Illia Polosukhin / NEAR Protocol

Illia Polosukhin on NEAR's privacy features and AI agents in 2026

InterviewOctober 1, 202624:51

In this episode

Illia Polosukhin, co-founder of NEAR Protocol, joins Ashton Addison on the Crypto Coin Show to break down everything NEAR has shipped since June: privacy by default across trading, confidential perps, IronClaw for AI agents, tokenized stocks through NEAR Intents, quantum-safe signing on mainnet, and the new Bitwise NEAR ETF.

We also cover the shield that stopped Bitget hack funds from moving through NEAR Intents, why liquidity is the key to tokenized stocks, the NEAR distribution for early confidential intents users, and what the team wants to ship before 2027.

Key takeaways
  • NEAR has implemented privacy by default for trading and confidential perpetuals, addressing concerns about exposing financial strategies to competitors and malicious actors.
  • NEAR Intents aggregates liquidity across multiple blockchains and traditional exchanges, enabling large tokenized stock purchases like $10M Nvidia shares without excessive slippage.
  • The Shield system prevents stolen funds from moving through NEAR Intents by intelligently evaluating incoming assets and flagging high-risk transactions to the network.
  • NEAR's privacy model is practical rather than sovereign, allowing government requests to decrypt information through multi-party infrastructure while protecting ordinary users from constant surveillance.
  • Early Confidential Intents users received a free token distribution to encourage actual usage of privacy features rather than passive token holding.

Chapters

Read the interview

NEAR's Polosukhin on 'pragmatic' privacy: lawful decryption, not Zcash-style secrecy

Illia Polosukhin, co-founder of NEAR Protocol, spoke to Crypto Coin Show in October 2026 about confidential trading on NEAR Intents, tokenized stocks, private AI and the day's NEAR ETF launch. He explained why NEAR is building compliance-friendly privacy rather than Zcash-style secrecy.

Pragmatic privacy, with limits

Ashton: NEAR Intents now lets users open Hyperliquid perps positions. What is private and what is public? 00:39

Illia Polosukhin: When you use near.com and NEAR Intents, you can take positions on Hyperliquid directly. Usually there's a long provenance chain showing where your money came from, because you're bridging from Arbitrum or another chain. With NEAR Intents you don't have that chain. You open positions on Hyperliquid straight from NEAR's confidentiality pool. The positions themselves are still public and visible for now. We're working on privacy for positions too, but that is a harder lift and will take more time.

Ashton: Why is there demand for privacy? Is it coming from enterprises that want to hide their strategies? 02:14

Illia Polosukhin: Crypto has matured from a playground for speculative trading into fundamental financial infrastructure. More real-world assets are being issued and more real-world payments processed, and that brings clear requirements. If you're a public company or a business and most of your transactions happen on chain, you're leaking all of your information to everyone, including competitors and adversaries. I think that's been the reason adoption has been slower than it could have been. People are still staying on centralized exchanges. Take on-chain cards. Every swipe is a public transaction, so somebody can track when you spend money and use that for malicious purposes. That is literally worse than the traditional system. We also saw perps traders whose accounts got linked. Somebody noticed a short position, and the trader had to apologize by showing they were actually long and running a delta-neutral strategy. Those things don't happen in a normal financial system. Privacy on its own isn't enough. You need a great product and great liquidity. That's why we built Intents first and put privacy on top, where liquidity already exists. We're not trying to facilitate crime. If we're building a new financial system, we can't say somebody else needs to figure it out. That's why we call it confidential. If there is a legitimate government request for a viewing key, our multi-party computation infrastructure can decrypt information for lawful purposes. This is not sovereign-level privacy like Zcash. It's pragmatic privacy, so that every transaction doesn't get telegraphed to everyone. If you're a regular user, you have no problem. If you're a criminal, this is not a place for you.

Read the full interview →

Clips from this interview

By the numbers

50MTransactions blocked by Shield system
$70MInitial Confidential Intents TVL target
$183MCurrent Confidential Intents TVL achieved

Transcript

Read the full transcript 3,861 words

I'm Ashton Addison from the Crypto Coin Show. And today on Blockchain Interviews, we welcome back Illia Polosukhin, co-founder of NEAR Protocol. When we last spoke in June, NEAR had just switched on Hyperliquid perps. They've now made privacy the default across trading, launched IronClaw for AI agents, brought tokenized stocks on board, put quantum-safe signing on mainnet, and the NEAR ETF, one of the ETFs, has just launched on the New York Stock Exchange.

