The Next Era of Validators

Transcript

Korea's validators… Yes, hello everyone. Today's panel discussion is titled "The Next Era of Ethereum Validators," and that's what we'll cover. There are quite a lot of people running validator businesses in Korea, and leaders from the companies at the forefront of the industry have kindly joined us for this session. I'm Hudson Jang, and I'll be moderating this session. I work at Four Pillars as a validator engineer so being able to bring these leaders together in one place for this conversation is a real honor for me.

Before we begin, I'd like to ask each of you to briefly introduce yourselves. Our company was formerly B-Harvest, and we recently changed our name to Altus. I'm Hyung Yeon Lee, founder of Altus. Hello. I'm Youngseok Yang, CEO of the validator company Nodeinfra.

Hello, I'm Joonkyo Kim, CTO of DSRV. Thank you. I've been in the validator industry for about 4 years myself, but with CEOs and CTOs… in earlier sessions it felt like chatting with peers, so I went in relaxed, but now I'm feeling a bit nervous. Whenever I do a session like this… since everyone here is friendly, I'll try to keep it relaxed as I go.

This panel will run for about 40 to 45 minutes. Since there are 3 panelists, please take about 2 minutes on each topic I raise, and if you have thoughts or questions for the other panelists, feel free to jump in. The first topic is the barriers to entry in the validator business and its concentration. AI tools have been advancing rapidly, and the know-how of running validator nodes has built up a great deal compared to the past.

So I think the technical barrier itself is getting much lower. In the past, there were lots of very small validator companies and startups popping up, but these days, more and more, the small players are dying off and the market is consolidating around big firms. Why is that? And for the validator business — the blockchain validator and infrastructure business — where do you see the real barriers to entry?

Well, the validator business moved away from the center of our company quite a while ago, so my speaking about the validator business may not be entirely appropriate. Technically speaking, our view is that from about 5 years ago, the technical barrier had already collapsed. When chains launched with validator competitions, we took part, often won 1st place, made many technical contributions, found lots of bugs on GitHub and built tools, but in terms of our business, none of that translated into delegation volume.

In the validator business, to generate revenue, you need delegations, and delegations ultimately come from token holders. It used to be but these days it's mostly VCs, so it comes down to how good your connections with VCs and institutions are, and how well you build relationships to win delegations — that's the business's real barrier to entry. I see it as networking; the technical barrier has largely collapsed. That's how I see it. Our company also shifted away from validators toward technology development.

Starting late last year. I strongly agree with what was just said. Networking matters more than technology, but these days, beyond that, with institutions and the like, there seem to be knowledge barriers networking alone can't solve. On top of that, far fewer new chains are launching than before, so the barrier to entry around new chains has risen sharply — that's how I see it.

We're still running validators, so I'm not quite sure how to answer this, but as mentioned, the technical barriers have dropped a lot, and since validators sit somewhere between DevOps and communications, people tend to see them as neither one thing nor the other, and that has made things a bit difficult. With AI coming in, we've also had to compress our operations further, which has been a challenge.

So what could become the moat? What could become the barrier to entry? When I think about it, the key issue is trust. Delegation means someone is entrusting their money to someone else, and as you all know, accidents do happen in blockchain. People delegate knowing they could lose their money or that accidents can happen, People delegate accepting that risk, so in the end I think they relied on 7. So in the end, track history, or how much potential a team has, became the easiest basis for choosing, I think.

On top of that, as mentioned, institutions are coming in. Retail is hard to revive right now, institutions are entering, and in Ethereum's ecosystem in particular, as institutions enter, validators that were running on delegation are declining, and on top of that, server costs have risen, indirectly because of AI. All of this combined has made optimizing operations a lot harder. And as institutions' grows, they're demanding more.

For example, in Korea, once revenue passes a certain level, you need rigorous ISMS certification, some clients require SOC certification, and with custodians, some require S2 or S5 you need to have and to get you need about 3 teams [?]. All of these combined are probably what make it hard for small validators to break in. Thank you all three for those great answers.

