Remember when everyone said Ethereum would never scale without losing its soul?
I do. For years we heard the same tired joke — “Ethereum is a decentralized world computer… that processes 15 transactions per second.” It stung because it was kinda true. Then rollups came, blobs came, and still people asked: when is the real sharding we were promised back in 2017?
Well, stop asking. It just happened.
Late Wednesday night, the Fusaka upgrade went live on mainnet, and Vitalik Buterin did something he rarely does — he celebrated out loud.
The Upgrade Everyone Waited a Decade For
Fusaka isn’t sexy like the Merge. It didn’t have a countdown timer or a panda on stage. But in the quiet way that actually moves blockchains forward, it might be the most important Ethereum upgrade since we ditched proof-of-work.
At its heart sits one EIP that changes everything: EIP-7594, better known as PeerDAS — Peer-to-peer Data Availability Sampling.
Let that name sink in for a second. Data. Availability. Sampling. Those three words just made sharding real for the first time in Ethereum history.
What the Heck Is PeerDAS, Anyway?
Imagine you’re at a huge potluck dinner. In the old days, every single guest had to personally taste every single dish to make sure nothing was poisoned. That’s Ethereum pre-Fusaka — every node downloads every blob, every transaction, every piece of rollup data.
With PeerDAS, the network randomly assigns tiny bites to different guests. If enough people taste their assigned piece and say “yep, looks fine,” the whole table agrees the dish is safe. No one person has to eat the entire lasagna.
That’s it. That’s the trick that took eight years to make trustless.
“PeerDAS in Fusaka is significant because it literally is sharding. Ethereum is coming to consensus on blocks without requiring any single node to see more than a tiny fraction of the data.”
– Vitalik Buterin, December 3 2025
When the co-founder of the chain says “we have it” after a decade of promising sharding, you sit up and listen.
From Dream (2015) to Mainnet (2025)
Let’s take a quick trip down memory lane, because context matters.
- 2015 – Vitalik sketches sharding on a whiteboard in a Miami hotel room
- 2018 – Phase 1/2 plans, beacon chain, stateless clients
- 2021 – “We’ll have data sharding by 2023” (narrator: we didn’t)
- 2024 – Dencun brings blobs, the appetizer
- 2025 – Fusaka serves the main course
Every delay hurt. Every “soon” meme aged poorly. But the crazy part? The final design is actually better than anything we drew on those 2017 whitepapers.
PeerDAS doesn’t need new validator sets. It doesn’t split history. It doesn’t even touch execution. It just lets nodes sample data availability like they’ve been sampling beacon blocks since 2022. Same math, bigger impact.
The Numbers Everyone Wants to Hear
So what does this actually unlock?
Right now Ethereum can handle roughly 1–1.5 million transactions per day. With full PeerDAS and the future increase to 16 blobs per block (already coded, just needs a flag day), analysts expect:
- Rollup throughput jumping from ~100 TPS today to north of 2,000 TPS by mid-2026
- Blobspace market becoming the new gas market — cheaper, more predictable fees for L2 users
- Average L2 transaction dropping below $0.005 permanently
And the best part? None of this requires trusting centralized sequencers forever. The data is on Ethereum mainnet, sampled by thousands of nodes. If a rollup lies, the network knows within seconds.
Why This Upgrade Feels Different
I’ve covered every Ethereum upgrade since Berlin. Usually the price pumps, dumps, and we argue about something else two weeks later.
Fusaka feels quieter, almost anticlimactic — and that’s exactly why I think it’s huge.
There was no “Shanghai withdrawal narrative.” No ultrasound money memes. Just a bunch of researchers shipping code they’ve been testing for four years. The fact that almost nobody outside core dev circles fully grasped what shipped yesterday tells you how deep this rabbit hole goes.
In my experience, the upgrades that change everything are exactly the ones that feel boring on launch day.
What Happens to ETH Price?
Short term? We saw the usual post-upgrade pump. ETH broke $3,200 within hours, currently sitting around $3,190 as I write this.
Long term is where it gets interesting.
More transactions → more blob fees burned → higher economic security → higher staking yield → more ETH locked → lower liquid supply.
We’re entering the part of the cycle where Ethereum becomes aggressively deflationary again, but this time driven by L2 activity instead of base layer spam.
The Road to One Million TPS
Fusaka isn’t the end. It’s the foundation.
With PeerDAS live:
- 2026 will likely bring 16–32 blob slots
- 2027 could see full danksharding with 64–128 MB blocks
- By 2028 the roadmap points to distributed block building and PBS
Add based rollups, parallel execution, and EVM improvements still cooking, and the “one million TPS” dream that sounded insane in 2021 suddenly looks… achievable?
Maybe not next year. But for the first time, I don’t laugh when someone says it.
Final Thoughts
Sometimes progress in crypto looks like a new dog coin pumping 100x. Sometimes it looks like a GitHub commit at 3 a.m. that nobody screenshots.
Fusaka is the second kind.
Ethereum just proved — again — that it can ship the hardest things in computer science while the rest of the industry argues about memes.
And yeah, the price will do whatever it does tomorrow. But the network just got a lot stronger, a lot cheaper, and a lot closer to the vision we’ve all been holding since 2017.
Sharding isn’t coming anymore.
It’s here.