So much has happened in the last two months and there's lots to talk about. Illia, welcome back to the show.

Thank you. Yeah, it's been a busy summer.

Definitely. And the market is rewarding us and rewarding NEAR as well. And I think the community is happy with the progress of the protocol and everything that's being developed. NEAR has been pushing confidentiality across balances, transfers, swaps, and leverage trading now. Can you talk about the opening of perps on NEAR and the confidentiality part of that and what's private, what's public?

Yeah, for sure. So this is something we announced in the last couple of weeks: when you use near.com, NEAR Intents, you can effectively take positions on Hyperliquid right now. Usually when you're doing this, there's a long provenance chain of where your money came from, you know, you're usually bridging it from Arbitrum or connecting to some other bridge from other chains.

And so with NEAR Intents, you effectively do not have that provenance chain. You are opening positions directly on Hyperliquid from the confidentiality pool of NEAR. Now the positions right now are still going to be public and visible. And so this is something we are working on to improve as well, and create actually privacy for positions, but that is a harder lift and it will take some more time.

Definitely. And can you talk about the demand for privacy? Is there a strong demand on the enterprise side and corporations looking to just hide their strategies? What's the intent behind privacy for all?

Yeah, I think there's a combination of things that are happening over the past year. Crypto has matured from being a playground for speculative trading to starting to be more and more used as fundamental financial infrastructure. More and more real world assets are being issued, more and more real world payments are being processed. And as part of this

there are very clear requirements that come. There was actually news yesterday where I think the Chinese government was highlighting that blockchain is not private, everything's transparent, you're effectively communicating to everyone everything you do. Which is kind of funny that they're highlighting that as a problem. The reality is if you're a public company, if you have funds,

if you're a business and the majority of your financial transactions are happening on chain, you're effectively leaking all of your information to everyone, both your competitors and your adversaries in a broader sense. We obviously see a lot of malicious actors operating as well.

I actually think that's been the reason for slower adoption than it could have been. People are still staying on centralized exchanges and not able to participate in the broader financial infrastructure we've been building on chain. For us that's really what we're enabling: going away from this world where every action you do is telegraphed to everyone and stays available forever for anyone to do any analysis and attacks later.

To give you a few examples, we now have a few different on-chain card solutions, but every time you swipe, it's a public transaction on chain. So somebody can effectively track when you're spending money and use this information, including for very malicious purposes. That is literally worse than the traditional system.

The same goes for trading, for strategy, for perps. We saw perps traders whose accounts got linked, so somebody noticed this person has a short position, and they effectively had to apologize by showing they were actually long and it was a delta neutral strategy to capture profit. Those are things that don't happen in a normal financial system.

So I think it's fundamental, as we build this on-chain economy, that privacy is an enabler. Now privacy on its own is not enough. You need a really great product, you need great liquidity. That's why we built Intents first and then privacy on top of Intents, enabling privacy where there is already liquidity.

And we see amazing growth, more and more funds are moving into this. I do want to be very clear: we're not trying to facilitate crime. This is something we are very strongly positioned on. There's a lot of controversy right now around this in the crypto space, but if we're building a new financial system, we cannot say we don't have any responsibility.

Somebody else needs to figure it out. No, this is on us to figure out and really deliver a product that's better than the traditional system. That's why we call it confidential: if there is a legitimate government request to receive a viewing key, our multi-party computation infrastructure can decrypt information for lawful purposes. So this is not sovereign-level privacy like Zcash.

It is very much pragmatic privacy, so that every transaction doesn't get telegraphed all the time to everyone. If you're a regular user, you have no problem. And if you're a criminal, then this is not a place for you.

Definitely. From what I understand, over the last two years the future is becoming compliant privacy, without too many fiery hoops to jump through. We want to make it as easy as possible, but if you need to be compliant for the sake of stopping crime and malicious actors, that can happen. I saw your long tweet thread about the responsibility.