As someone who's been in this field, and in a way one of the people hit hardest by the validator industry's contraction — the company I used to work at suddenly shut down. To sum up: customers are being reshaped around institutions, and running costs are rising due to AI and more, while expected returns are falling sharply. As customers shift to institutions, the cost of meeting their standards rises, and track record matters more, so new validators find it hard to emerge.

That's what the three of you said, and I strongly agree. I understand Altus and Nodeinfra, and DSRV as well, have been expanding substantially into businesses beyond validators. The next question I prepared is about the issuance problem. You may be scaling back your validator businesses, but the validator business's revenue comes almost entirely from rewards. The most critical recent issue in Ethereum — or the most controversial — is a proposal called EIP-8363.

In Ethereum, issuance only grows with the square root as the staking ratio rises, but as inflation keeps growing it becomes a burden, so there's a proposal to bring it down toward 0. As people who have run validator businesses, what do you think reducing issuance means, and how would it affect your business? by changing the curve the issuance adjustment would be phased in linearly over about 1 and a half years, as I understand it.

If that's adopted, our revenue would immediately be cut in half. Looking at the rationale behind it, in the long run, as the Ethereum ecosystem keeps to make the ecosystem healthier, it's something we should consider and apply if needed, but I think it's too early. The reason is that in a blockchain system, everything runs on incentives.

Security, too, is built on incentives, so if you change the incentives of validators at the very base layer, we need to think hard about what happens to solo stakers and small validators. Things could even end up, around the Ethereum Foundation more centralized — no one knows how it'll play out, but it could become more centralized. As I recall people in the forums and economics and there's a lot of that — if you change the curve,

could it become something that abruptly brings down the market itself? That's my concern. Fortunately, is coming. Among the many proposals for making this ecosystem thrive, cutting rewards is easy to understand, and the effect seems immediate and visible, so it's and gets discussed a lot.

As was said earlier, security is the foundation, so I think it's better to use inflation to invest in security — that's my personal, business view, but among the many options, I personally wonder why this one needs to be emphasized so much. But I also understand it, because the numbers show up right away.

With other measures, it's quite hard to predict what results you'll get. When you spend money, there are many uncertainties, but this money looks like it's leaking right now, and cutting it seems to save money, so it's easier to judge and evaluate, which is why people lean that way. But personally, from a security standpoint, I think it might be better to protect this through other measures — that's my view.

In our case, regarding Altus's Ethereum validator operations, we're not really in that position ourselves. What we've felt watching other chains is that as validators' customers keep shifting from retail to institutions, it's natural for the market to reorganize around supply and demand that fits institutional needs.

For retail, whether 100 or 200 people delegate, charging fees by delegated amount makes sense. But for an institution, whether it delegates 100 or 200, the validator's cost is the same, so why should it pay 2 times the fee? They'll start asking questions like that. What happened behind the scenes — white-label validators, node services and so on — in Web2, these became institutional contracts priced in proportion to actual costs.

So I think if institutional demand for staking on Ethereum continues, supply will emerge to meet it, and supply and demand will find their meeting point — likely not through on-chain staking yield, but through node services or white-label arrangements via contracts. Rather than saying what's right, if I had to predict the future, that's how I'd… Thank you for all three perspectives.

I'm not sure if it's now or at another time, but the person who made this proposal is apparently speaking about it in an English session. On this very topic. I'd love to hear that, too. He's a researcher rather than a validator operator, so hearing from people who actually earn revenue from this offers a lot of insight. I actually feel the same way. As was just said, demand will probably find the right balance in the end. But the validator business requires upfront spending on infrastructure, and many say constantly changing returns

make that hard, so I hope the Foundation reaches consensus with the community. One request from me: there are many validator companies in Korea, too. It would be great if they spoke up on issues like this. The third question is about running validators in Korea. There have actually been quite a few Korean builders in the global market. Most of those builders are struggling.

