Hexagon AEON Robot

How much does a Hexagon AEON cost?

There is no published price. Not a starting price, not a range, not a lease rate. Hexagon deploys AEON exclusively through enterprise partnerships, has no online store and no research or education channel, and has disclosed no contract value for any deployment - including the Schaeffler commitment, which is the largest confirmed humanoid fleet deal in European manufacturing and whose financial terms neither party released.

Our estimate of $100,000 to $250,000 per unit is exactly that: an estimate, and a weak one. It is the published band for comparable industrial humanoids of similar specification - roughly 1.6 m, 60 kg, 34 degrees of freedom, five-fingered hands - used as a proxy. We give it low confidence, and we would rather you knew that than saw a confident-looking number.

Hexagon reports its financials in Swedish krona and has disclosed no AEON figure in any currency, so no conversion has been performed on anything above.

Why does it matter that we say this plainly? Because 'Inquiry' or 'contact sales' is the weakest possible answer on a page people arrive at wanting a price, and the temptation to invent one is enormous. Several sites have. A comparable-band estimate with its reasoning attached is the strongest defensible substitute, and the reasoning is the part that matters.

Why there is no price, and probably won't be one

It would be easy to read Hexagon's silence as a startup hiding an uncomfortable number. It is more likely to be ordinary industrial practice.

AEON is not really sold as a unit. It is sold as a capability inside a plant: a defined inspection or machine-tending task, integrated with the customer's existing systems, commissioned by engineers, supported under a contract. The price of that is the price of a project, and it varies by site, by task count, by integration depth and by how many units the customer eventually takes. Hexagon has been selling coordinate measuring machines and metrology systems into these same plants for decades on exactly this basis. Nobody publishes a price for a CMM installation either.

Which means two practical things for a reader. First, if you are a manufacturer, the number you should be asking for is cost per inspected part or cost per shift covered, not the unit price - the unit price is the least informative figure in the deal. Second, if you are not a manufacturer, there is no route in. There is no research edition, no developer kit and no path for a university to buy one, which is unusual for a humanoid this prominent and explains the near-total absence of community around it.

The BMW Leipzig deployment, in order

This is the best-documented humanoid deployment in Europe and worth setting out as a timeline, because 'deployed at BMW' is doing a lot of work in a lot of headlines.

It started with a theoretical evaluation phase, then laboratory testing. A first test deployment ran at BMW Group Plant Leipzig in December 2025. In March 2026 BMW showed the programme publicly at its Munich campus. Then on 12 June 2026, Hexagon and BMW announced that AEON had begun performing production tasks at Plant Leipzig while simultaneously being trained on future manufacturing applications - reported as the first humanoid robot deployed in production in Germany. Two AEON units work across two use cases, with BMW expecting both to be in production by the end of the year.

What makes this credible is the shape of it. Eighteen months from evaluation to production tasks, with a test phase in between, is what an actual industrial integration looks like. It is not a launch video. The work is inspection and measurement rather than heavy handling, which matches both the 15 kg payload and Hexagon's core competence.

What it is not: proof of scale. Two units on two use cases in one plant is a pilot that graduated, not a fleet.

Schaeffler's 1,000 units - what is and isn't confirmed

In April 2026 Schaeffler, the German bearings and precision components manufacturer, committed to deploying at least 1,000 AEON units across its global factory network by 2032. On the numbers available, that is the largest confirmed humanoid fleet commitment in European manufacturing.

Here is what is confirmed: the commitment itself, the unit count, the 2032 horizon, and that the rollout is multi-site. Reporting has put the start of that rollout at the end of 2026.

Here is what is not: the money. Neither Hexagon nor Schaeffler disclosed the financial terms, which is why we cannot back into a unit price from a contract value the way you sometimes can with public procurement. Nor is a commitment a delivery. Six years is a long horizon, the number is a target rather than a purchase order, and industrial automation commitments of this shape get revised - upward and downward - as pilots produce data.

Still, treat it seriously. Schaeffler is not a venture-funded believer in humanoids; it is a components manufacturer with a large plant estate and a long history of automating it. A firm like that committing to a fleet is a different signal from a technology company buying two robots for a press release.

Pilatus, and the other named partners

Beyond BMW and Schaeffler, Hexagon has named Pilatus, the Swiss aircraft manufacturer, as a partner. Aerospace is an obvious fit for a robot built around measurement: aircraft assembly is a world of dimensional tolerance, documented inspection, and parts too large or too awkward for a fixed measuring machine. A mobile robot that carries a metrology-grade sensor package to the part instead of the reverse is a genuinely interesting proposition there.

Trade press has also reported a 2026 factory deployment with an Austrian machine builder, covering inspection and machine tending - the two tasks that recur across every AEON deployment described so far.

What is missing from all of this is independent performance data. Every figure about how well AEON does its job comes from Hexagon or from a partner announcing a partnership. There is no third-party benchmark, no independent hands-on, and no owner community publishing results, because the owners are manufacturers with confidentiality agreements. That is normal for industrial equipment and it is still a real limit on what anyone outside those plants can honestly say.

