Decisions

Nine Ions on a Chip Raised 25 Million Dollars

An ETH Zurich spin-out raised six times its lifetime funding on a demonstration of nine trapped ions. The qubit count is the least useful number in the announcement, and knowing why changes how you read every vendor roadmap.

DecisionsBy Servola Tech Desk2026-07-286 min read

AI-assisted, edited by humans. | Editor-in-Chief: Leon Soliman | Editorial standards

Nine Ions on a Chip Raised 25 Million Dollars

Key takeaways

ZuriQ, an ETH Zurich spin-out, raised 25.5 million dollars on 28 July led by Quantonation, which is roughly six times the 4.2 million it had raised in total before, on the strength of a machine holding nine ions.
Nine is not the selling point and reading it as one inverts the story: the claim is about the rate at which qubits can be added, not the number reached, so compare architectures on slope and never on today's level.
The physical change is that ZuriQ confines ions with static magnetic fields in Penning micro-traps rather than the oscillating radio-frequency fields of conventional Paul traps, which is what permits a two-dimensional grid instead of a line.
The underlying physics is European public research: the 2024 Nature paper behind the approach came out of ETH Zurich with Leibniz University Hannover and Germany's national metrology institute, and the company was built on top of it.
Do not move your post-quantum migration date because of a funding round. Set that date from how long your data must stay confidential and from the rules you answer to, then review it only when a machine runs a real workload.

Nine ions, and a round six times the size of everything before it

Pavel Hrmo runs a company whose most advanced machine holds nine ions. On 28 July, ZuriQ, the ETH Zurich spin-out he leads as chief executive, announced a seed round of 25.5 million dollars led by the Paris quantum specialist Quantonation, with Forward.one, Extantia, Firgun Ventures and every previous backer taking part. The company had raised 4.2 million dollars in total before this. The new round is about six times its entire funding history.

What the money was raised against is a three-by-three array of nine individually controlled calcium ions on a single microfabricated chip, which ZuriQ describes as the first natively two-dimensional trapped-ion array and the largest of its kind demonstrated so far. The team has gone from four people to eighteen. The stated goal for the new capital is hiring, research and chip fabrication, with the aim of putting hundreds of qubits on one chip.

Set that against the field and the number looks absurd. Commercial trapped-ion and superconducting systems are quoted in the dozens, hundreds and thousands. Nine would not register on any leaderboard, and no serious investor mistook it for one.

Why it matters: a room full of professional investors looked at nine and committed 25.5 million dollars. Either they are wrong, or the qubit count is not the thing being priced. The second reading is the useful one, and it is a lesson that transfers well beyond quantum computing.

The number in the headline is the wrong number

ZuriQ is not selling a level, it is selling a slope. Hrmo's own framing is that the architecture is two-dimensional from the ground up, so the number of qubits that can be placed on a chip should grow far more readily than in systems built on an older blueprint. That is a claim about the derivative, not the value: how much it costs to add the next qubit, rather than how many exist today.

This is why a nine-ion demonstration can be worth more than someone else's hundred. A hundred qubits on an architecture whose cost per additional qubit climbs steeply is a machine near its ceiling. Nine on an architecture that adds them cheaply is a machine near its floor. The two numbers are not comparable, and putting them in the same sentence, as most coverage of quantum computing does, produces a ranking that means nothing.

The bottom line: whenever a technical field reports a single scalar, ask whether the seller is claiming a position or a rate of change. Almost every credible early-stage hardware bet is a rate-of-change claim wearing a position's clothing, and almost every misleading one is the reverse.

Static fields instead of oscillating ones

The reason the slope might differ is a genuine change in physics rather than in engineering polish. Conventional ion traps, called Paul traps, hold ions using rapidly oscillating radio-frequency fields. That works well, and it is how most trapped-ion machines operate, but it pushes designers toward lining ions up in one-dimensional chains connected by junctions. Growing such a system means adding junctions, and junctions are where the difficulty concentrates.

ZuriQ uses Penning micro-traps instead: a large uniform static magnetic field combined with voltages on a microfabricated electrode array. Because nothing is oscillating, an ion can be held and moved anywhere above the chip surface, which is what makes a genuine two-dimensional grid possible rather than a line with corners in it.