Some collapsed, like Terra-Luna. By comparison, validator infrastructure companies have survived longer than expected, and are still going strong. So what are the strengths of doing business in Korea that let infrastructure firms survive? And yet, what's the biggest constraint keeping them from reaching the global top, like Figment or P2P? From the start… we did run validators, but from the start we were a company that researched blockchain core, and that's how we grew,

so in that sense we do fit what you described. Personally, I don't think there are many validator companies in Korea that are surviving and growing well. But in terms of scale, why Korea has no company as big as Figment or P2P, or the big Western firms, ties back to the topic we discussed at the start. And it's VC connections, in the end… we met a lot of VCs. We went to the US and to conferences and tried hard to persuade them.

Asking why they delegated to so-and-so. Why couldn't they move it to us? What's the reason? When we asked why they chose them, the person said the previous manager decided, so they didn't know. Or, "we were just close." Reasons like that come up a lot. The reasons behind these choices are so unclear that it's really hard to convince them to switch. The person in charge is just following a decision that's already been made, and if they change, they have to justify why, which is hard — things are already working fine, so why change?

We tried that many times and eventually realized it just isn't easy. That's why validators that built institutional ties early on — since most VCs were in the West, validators with good connections there, like Figment, Chorus One and others — were able to succeed. As for why companies that started as validators in Korea are surviving, probably because they pivoted well?

That's honestly what I think. If they'd stuck to validators alone, it would have been tough. If you split things into B2C and B2B, in crypto, B2C has high peaks but also deep troughs. is B2B, so it doesn't soar in good times but it doesn't crash in bad times either, so relatively speaking it feels more sustainable long-term.

I strongly relate to the overseas examples, too. Plus, infrastructure firms like Fireblocks, BitGo, Anchorage, Ledger, — most of them came out of the US or Europe, and there were validators alongside them when they emerged, and they built the market from then on. We also tried to work with overseas custodians. At first we worked together, but later once those companies grew and then it's "why would this company be working with a validator…"

If companies like Ledger had emerged from Korea and expanded globally, validators would have gone along with them, but favored the US and Europe, I think. And even big validators that didn't partner with institutions or custodians have also shut down — that happens too. So it ties back to the institutionalization we discussed at the start, and to where the custodians were. It's all connected.

As for what the advantages are of running validators in Korea — honestly, I'm not sure how much you'll see it as an advantage. In reality the time zones are always different. At 2 a.m. we'd gather on Discord to run operations, and the Asians ask "when's it done?" It's their working hours, so they're fine, but we need to sleep.

I remember some really painful times because of that. Still, if I think about the advantages, in the early days of PoS, there was demand for geographic decentralization. Since most validators were concentrated in the US and Europe, more thoughtful players wanted some operations in APAC, and there weren't as many validators running in APAC as you'd think, so I think we were able to benefit from that.

As for the second question — why there's no like P2P or Figment — I think it ultimately comes down to liquidity. In Korea, liquidity just isn't as deep as you'd think. And as mentioned, overseas custodians and large VCs are concentrated in the US and Europe. When they ask, "Why should I work with a company in Asia — different language, different time zone, already hard to deal with?"

realistically, we found it hard to give them a good answer. In the end, the infrastructure for moving large institutional volumes needs to be built more in Asia — that's the issue, I think. Thank you for those perspectives. Let me try to sum it up. In the past, was mostly foundations lots of new chains were launching, so with the push for geographic diversity, there was demand for Korea, and firms survived. But now there aren't many new chains, customers are mostly custodians and institutions, and liquidity sits in the West rather than Korea, so Korean validators

are finding it harder. That's why it's good that all three of you are also running other businesses beyond validators — and you're still going strong. The next question — and views on this really differ from company to company — is about the role of validators. When it comes to running blockchain nodes and a validator business, I wondered where the validator's role really ends.