Physical AI and the 20-demonstration claim

Hexagon describes AEON's autonomy as four layers of physical AI, with imitation learning as the mechanism for teaching new tasks, and states that roughly 20 demonstrations are enough to train an autonomous operation.

Twenty is a striking number, and it deserves an asterisk rather than a dismissal. In constrained industrial settings - a fixed station, consistent lighting, the same part arriving in the same orientation - small-sample imitation learning genuinely works, and the constraint is doing most of the work. The same claim about an unstructured environment would be extraordinary. Hexagon is careful to make it about factory operations, and the deployments described so far are exactly the sort of repeatable, well-lit, fixtured tasks where it is plausible.

What has not been published is the definition. Twenty demonstrations to what success rate, on what task complexity, with how much engineering setup around it? Those numbers exist inside Hexagon and inside BMW. They are not public, and until they are, the claim sits in the category of credible-but-unverified.

The honest summary: AEON's autonomy story is more believable than most, because its task domain is narrower than most. Narrowness is the feature.

What AEON can do, and what it cannot

It can move quickly across a factory floor - 2.4 m/s is faster than most humanoids manage - on wheeled legs that suit flat industrial surfaces. It can carry and handle up to 15 kg. It can perform dimensional inspection and measurement using a sensor package denser than anything else in this class. It can swap its own batteries, which removes the most common reason a mobile robot stops being useful. It can be taught new operations by demonstration, within limits. And it has been integrated into a live automotive production environment, which almost nothing else in this catalogue can claim.

It cannot be bought by an individual, a university, a developer or a small business - there is no channel. It cannot climb stairs or handle rubble in the way a true biped can, by design. It cannot be independently benchmarked, because no independent party has had one. It cannot be priced from public information. And it cannot do the general-purpose household or logistics work that dominates humanoid marketing, because that was never the brief.

The list of things AEON cannot do is longer than for a general-purpose humanoid. That is because Hexagon made choices, and choices exclude things.

Does the July 2026 US rule affect a Swedish robot?

This is the point where most coverage gets the law wrong, so it is worth being precise.

On 28 July 2026 the FCC issued Public Notice DA 26-786, adding equipment categories to the Covered List. The press reported it as a ban on Chinese humanoid robots. The text is not about China. It turns on 'foreign-produced' - failing a Buy American test of US final assembly plus more than 65% domestic content. Law firms reading the notice have been consistent that British, EU and Japanese equipment is caught by the same wording. It blocks new FCC certification only; it is not a customs ban and not a recall, and equipment certified before 28 July can keep being sold and updated until 1 January 2029.

So: yes, on its wording the rule reaches AEON. A European-manufactured mobile robot meets the foreign-produced definition, and the listing also covers autonomous mobile robots as a category. Hexagon has published no authorisation position and no disruption has been reported, which is unsurprising given that AEON runs through industrial pilots rather than open sale and its announced deployments are in Europe.

For a US manufacturer considering AEON, this is a procurement question rather than an academic one. Ask Hexagon directly about US certification status before scoping a deployment. European industrial pilots are unaffected.

AEON versus the humanoids it is actually competing with

AEON is not competing with Tesla Optimus or 1X Neo, whatever the comparison tables say. Its rivals are the industrial humanoids being placed in plants right now: Agility's Digit in logistics, Apptronik's Apollo in materials handling, Figure's machines in automotive pilots, and the wheeled dual-arm platforms from Chinese manufacturers.

Against those, AEON's differentiators are specific. Metrology - nobody else brings a measurement company's sensor stack or its credibility with quality departments. Wheeled legs - faster and more enduring than bipeds on the floors that actually exist in factories. Self-swapping batteries - a genuinely underrated feature that decides whether a robot covers a shift. And an existing customer relationship: Hexagon is already in these plants selling measurement systems, which is a shorter path to a purchase order than any amount of demo footage.

Its weaknesses against that field are equally specific. The scope is narrow - inspection and machine tending, not general handling. Payload at 15 kg is mid-pack. There is almost no independent visibility into performance. And the total absence of a developer or research channel means the ecosystem effects that help Unitree and Booster simply don't apply here.

If your problem is measuring and inspecting things in a factory, AEON is arguably the best-aimed humanoid on the market. If your problem is anything else, it is the wrong machine.

Who AEON is for

It is for large manufacturers - automotive, aerospace, precision components - with existing quality and metrology infrastructure, a plant estate big enough to make a multi-site rollout worth the integration effort, and a defined inspection or machine-tending problem they can measure the cost of today.

It is not for anybody else, and there is no polite way around that. There is no research edition, no developer programme, no reseller, no price and no store. If you are a university lab, a startup, a systems integrator without an enterprise relationship, or an individual, AEON is not available to you at any price, and the closest equivalents you can actually buy are wheeled dual-arm platforms from Chinese makers at a fraction of the estimated cost.

For the manufacturers it is aimed at, the buying process is a conversation with Hexagon about a task, not a product comparison. The relevant questions are cost per inspected part, integration timeline, who owns the training data from your demonstrations, and what happens to support if the programme is restructured. None of those appear on a spec sheet.