Worth noting: this is not a company that invented a story to raise money. The approach was published in Nature in March 2024 by a group at ETH Zurich working with Leibniz University Hannover and the Physikalisch-Technische Bundesanstalt, Germany's national metrology institute, and Hrmo is among the authors. The commercial entity was built on top of already-refereed physics, which is a materially different risk profile from a claim that exists only in a pitch deck.

The base rate this bet is running against

Architecture bets in computing hardware fail more often than they succeed, and they usually fail late. The pattern is familiar: a new approach shows a clean small-scale result, the scaling argument is sound on paper, and then a problem that was negligible at nine becomes dominant at nine hundred. Nothing about ZuriQ's round tells you it has escaped that pattern. A seed round is evidence that specialists found the argument credible, not evidence that the argument is correct.

What would actually update the picture is narrow and worth writing down in advance. Does the array grow by an order of magnitude while keeping individual control of every ion. Do the error rates at that larger size stay in the same range as at nine. Does the multi-layer fabrication needed for bigger chips work outside a research setting. Until those land, the honest description is a promising architecture with one demonstration behind it.

Yes, but: the fact that a bet is unproven does not make watching it pointless. It makes the watching cheap. You are not being asked to buy anything, and the cost of tracking three specific milestones is a few minutes a year.

Christophe Jurczak of Quantonation put the investment case plainly: a common misconception is that the quantum race is already decided, and it is not. That is a fair statement of why money is still moving into new architectures. It is not a statement that any of them work at scale yet, and it should not be read as one.

What this does not change on your side

Nothing in this round should move a single date in your security planning. The temptation after a quantum funding headline is to treat it as a countdown signal, and vendors of migration services will present it that way. It is not one. A nine-ion research chip is many orders of magnitude away from anything that threatens the cryptography protecting your systems today.

The dates that should drive post-quantum planning come from two places, and neither is a press release. The first is how long your own data must stay confidential, because encrypted traffic captured now can be stored and attacked later, which means information with a fifteen-year secrecy requirement is already exposed to a machine that does not exist yet. The second is whatever your regulator or your largest customer puts in writing. Both are knowable today and neither depends on which architecture wins.

One instruction: take your three most sensitive data flows, write down how many years each must remain secret, and let that number set your migration order. Then revisit the field only when a machine runs a real workload, not when one raises money.

Frequently asked questions

Is nine qubits a meaningful result?

It is meaningful as a demonstration of a two-dimensional arrangement with individual control of each ion, which is what ZuriQ set out to show. It is not meaningful as a measure of computational power, and the company does not present it that way. Comparing it to the qubit counts of larger commercial systems is comparing two different kinds of claim.

What is the difference between a Penning trap and a Paul trap?

A Paul trap holds ions with rapidly oscillating radio-frequency fields and is the conventional approach. A Penning trap uses a strong static magnetic field together with static electric fields. Because nothing oscillates, ions can be positioned and moved freely across a chip surface, which is what allows a two-dimensional grid rather than a chain of ions joined by junctions.

Should this change our post-quantum cryptography timeline?

No. A funding round is not a capability milestone, and a nine-ion research device is far from anything that threatens current encryption. Set your timeline from how long your data must remain confidential and from your regulatory obligations, both of which you can determine now, and revise it on demonstrated capability rather than on investment news.

The instructive part of this round is not the physics, it is the reading. Nine ions attracted six times a company's lifetime funding because the people writing the cheques were looking at the cost of the tenth ion rather than at the existence of the ninth. Most technical claims that reach an owner have been flattened into a single number somewhere along the way, and the number that survives is usually the one that sells rather than the one that decides. Ask what is being claimed, a position or a rate, and most of the noise sorts itself out.

ZuriQquantum computingQuantonationETH Zurichtrapped ionspost-quantum cryptographydeep tech

The Quantum Deadline Hiding in Your Certificates

Cloudflare will move its signatures to ML-DSA now rather than wait for NIST's nine new candidates. For owners, the trust layer is the slow half of post-quantum.

2 min read

Decisions

Cursor's Cheapest Seat Runs a Different Model

Cursor launched a 649 rupee India tier on 28 July. It keeps the two models Cursor does not buy and drops the two it pays for. That is a bill of materials, not a discount.

5 min read

Law & Regulation

The EA Buyout Now Hinges on Brussels

The largest take-private in history has cleared EA shareholders and US antitrust. The last real gate is the EU - and not classic merger review, but the Foreign Subsidies Regulation, facing its first blockbuster test.

3 min read

← All articles