Is it enough to run nodes reliably and keep the chain running, or does the future role extend to contributing to the community and the technology? I'd like to hear each of your views. I'm a developer by background, so naturally I think we should make those kinds of contributions. In the past, validators were seen as the teams that protect the network.

We used to — though we can't do as much now — patch things at the code level, help fix bugs, and in Ethereum's case, when there's a chance, we actively work on tooling and such, directly or indirectly, and do research too. That kind of thing. Contributing to the network that way helped grow the ecosystem, and that's what's needed from those guarding the base layer. The business side is a good thing, of course.

But ideals and reality seem to diverge. I think it depends on who's delegating. If it's a foundation, it wants the ecosystem to thrive, so contributing to that is part of the role. If it's an institution, as I said, custody, licenses, and track record matter. There were some unusual validators too — some focused on retail running YouTube channels to teach people how to stake.

Especially around. So everything needed to win delegations is what a validator does, and those activities vary depending on what type of client is delegating. While running our validator business, we struggled a lot with what validators should do — what their duties and role really are. At first, we thought of a validator as a tech company, so we focused on technical contributions. But it turned out that validators

with more delegation and those with less didn't differ in proportion to technical skill or contribution at all. We came to realize that our customers weren't the network or users, but the delegators. We'd done a lot of that in theory, and as a company that researched and invested heavily in it, we wanted to keep investing in that direction. But since that wasn't the case, we ended up doing networking, supporting marketing — that sort of activity.

Doing that, as a tech company, we kind of hit a reality check, you could say. We wondered what we were even doing. And in a way, one reason we scaled back the validator business was realizing we should focus on what we do best. As the market shifts toward institutions, I see it positively — it's becoming an environment where companies like ours can focus on what we do well and be valued for it. That's how I see it.

The basis of delegation and the revenue it generates kept drifting away from technology on public chains, which was hard for us to adapt to. But institutions will value network reliability and security more than retail does, and that aligns better with our direction. So we view this change rather positively.

Thank you. Summing up what I've heard: from the validator's perspective, the engineers and businesses running validators hope the validator's role will be more technical and contribution-focused, but listening to you, it seems that's not what the market wanted from us. Going forward, spaces where these capabilities can shine may grow on blockchains, so I think what you've been doing may finally pay off.

This will probably be the last question. It came up in the first question: as profitability keeps declining and business gets harder, efficiency will matter even more. The first lever is AI. Thinking back, at my previous company just about 1 and a half years ago, as you said, people had to wake up at dawn and stay up — the work was powered by grinding people down.

Now AI does the first pass, analyzing the root cause and telling you whether a human really needs to step in, and part of operations is actually handed over to AI. Using AI heavily like this is cutting costs. So how much do you actually use AI in infrastructure operations or core development, and in what ways? That's my final question. Compared to the others, our validator operation was… at its largest we ran 30-40, but even before AI,

about a thousand people [?] handled it. So we didn't experience it as a cost-heavy business. With over 100 networks you might need about 10 people, but what actually costs us far more than engineering skill is ecosystem costs. You have to attend conferences, host events, help with marketing, help with GTM — those things actually took more resources. Where we've recently been using AI a lot… even while running validators, if something looks off, we go hunting for bugs.

It's just instinct for us… even without knowing why, if it's odd, we dig in. So we use AI heavily in bug hunting, and among AI auditors, there are about 3 firms, like Cantina, that we work closely with, giving model feedback almost at a co-research level. The better the model, the better our audit results, so we collaborate in that spirit. With AI, our sense is that it improves security more than productivity.

Attackers are improving too, so it feels like we're just stepping up to keep pace, but we strongly feel we'd be in trouble without it. So auditing, bug hunting, — that's where we use AI the most. 3 years ago, at a large one co-founder said they were developing with AI, and I thought it was kind of pointless.

That's what I thought back then, but it worked so well that I wondered whether. At the time it was a research topic, but now it's something you can just do with a bit of. So we use it too. We're not going so far as to aggressively change everything, but in management it has clearly improved efficiency a lot.