The responsibility is on us to build this technology, and I agree with that. I'm glad somebody's taking responsibility for the software, and as a community, crypto needs to step up. On stopping malicious actors, the recent hack through Bitget and a vulnerability lost hundreds of millions, and from what I understand some of that was attempted to move through NEAR Intents and the confidentiality, and there's actually a software shield that stopped that.

Can you talk about that?

Yeah, exactly. The way we approach this is we want code as law, but that doesn't mean the code needs to be dumb. We want an intelligence system that is able to assess the assets coming in and decide not to participate in a transaction when there's really high risk or known information that these are stolen funds.

And we can telegraph this information to other partners so they can decide as well whether they want to participate. That's the purpose of Shield. We're still working on it and rolling it out, but it also served us when there was a Litecoin attack and a double spend on Litecoin. We detected the malicious transactions with an earlier version of this system.

It actually works every day, and we're going to show more reports on this. It's stopping a lot of funds that have clearly been stolen and that people are trying to launder. The default path is to decline the transaction and inform the solvers and the network that these are high-risk funds coming from malicious sources.

That is the headline number, fifty million declined. The big part here is that as we do this, we signal to criminals not to use this platform, and so all the users who are using it are actually safer. They don't get in trouble when they try to off-ramp or pay for something and the other party says your source of funds is tainted.

That's a huge benefit for anyone using Intents: we're trying our utmost to make sure your funds are safe and secure and you have a clean source of funds when you go to banks or other financial institutions.

Definitely. And I'd love to flip the script to the legitimate side. You mentioned the tokenization of real world assets. I saw the recent partnership of NEAR with Ondo for tokenized stocks, which I'm sure will see skyrocketing value locked as more major players move into tokenized stocks. We're seeing that on Robinhood chain as well. So what's capable through NEAR Intents right now with regards to tokenized stocks? Can you buy those assets with Bitcoin? Can you do it all privately?

Exactly, yeah. You have the ability to buy any of the stocks with any asset we support: Bitcoin, Zcash, USDC, USDT across all chains, Ethereum, Solana, whatever else, Doge, Shib, Trump. You can buy all that directly, you don't need to bridge, swap, or figure out gas fees. And if you're on near.com or other front ends that

work with confidential intents, all of this stays private. Nobody needs to see your portfolio allocation. The important part of why we think Intents is the right mechanism for RWAs is that one of the problems historically limiting growth of equities on chain is liquidity.

And it's been getting more and more fragmented. There's a little bit on Solana, a little bit on Ethereum, a little bit on Base, a little bit on Robinhood, and we see very thin AMMs. If somebody wants to buy, say, ten million dollars worth of Nvidia, which is a highly liquid stock, ten million is super easy to buy on Nasdaq, but almost impossible to do on chain.

Or you need to go to Ondo directly and become KYB'd and do all of the work there. So with Intents you describe your intent: I need ten million worth of Nvidia. And our solvers can go to Ondo and mint that ten million and give it to you within that one intent fulfillment process. Again, this is an example where something like Shield is very important, because we need to make sure the funds coming in are legitimate.

But this now allows anyone to tap into the liquidity of Nasdaq on chain. You don't need to be on Nasdaq, and NEAR Intents becomes an aggregator not just of on-chain and centralized exchanges, but now Nasdaq and other liquidity trading these RWAs. That is the biggest benefit of why NEAR Intents is a better approach, because we're unifying liquidity across all of those venues.

Yeah, I think that's probably the most important factor. I've been talking about security tokens since 2018, and the main issue has always been liquidity. Everyone wants to tokenize, but when it comes to trading, you're the only stock on the tokenized exchange and there's nobody to trade with, and the project just fails. That's why it's been a ten-year hurdle. Combining the liquidity matters.

It was a similar issue in crypto as well. We've seen that with Hyperliquid and a merging of different platforms agreeing that we need to increase liquidity between chains. I think we've solved a lot of hurdles in that sense. Now combining digital assets with tokenized stocks only makes sense and it's the next step.

Yeah, and similar to how the intent approach has been a robust way to tap and aggregate liquidity across chains, we can see more and more real world assets coming online and tapping into that. And it's not normal that your portfolio is public to everyone. I use this example: I have over a hundred accounts on NEAR and dozens of accounts on other chains,

because for every asset I buy, every investment I make, every payment I make, I have a separate account, and I used to have a spreadsheet to track all of this.

