeela: and the same premise can actually be, applicable in, in AI again, which is I can do so and so with AI without actually exposing my data. Of course, everybody has sensitive data, whether it's patients' records, transactional, records,
[00:02:03] Rami Akeela: But the idea is how can I actually build the system on top of this? It's not a question of can I actually use it, securely and privately? No. We know we can. Again, it's a matter of, how do we actually , build, , systems that would allow for AI intelligence to be extracted, while protecting our data?
[00:02:27] Rami Akeela: Instead of actually saying no. Everybody is basically saying, "If I-- If this is too sensitive, then I shouldn't be using it with AI." So the, the answer is basically either/or.
[00:02:38] Rami Akeela: And that's, that's what we're trying to crack. It's doesn't have to be a paradox.
[00:02:43] Pete Townsend: Almost like saying science has proven that we are able to get a starship from Earth to Mars. What do we actually do when we get there, right?
[00:02:55] Rami Akeela: Exactly. Exactly. Yeah. Because,
[00:02:57] Rami Akeela: it is a problem of people not realizing, that such solutions,, exist. And so what is happening, Pete, now is the fact that we have locked data. This is too sensitive then-- and we can't use it with AI because it then will be exposed.
[00:03:12] Rami Akeela: Now that they know that this can be done, all this gets unlocked, right? So this is data we can process. This is insights that we can get from most, our most sensitive data. And now that they know that this can be done, the opportunity becomes a lot bigger.
[00:03:28] Pete Townsend: Give me a, a couple of examples. You know, when you and I first met, Rami, the things that immediately jumped out at me were the sensitivity of financial data in terms of retail customer financial services data, banks, fintechs, whatever, and also healthcare data. So your most sensitive patient data that you have when hospitals are tracking everything about you when you go to see them to be treated for a certain thing.
[00:03:59] Pete Townsend: And that is incredibly sensitive data. Now, most people don't think about the fact that, hey, could this healthcare worker just take all of my medical files and pop them into Claude and say, "Hey, can you look at everything else here and see what's really going on with this person?" Or that a large bank would take millions and millions of lines of data and feed it into ChatGPT with all of their most sensitive client data and say, "Can you give me some analytics here on the demographics of all the people who might be eligible for spending more and creating more of a revenue opportunity for us on interchange on credit cards, or give us an opportunity to cross-sell?"
[00:04:46] Pete Townsend: Banks, financial services providers, healthcare wouldn't just drop this data into ChatGPT and/or Claude today, because they know that once that data is with inside the walls of OpenAI, in ChatGPT's case, or Anthropic, in Claude's case, that it's open season, and so you're not gonna do that. Are these kind of the two biggest use cases that you're pointing to for being able to run AI on datasets securely and privately?
[00:05:16] Rami Akeela: So yes, if you're in a bank or some financial firm, then y-you're absolutely right. You shouldn't be uploading these records to ChatGPT and Claude to get insights,
[00:05:26] Rami Akeela: The reports are basically saying it's close to seventy, eighty percent, of employees use ChatGPT secretly. So we know this is happening. Employers know this is happening.
[00:05:37] Rami Akeela: In hospitals, I know I have friends in the medical field, and they tell me they... Nurses, doctors are always uploading, stuff to, ChatGPT to get answers in seconds
[00:05:48] Rami Akeela: because they're under a lot of pressure But they do that violating HIPAA, o-obviously,
[00:05:52] Rami Akeela: We know this is happening. But those are the most obvious ones. But then you have cases where, we're talking to accounting firms, and basically an accountant needs to run, say, taxes on their data.
[00:06:04] Rami Akeela: They know they can just use AI and save, a lot of time and actually take on more.
[00:06:09] Rami Akeela: Again, some people are just doing it, violating all kinds of client data confidentiality rules around that.
[00:06:16] Rami Akeela: So this is a solution for them. This is a solution for private equity firms where they need to run diligence.
[00:06:22] Rami Akeela: We worked with CPG firms, on their retailer data so they can, then sell to manufacturers
[00:06:29] Rami Akeela: you can see it playing,
[00:06:30] Rami Akeela: out in every vertical because, again, if that solution exists, then I gotta think of ways I can utilize this solution without having to,, make, or accept any compromises.
[00:06:45] Rami Akeela: When you talk about sensitive data, you're no longer talking about banks and hospitals and defense. You're talking a lot about, about every single business, and how they should protect it.
[00:06:55] Pete Townsend: There are big companies who have
[00:06:57] Pete Townsend: their own internal AI models, their own LLMs that they're running, and that where all of the data stays within their own closed walled garden. But only the biggest and best companies with the largest tech budgets can really afford to do that.
