EV battery packs are still taken apart by hand. Disassembly — dangerous, high-mix and unautomated — is becoming the bottleneck of the UK's battery circular economy, and perception-driven robotics is what unblocks it.
The UK has spent the last five years building the front and back ends of a battery circular economy. Recyclers can now recover lithium, nickel, cobalt and graphite from end-of-life cells at impressive purity. Cell and pack makers are standing up UK production. Government funding through the Advanced Propulsion Centre has backed battery recycling and reuse repeatedly — hydrometallurgical recovery, graphite regeneration, second-life storage, battery health diagnostics.
But between the retired battery pack and the recycler's shredder sits a step that has barely changed since the first Nissan Leaf came off the road: a person in insulated gloves, with hand tools, taking the pack apart.
- The problem: EV traction packs are still disassembled by hand — high-voltage, high-mix and hazardous.
- Why it matters: recyclers cannot recover materials from packs nobody has taken apart, so disassembly caps the whole circular economy.
- What changed: perception-driven robotics — machine vision, 3D scanning and force-controlled tooling — can now handle the variety that defeated classical automation.
- The bigger prize: the same capability serves battery repair, remanufacturing and second-life triage — markets that exist at scale today.
Why disassembly is still manual
An EV traction battery is one of the most hostile objects in industrial dismantling. It arrives at several hundred volts DC and stays dangerous until verifiably de-energised. It may be damaged — crash-recovered packs carry a real thermal-runaway risk. And unlike almost every other product that gets dismantled at scale, there is no standard design: every manufacturer, model and model-year has its own enclosure, fastener pattern, busbar layout, adhesive strategy and cell format. A dismantler might see a bolted steel enclosure in the morning and a structurally bonded cell-to-pack design in the afternoon.
That variety is why automation hasn't simply been bought off a catalogue. A conventional robot cell is programmed for one known product in one known state. End-of-life packs are the opposite: high-mix, degraded, sometimes deformed, with corroded fasteners and undocumented field repairs. The task demands perception and judgement, which is precisely what classical automation lacks.
So the work stays manual — slow, expensive, dependent on scarce high-voltage-trained technicians, and hazardous enough that insurance and safety cases constrain throughput. As end-of-life volumes climb through the late 2020s and into the 2030s, hand disassembly becomes the choke point of the whole circular economy: recyclers cannot recover materials from packs nobody has taken apart.
What has changed: perception-driven robotics
The technology picture has shifted in the last few years, and the interesting developments are not in the robots themselves but in what surrounds them.
Variant identification. Machine vision, backed by a growing database of pack architectures, can now identify a pack's make, model and revision at intake and retrieve the corresponding disassembly model — which fasteners, in which order, with which tools. Where no model exists, 3D scanning builds one, so the system learns the fleet as it works. And from 18 February 2027, the EU's battery passport becomes mandatory: under Regulation (EU) 2023/1542, every EV and industrial battery above 2 kWh placed on the EU market must carry a QR-code-linked digital record of its construction, materials and history. Comparable data expectations are emerging in the UK. Over time, packs will increasingly arrive with machine-readable identity rather than needing to be recognised from scratch.
Fastener localisation and adaptive unfastening. Modern vision systems can locate bolts, welds and adhesive lines on a specific, possibly deformed individual pack — not a CAD ideal — and feed robot paths that adapt to what is actually there. Force-controlled tooling handles seized and damaged fasteners that would defeat position-controlled automation.
Safety by design. Automation changes the safety case fundamentally. Robots do not suffer arc flash. A properly designed cell keeps humans outside the high-voltage envelope entirely: automated de-energising verification, continuous thermal imaging watching for the early signature of thermal runaway, and interlocked enclosures rated for the failure modes that make manual work so heavily constrained. The safety engineering — functional safety, interlocks, safety-rated control — is as much of the system as the robotics.
Triage, not just teardown. Disassembly is also the moment of decision: is this module fit for second-life storage, or is it feedstock for recycling? Automated measurement at module level — impedance, open-circuit voltage, thermal behaviour, visual defect detection — turns disassembly into grading, and grading is where the economic value concentrates. A module diverted to second-life use is worth a multiple of its shredded material value.
The UK moment
The direction of travel in the UK is clear. Government funding has now backed what is set to be the country's first automated EV battery disassembly plant — one of ten early-stage projects sharing a £9 million award. The Manufacturing Technology Centre has stood up a prototype vehicle battery teardown facility, and Altilium has secured £18.5 million to build the UK's first commercial EV battery recycling facility. European robotics companies are scaling automated disassembly platforms with serious venture backing. The category has moved from research topic to industrialisation race.
But one facility does not serve a national fleet, and the deeper opportunity is broader than end-of-life: the same perception-driven disassembly capability serves battery repair (replacing one failed module instead of scrapping a pack), remanufacturing, and warranty returns — markets that exist today, at scale, while end-of-life volumes are still ramping.
For the UK, which has committed to a battery supply-chain strategy built substantially on circularity, disassembly automation is the missing middle: without it, the recycling capacity being built upstream will be fed by a manual process that cannot scale to meet it.
Where DevSpark fits
DevSpark's background is the exact discipline this problem needs: industrial robotics and machine-vision integration, edge inference, thermal imaging, and embedded, safety-rated control systems, built over more than a decade in industrial automation and vision. This is the engineering that turns a robot arm into a disassembly cell: the perception layer that identifies a pack and finds its fasteners, the control and safety layer that keeps people outside the high-voltage envelope, and the measurement that turns teardown into triage.
If you are a recycler, dismantler, OEM or fleet operator thinking about what happens to your packs at end of first life — or an automation company looking at the UK market — we would like to talk.
If you are a European technology company weighing the UK market, the funding route is worth understanding first: how EU cleantech companies can enter the UK with government co-funding →
Sources: EU Battery Regulation (EU) 2023/1542 and European Commission Digital Battery Passport guidance (mandatory 18 February 2027 for batteries above 2 kWh); Advanced Propulsion Centre award announcements (automated EV battery disassembly project; £9m to ten early-stage projects); The Manufacturing Technology Centre battery teardown facility; Altilium £18.5m EV battery recycling facility. Figures and dates current at publication; verify against primary sources before acting.
More from Info
Related reading
20 Aug 2026
How EU Cleantech Companies Can Enter the UK Market with Government Co-Funding: A Guide to DRIVE35 Partnership Structures
The UK's £2.5bn DRIVE35 programme co-funds up to 50% of industrialising zero-emission vehicle technology in the UK — and…
16 Aug 2026
Embedded Systems in Harsh Environments: What Actually Fails
Roadside, trackside and factory-floor electronics rarely fail the way the datasheet suggests. The failures come from con…
16 Aug 2026
Case Study: CCTV and ONVIF Platform for UK National Infrastructure
An embedded CCTV and video communications platform for a UK national infrastructure environment: embedded Linux and Open…
Bring us the challenge
The one that has been handed back, sits between two suppliers, or nobody can say is possible yet. A short call costs you nothing and you will speak to one of our consultants.
Engineer to engineer. No handoffs.