Yeah, it's too much work. We just need the confidentiality so that we don't have to do that.

Exactly. And near.com makes this experience so easy that I don't need to think about it. I need to pay someone, I pay someone. I don't need to think about how it's going to be tracked or who else I paid or what else I invested in. The experience is so much smoother.

It's a true entrepreneur's journey: you're solving your own problems. As the NEAR token has been appreciating over the last month and the whole industry has been moving up, there was a distribution of some sort at the NEAR 333 level for early users that actually used confidential intents. Can you talk about that and why it's important to actually use the software and test it out, not just hold NEAR?

Yeah, so there were two thoughts there. One is NEAR never had an airdrop-type thing, and we really wanted to engage the early community and let them participate in something where they actually try the product. As near.com was launching, the idea was to have a place where people can do something,

interact with the product, but also have a common goal. There were effectively two pieces in this program. One is we needed to hit a confidential TVL of seventy million dollars. As we were launching confidential intents, you do need a reasonable amount of liquidity, funds in the confidential pool,

so that it actually provides you with this confidentiality. The bigger the pool, the more money you can stick there, and people will not be able to see where things came in and out. So the first milestone was seventy million, and we've actually just crossed 180, I think, as of today. Yeah, 183 right now.

So that was the first milestone, and creating this as a goal for the community was important: create common goals and drive attention around them. It served as a way to try the product, but also to band together and talk about this as one of the core metrics we want to measure and grow and use to attract people.

Definitely. The NEAR holders I know were happy to see that their usage of the platform pays off. And from what I understand, there are more thresholds, NEAR at 666 with other qualifications, so people should still try out the platform, and there's more of that to come.

Yeah, at least one more. The way we think of this is we want people to continue exploring the product as we launch more features and then continue hitting a higher and higher bar on overall confidentiality and usage of the product.

And on the AI agent side, last time we spoke IronClaw 1.0 was launching. I feel like AI agents are gaining more adoption as the connectors for AI become easier for non-technical people to have agents trade and pay for them. What's the stage of AI agent usage in NEAR right now and how can we grow it?

The NEAR AI team has been focused on private AI inference. We just also listed on OpenRouter, so demand is starting to come from there. Brave has been rolling out and we're going to have an announcement shortly with them. We see more and more people demanding private AI across all these venues, because the news is not looking positive: your data is being used, we've seen mathematicians concerned about that,

we've seen corporations worry that their work goes to these frontier companies and we don't know if they're going to train on it, use it themselves, or build an alternative company. So I think privacy and sovereignty have become important topics for all of these participants. At the same time, open-weight models have gotten way better.

Compare this to even three months ago, but especially when we launched at the end of last year and beginning of this year: everybody was still in the prototyping stage, so privacy didn't matter as much and open-weight models were pretty far behind. Now I'm using open-weight models for NEAR AI in OpenCode, for example. You have GLM Flash, you have DeepSeek Flash, and these are astounding models for a lot of use cases.

They're fast and they're fully private, so I don't need to worry. Similarly with crypto, I use Opus as well, and I do have cases where they decline to answer. I was investigating a report I got and the model said this is touching cybersecurity, I don't want to do this. So I need to go and use another model anyway.

It's the same mental overhead: now I need to model in my head what OpenAI and Anthropic think I'm not allowed to do, and instead with open models and privacy, I have the guarantee that I can just use it. Usage is scaling and we'll be publishing more stats on this. Obviously compute is the next frontier, compute and energy, so we're also starting to look more into that.

Private AI is now critical for any enterprise and individuals should also care about it. You need a great product and then privacy is the baseline.

I agree. I've been getting deep into the centralized models, and many people do for convenience, not knowing the alternatives. These open source models don't have a marketing team behind them, so regular people, even crypto and tech folks, don't know there are private options until it's too late, until these centralized models start copying your business and do it better than you. So I'm glad we're working on that in the blockchain industry. With IronClaw and the AI agents, how easy is it for people to whip something up with NEAR AI without having to understand all of the different open models?