[00:07:18] Pete Townsend: Looking at the personal journey that you've gone through, Rami,
[00:07:21] Pete Townsend: with your earlier startups in mind, was there a certain point where you realized that you've been training for this problem the whole time?
[00:07:28] Rami Akeela: You gotta trust the process. Just, just go through this journey. Learn, make mistakes, but keep going. Every time you fall, get back on your feet and keep going. That's the main lesson.
[00:07:40] Rami Akeela: I'm happy that I, I stayed, on my path because, I didn't realize that I was building to this point.
[00:07:47] Rami Akeela: How I, I got to understand privacy en-enhancing technologies, how I, I got to build systems for different applications and domains,
[00:07:54] Rami Akeela: And I was doing that while I doing PhD. It was like, how do I take a really complex problem and, and put it in simple terms so people can understand?
[00:08:02] Rami Akeela: How do you actually go from research papers to actual products that you can ship?
[00:08:08] Rami Akeela: It all led to one point. It was like,
[00:08:10] Rami Akeela: you need to build this, before it even, becomes, a tangible problem because you can see it, because you've been trained to see problems in the early days.
[00:08:21] Rami Akeela: And so yeah, every day I'm reminded that something happened in the past, had a purpose, had a reason, and it's all to get me to this point
[00:08:31] Pete Townsend: Through this whole process of the prior startups, your time in academia, everything, that, you know, getting to this point now with Nera, what's kept you building through the parts that had nothing to do with cryptography, all the other things that you need to do as a founder with fundraising, with pitches, with the grind underneath all of that?
[00:08:51] Pete Townsend: What's kept you going?
[00:08:53] Rami Akeela: The fact that I'm doing something of value outside of my comfort zone. But you know, the more I do it, the more I enjoy it.
[00:09:00] Rami Akeela: Connecting with people, pitching, handling the boring stuff like HR and payroll and accounting and, and filing taxes for the company.
[00:09:09] Rami Akeela: Doing the necessary, part of what would allow for, for me and Nera to make this happen. On a personal note, for example, , I'm a husband, , I'm a father.
[00:09:22] Rami Akeela: That's what you gotta do. And I love it. And sure, it's a lot of work, right?
[00:09:26] Rami Akeela: Not everything is a Kodak moment. But it, but for that Kodak moment to happen, a lot of work needs to, be put into it.
[00:09:35] Rami Akeela: You start to feel joy,, doing the mundane stuff and the boring stuff and the hectic stuff and I think that's what I'm experiencing,, building these startups. I feel like it, it has a purpose.
[00:09:48] Rami Akeela: No matter how tired, you keep going, even if you're running on fumes, doesn't matter.
[00:09:53] Rami Akeela: 'Cause you believe in this thing, you wanna see it, in the hands of more people. You wanna see it add more value, fix real problems. So you keep going. , And it doesn't matter how tough it gets. And you, you know some of the stuff that I went through and still going through.
[00:10:08] Rami Akeela: This all means nothing when you see the fruit of your work.
[00:10:13] Pete Townsend: Yeah, I know what you mean. And it's like I'm staring at this book right now in front of me, which I've read only a bit of. It's David Whyte, "Essentials." He's a, he's Irish poet, and that word essentials, what are the essential things you must be doing? I keep that on my desk so that I focus on that, because it's sometimes that extra 5% that you've gotta put into your day that makes 50% of tomorrow happen a lot better.
[00:10:40] Pete Townsend: And so it's just whatever works to keep moving is what you need, right? Very good. Very good. Well, Rami, listen, that's a great place to end it.
[00:10:52] Pete Townsend: Where can people find out more about you and Nera Systems?
[00:10:55] Rami Akeela: Nera.systems, N-E-R-A.systems. - The website should have, plenty of information of what we're doing. And it also has under solutions, you'll find our chat app, where people can try it out for two, three weeks and see, the value it can add to your data. But me personally, I'm, happy to connect with people over LinkedIn.
[00:11:16] Rami Akeela: It's under my name, Rami Akela, R-A-M-I, , A-K-E-E-L-A. And feel free to reach out, to talk about anything data and AI related or really anything else out because al-I'm always happy to talk about more stuff.
[00:11:29] Pete Townsend: I could attest to that. I could attest to that. Wonderful. Well, listen, thank you Rami. I really appreciate you joining the show.
[00:11:35] Pete Townsend: And to all of you out there, thanks for watching and listening. Don't forget to follow or subscribe wherever you get your podcasts. Helps others to find the show, and it means a heck of a lot to me Till next time. See ya