So in DevOps, too, they use AI to solve problems and build platforms. Right. Right. Setting aside 3 years ago, compared with 1 year ago, efficiency has improved a lot. Our dev team is also seeing big gains from AI.

Speaking as a validator company, since we run RPC, we have about 700 instances. One person would have to watch all of them, and with so many time-zone-related incidents and various issues, it's physically and mentally exhausting work. we need to automate more so we're applying AI heavily with that goal.

First, everything with IaC code and when an upgrade notice comes out, it's automated into a ticket, and once the ticket is created, the code gets written, and then we review the code. up to 100%, and we're in the process of rolling it out. Demand clearly varies a lot depending on infrastructure scale, but past a certain point, you really feel you couldn't survive without it.

At Four Pillars, we don't run that many validators, and we don't really have an engineering dev team — just a few DevOps people. For that kind of setup, it's great. For example, if a client asks for a dashboard on their staking, in the past developers would have built it over a month or so. Now even DevOps engineers can turn it around in a day or two with vibe coding. In terms of tooling, it has helped a great deal.

As tech companies, you're clearly thinking hard about how to use not just blockchain but AI going forward. Thank you for your insights today. These are probably the people who've thought hardest about the validator business in Korea's entire blockchain industry, so if anyone here has questions about the validator business or the infrastructure business, the CEOs are here now, so why not ask?

Any questions? Yes. Actually, back in the days and I followed you and read your tech blogs, and I really enjoyed them. And there's the institutionalization of the ecosystem, and the security side — validators being responsible for security. Recently, with falling token prices and the declining you've been talking about I get the sense things are shrinking.

It almost feels like the decline of PoS. We moved from PoW to PoS, and countless chains ultimately relied on validators as the ones responsible for security and supported their operations. If validators really disappear for these reasons, critical issues will increasingly surface not just in Ethereum but across PoS chains. They say infrastructure shines brightest when nothing breaks. But if that role shrinks and the gaps start to show, I think real incidents could happen. How should blockchains evolve from here? If validators' starts to disappear

I'd like to hear your thoughts on that… Please go ahead. Revenue keeps falling, but do you think it will hold at some level? Many validators, ourselves included, are using this as a tool to generate additional revenue. We're doing custody and such. The reason I think the validator business won't disappear is that validators keep PoS chains running permanently, and there's Ethereum's so incentive adjustments will happen.

I think they'll settle at a reasonable level, and where is difficult validators will take on additional meaning and pursue various approaches, I believe. Our work on collateral and similar things should be seen in that light. Thinking about my answer, networks that can't provide enough PoS revenue could disappear, I think. Many are already disappearing, and some are moving onto Ethereum as.

But for Ethereum and a few other networks that can sustain themselves, I think security could actually get stronger. Because delegators and customers are shifting from retail to institutions. Before, to run a validator you'd just use the cloud or buy a machine, but now to win institutions, you need to obtain SOC, ISMS and so on. In practice, you'll have to pay much more attention to.

They'll demand that much, and where that demand exists, I don't think the problem will be that critical. Let me explain how our company's focus is changing. We see blockchain use cases not disappearing as the public chain market shrinks, but moving toward institutions. On public chains, the topic was the security of public infrastructure, while in institutional blockchains, internal infrastructure security becomes the focus.

The security themes we've researched carry over. With institutions, the context changes — it's not security seen from outside, but security seen from inside: procedures, and. So lately we've focused on similar technology but with a different approach: how to safely manage large amounts of client assets inside institutions. That's where we keep contributing. In a way, even if public blockchains shrink, the field of researching and securing them isn't shrinking.

That's how I'd explain it — we're continuing along the same line. That's how I'd put it. Thank you. Are there any other questions? If there are no more questions, thank you all for being here today. I hope you'll continue to look after Korea's blockchain ecosystem, and with that, I'll close. Thank you for joining us.