Super easy. It's compatible with all the standard APIs, so whatever agent you're using, including Codex or Claude Code, you can switch them to use NEAR AI. OpenCode too, and on IronClaw it's basically the default. You can set up API keys, and we have a few different options for how people can participate.

The key idea is making it really straightforward for anyone who's using agents. And you can always ask your agent to configure your NEAR AI setup.

Definitely, until it decides that it's going to retire itself and doesn't want to. Then it has a brain of its own.

Well, there's actually one thing we're considering as a product. Because you have not just privacy but also verifiability coming from our stack, you have full provenance: a tamper-proof log of every AI interaction running through us, which is super useful for compliance and regulatory work, for insurance and healthcare companies, which need all that information.

Given that we don't even know if our models are hacking things, we can provide a service that authenticates, privately over tool outputs, that your agents are not going off and hacking things. I call it the anti-swarm: detecting if your agents are starting to swarm in your organization.

For any corporation that's worried about this, this is effectively a solution. If you use private verifiable inference, you can also have services attached that provide you with protection and a compliance and regulatory framework around this that is in your control and sovereign.

You mentioned at the beginning that today was the launch of one of the NEAR ETFs from Bitwise, I believe. Can you talk about what access that creates for investors looking for exposure to NEAR? And Bitwise was describing NEAR as AI money. Do you agree with that?

On the access side, there are still a lot of people who are not on chain and don't have a good way to custody funds. There are a lot of indices and pension funds, Roth IRAs, 401(k)s that are by default only able to buy assets on Nasdaq and the New York Stock Exchange.

So this opens up the opportunity for all this capital to participate in the crypto economy. It's a massive opportunity, and there are a lot of financial advisors and the broader financial ecosystem getting educated as part of this about what NEAR is building and how it's going to change the world. We should not discount that as a really big and important piece.

On the messaging side, our global vision is that AI is going to be the front end and blockchain is the back end: the markets, the identities, the trustless layer of value coordination. It needs to be private, it needs to be user-owned, the whole stack needs to work together. And the NEAR token is in between, facilitating that.

So in a way it is money for AI, because we're all going to interface with computing through AI. It also is very much the programmable money concept, allowing you to build programmable markets directly, and we see Intents very much like that. It's also the agentic economy, because everybody will have an agent using NEAR AI for inference and NEAR Intents for making actions across different parties.

All of these concepts are the same at the core. It's about creating this ecosystem of value for people and agents to interface in one place.

Definitely. I think it's great that the financial advisors are getting educated and so many traditional investors are looking for different ways to get exposure to AI but don't want to get it on chain, because they've been in traditional markets for thirty or fifty years. It opens the floodgates for more capital to get interested and slowly make its way on chain, and slowly use blockchain on the back end without knowing it. They're just using AI.

I think it's a great, exciting step and milestone for today, so congratulations. There's so much your team has worked on in the last two months since we last spoke, and so much you're doing simultaneously. What are you most excited about, and what do you want to get done before the end of this year, moving into 2027, which will probably be huge for both AI and blockchain?

There's a bunch on the roadmap that we're going to be launching, and I don't want to spoil too much. Overall we continue growing across all of these dimensions. Near.com becoming more and more a single place where you can have not just all your financial activity but also use AI to do it will be a big piece. Making sure it's available on mobile, available where you are.

And then continuing to expand on the private AI side: providing tools that make sure people own their data and their interactions with agents, and are able to control, detect, or monetize them. That's a big dimension we're looking at.

And continuing to expand what assets Intents provides, not just blockchains but also real world assets. How do we bring more of the real economy in, both work, like hiring agents to do work, and real physical goods. The more blockchain is interleaved with real world interactions, the more we achieve the vision that

we all set out to do.

Amazing. Looking forward to it. Thank you so much for the time, Illia, and congratulations on all the success so far. Let's continue to build the future of private AI and private, confidential trading, Intents, all of that in the blockchain industry. I'll be following NEAR closely and would love to have you back on in the near future. Thank you so much for the time.

Sounds good. Thank you for having me.

Related coverage

More interviews

Browse all 1,088 interviews

Get new interviews firstCCS Insider, the free newsletter from Ashton Addison. Twice a week.