Ivan:

Hello everyone, welcome to the Behind the Screens podcast, where we analyze the realities behind building the new generation of blockchain infrastructure.

Today, in about 15 minutes, we will discuss a hot question: how expensive is decentralization really? Everyone loves decentralization in crypto — it's almost a given. But this episode will “unpack” its downsides, along with Anit – the CTO of Altius.

If you've ever thought that more nodes = more security, this episode might make you reconsider.

Anit, my first question is: everyone claims to be decentralized, so in practice, what does it take to achieve that?

Anit:

Hello everyone, I am Anit. Thank you Ivan. That's a great question.

I think large chains like Bitcoin or Ethereum are truly decentralized — one node failing does not affect the entire system. But as you move away from these large chains, everything starts to… fade.

Oh, we should add Solana too — but that's a special case, because to run a Solana node, you need extremely powerful hardware. They choose to sacrifice decentralization to achieve performance.

That is: the further you move away from the original ethos of Bitcoin and Ethereum, the closer you get to centralization.

Even with Bitcoin — using Proof-of-Work mechanism — only those with enough powerful mining rigs can mine it. So decentralization is somewhat limited by hardware resources.

As for Ethereum — it's truly decentralized, but now we see many L2s and L3s emerging. And most L2s currently are absolutely centralized, as there's only one sequencer.

If that sequencer fails, the entire system stops functioning. Although you can run a node to read state data, control lies at a single point.

Currently, Ethereum is still maintained by a diverse network of nodes — each person runs their own software. If someone has a bug, it doesn't affect the entire network.

But the reality is about 50–60% of Ethereum nodes run Geth. If Geth has a bug, only the remaining 40% can “carry the team.” Luckily, Ethereum has diverse clients like Nethermind, Besu...

However, that raises the question: is the system truly decentralized, or is it just an illusion?

Ivan:

Wow, I never thought about that — stability depends a lot on the community.

So why can Web2 systems like Google or AWS ensure almost absolute uptime (99.999%), while blockchain struggles to reach that level?

Anit:

That's due to differences in the model.

Google is a centralized system. They have thousands of engineers, hundreds of people specifically working on QA and testing for Gmail. They test every scenario, have standby hardware, specialized infrastructure...

If Gmail has an issue, users don't even notice — you might just feel that the email takes 1 second to send instead of 300ms. Meanwhile, the backend may have automatically recovered from the crash.

In contrast, with L2 in Web3 — resources are limited. If the system crashes, you might have to log into AWS, rebuild the server manually, restore an old version... Too many steps, too few people.

Additionally, many other L1s besides Ethereum suffer from “software homogenization” — usually only one client operated by the foundation.

With Ethereum, anyone can run a node, so it's more decentralized.

Ivan:

It's true that I often underestimate the current level of stability. Looking back, Web3 has come a long way. In the past, errors were much more frequent.

And indeed, Gmail and Facebook also had downtime, we just got used to it.

Final question: if we choose complete decentralization, what are we sacrificing in terms of performance? And what conditions are needed to achieve true decentralization?

Anit:

Tough question. I think there is no silver bullet — unless we have quantum computers.

We can optimize software, upgrade infrastructure, shorten block times... but there are always physical limits.

For example: when trying to achieve global consensus, you always hit the speed of light barrier — data can't be transmitted faster than that.

Maybe we will find a "balance point" — maintaining the spirit of decentralization while achieving practical usability.

It could be through redesigning data structures, changing consensus methods, requiring stronger hardware...

But then the question arises: who is qualified to run a node? What about those who only have smartphones?

Ivan:

Absolutely — 60% of the world only has phones, no laptops. If the hardware requirements are too high, we're excluding the very users.

Anit:

And that's the crux. We can't “maximize decentralization” at any cost.

We need to design systems that know when to decentralize and when to optimize. Because for Web3 to be widely adopted, it has to operate stably, not just look good on paper.

Ivan:

Awesome. If this podcast episode makes you rethink decentralization, then we've succeeded.

Thank you Anit for joining! Thank you all for listening 🎧