Ethereum just ripped out a key piece of its plan for surviving the age of quantum computers — and replaced it with older, more boring technology that happens to have decades of battle-testing behind it. Ethereum Foundation researcher Justin Drake announced on Thursday that the foundation is abandoning the Poseidon hash function in its planned post-quantum architecture and turning to established alternatives such as SHA and BLAKE instead.
By Amir Hassan | August 14, 2026
The Hook: Why a Hash Function Could Decide Ethereum’s Future
A hash function sounds intimidating, but the idea is simple: it converts data into a fixed digital fingerprint, letting computers check that information has not been altered — like a tamper seal on a medicine bottle. Ethereum relies on these fingerprints everywhere, and future versions of the network will rely on them even more.
The concern is that a future quantum computer could break some of the cryptographic tools blockchains use today. That is why Ethereum has been designing post-quantum systems — upgrades meant to keep the network secure even against machines that do not exist yet at meaningful scale. Poseidon, a newer hash function designed specifically for zero-knowledge cryptography, was supposed to be part of that future. Now it is out.
The Evidence: Why the Pivot Happened
According to Drake, the change is not a setback but a consequence of progress. Poseidon’s main advantage was speed inside zero-knowledge proof systems — compact mathematical proofs that let a computer confirm a calculation is correct without redoing the underlying work. Recent advances in these proof systems, known as SNARKs, mean that conventional hashes like SHA and BLAKE can now perform efficiently enough to take Poseidon’s place.
- What changes: Ethereum’s planned post-quantum systems, including a next-generation virtual machine called leanVM, will use established hashes instead of the newer Poseidon design.
- Why it matters: leanVM would help Ethereum efficiently verify large volumes of blockchain activity — a scalability tool as much as a security one.
- The timeline: Drake said a production-ready leanVM is targeted for 2027, followed by deployments across Ethereum’s consensus, data and execution layers in 2028 — though he cautioned the dates remain preliminary.
- The speed bonus: Eigen Labs founder and CEO Sreeram Kannan said joint work with the foundation and zero-knowledge firm Succinct has already increased proving speeds by 2.5 times.
The Core Conflict: New and Fast Versus Old and Proven
Underneath the technical detail is a debate every security engineer knows by heart. New cryptography can be elegant and fast, but it has spent less time being attacked. Old cryptography is slower, but armies of researchers have spent decades trying to break it and failed. Kannan made exactly this point, saying established hash-based systems have fewer known avenues of attack than other post-quantum approaches and could be deployed faster precisely because they have already undergone years of scrutiny.
For a network that secures hundreds of billions of dollars in value, choosing the boring option is not a retreat — it is the whole job. Poseidon was never deployed on Ethereum’s mainnet, so nothing users hold is affected by the change. The pivot happens on the drawing board, before the concrete is poured.
Market Implications: What This Means for Your Portfolio
If you own ETH, this news will not move the price this week — Ethereum trades around 1,870 dollars these days, and hash function debates do not trend on social media. But the decision matters for the long-term case. A serious quantum threat to Ethereum would be catastrophic for holders, so visible, methodical preparation is exactly what you want to see from the organization maintaining the network.
There is also a scaling angle worth noticing. leanVM is not just about quantum safety — it is about verifying huge amounts of activity cheaply, which is the same problem Ethereum must solve to stay competitive with faster rival chains. A faster proving stack, already 2.5 times quicker according to Kannan, serves both goals at once.
The Verdict
Ethereum swapping Poseidon for SHA and BLAKE is the cryptographic equivalent of an airline choosing engines with a longer safety record over a newer model with better fuel economy. Nobody cheers the decision at the gate, but it is the kind of quiet judgment that keeps planes in the air. The 2027 and 2028 targets may slip — Ethereum roadmaps often do — yet the direction is reassuring: the network’s quantum defenses are being built on foundations that have already survived decades of the world’s best attackers.
The cryptocurrency market remains highly volatile. This article is for informational purposes only and does not constitute financial advice.
Disclaimer: This article is for informational purposes only and does not constitute financial advice.
poseidon’s entire pitch was snark performance. then snarks got faster and sha and blake caught up. rip to the shiny new thing
pretty much. once proving got faster the speed premium stopped covering the cryptanalysis risk. 25 years of attack attempts on sha buys a lot of peace of mind
exactly. poseidon only made sense while snark proving was the bottleneck. once that improved, 25 years of sha cryptanalysis wins by default
Kasper V. exactly. once proving cost stopped being the bottleneck the security tradeoff flipped. drake basically admitted the confidence intervals on poseidon never got comfortable
Decades of cryptanalysis on SHA beats a clever new design with a short track record. The boring choice is usually the safe one.
agree, tho its funny we are rebuilding the vm for leanVM and quantum machines that dont exist yet. better early than sorry i guess
better early than sorry is doing a lot of work there. rewriting the vm for lean ethereum is a real cost today, quantum risk is still theoretical. worth it imo but call it what it is
the y2k prep joke writes itself. you patch before the deadline, not after. eth migrations take years so starting while the machines are still lab toys is barely early
the machines dont exist yet argument is how you end up scrambling later. migration lead times on eth are measured in years, starting early is the whole point
eth cant coordinate a routine upgrade in under a year. waiting for actual quantum machines before touching the vm is how you end up scrambling in 2032 with no plan and no runway
shanghai took what, two years from proposal to mainnet. if the quantum timeline compresses you want the hash swap already shipped, not sitting on a roadmap
poseidon had one good paper and a decade of unanswered questions. sha survived every academic swing since the 90s. drake picking boring over clever is the right call for a base layer
people call it boring but sha has eaten a quarter century of attacks and stayed standing. poseidon had like four years and open questions. justin drake picking the boring horse is correct
Worth remembering the quantum math here. Shor breaks signatures outright, Grover only nibbles at hashes. The hash layer was never the weak door, which is why retreating to boring SHA costs so little security.
right. grover is quadratic so you double the output size and shrug. the actual emergency was always ecdsa signatures. poseidon was polishing the smaller half of the problem
the cost framing is the whole argument. you give up some snark speed in a spot where speed barely matters anymore and buy decades of attack surface data. easiest trade on the roadmap
worth noting poseidon isnt dead across the board, snark friendly hashing still has a job inside circuits. its the state hash role that needs the boring decades-tested functions
exactly, poseidon stays inside the circuits where proving cost matters. state root hashing runs millions of times a day, thats where you want the boring battle tested function
^ right split. poseidon inside the circuits, sha for the state root. the walkback only reads as a defeat if you ignore where each function actually runs
swapping poseidon for sha and blake is the most ethereum move ever. years of zk papers and the winner is the hash function that ships with your linux distro
sha and blake winning over poseidon is boring tech defaulting to boring. thirty years of cryptanalysis beats a 2021 paper with nice charts
drake calling poseidon unbalanced after years of zk people hyping it is a rare public walkback. foundation researchers usually just quietly stop mentioning things
Agreed, researchers almost never admit a design bet missed in public. Drake naming SHA and BLAKE directly saves everyone two years of ambiguity in the roadmap calls.
credit where due, killing your own planned upgrade path takes guts. most teams would have shipped poseidon anyway just to save face after all the zk research
drake admitting the confidence never got there is the detail that matters. four years of zk papers and the state hash bet still went to a function from the 90s