Beyond the gadgets: Six deeptech companies to watch at IFA 2026
Image by EngineAI, the Chinese company that will participate in IFA’s “Robots on the Runway”, where humanoids, quadrupeds and companion robots will dance, pose and perform other demonstrations
IFA Berlin opens tomorrow with its familiar mixture of televisions, laptops, smart appliances and unexpected solutions to household problems most people did not know they had.
Look beyond the gadget launches, however, and IFA 2026 begins to resemble something more ambitious: a proving ground for deeptech companies trying to enter mainstream markets.
IFA Next, the event’s innovation hub, expects around 300 exhibitors from startups, scale-ups, research organisations and international pavilions. Its programme ranges from humanoid robotics and computer vision to exoskeletons, smart glasses, advanced manufacturing and health technology.
The expansion reflects a wider change in deeptech commercialisation. Technologies previously designed for factories, hospitals or research facilities are being adapted for homes, offices and individual users. That transition introduces a demanding set of requirements. Products must become affordable, intuitive, reliable and safe around people who have received no specialist training.
Among the companies to watch are Germany’s NEURA Robotics and Repartly, Chinese robotics developers EngineAI and Hypershell, Taiwan’s Next Integral and the PrimeBOT brand of Swancor New Materials. Together, they span physical AI, humanoid robotics, industrial computer vision, circular electronics and powered exoskeletons.
Not every robot dancing on a stage or AI device generating a summary will survive the journey into everyday use. Most specifications and performance figures available before IFA come from the exhibitors themselves; live demonstrations will provide an opportunity to examine some of those claims more closely.
Here are six deeptech companies worth watching when IFA 2026 runs from 4 to 8 September.
1. NEURA Robotics: building a European physical AI ecosystem
German scale-up NEURA Robotics is one of Europe’s most closely watched robotics companies. It combines artificial intelligence, sensors and robotics to develop machines intended to perceive their surroundings and work alongside people.
At IFA 2026, founder and CEO David Reger will use his keynote to argue for a European ecosystem around “physical AI”: artificial intelligence embodied in machines that can perceive and act in the physical world. The larger question is whether Europe can build the hardware, software, manufacturing capacity and markets needed to compete with robotics companies from the US and Asia.
As MTN previously reported, NEURA announced a Series C financing round of up to $1.4 billion in June 2026. The company says the capital will support serial production, its Neuraverse platform and a planned network of real-world robot-training environments called NEURA Gyms.
IFA places that industrial ambition before a different audience: consumers, retailers and appliance manufacturers.
What to watch: whether NEURA can connect its broad physical AI vision to convincing applications outside controlled industrial environments.
2. EngineAI: separating humanoid capability from choreography
Founded in Shenzhen in October 2023, EngineAI develops humanoid robots and embodied AI systems. The company says its in-house capabilities extend from robot design and actuators to motion-control algorithms and AI.
Its existing platforms include the compact PM01 humanoid and the larger T800. The PM01 is designed as an open development platform, while the T800 is being positioned for applications including logistics, hospitality and human-robot collaboration. Those use cases and performance specifications remain company claims rather than independently demonstrated commercial deployments.
EngineAI is also confirmed for IFA’s “Robots on the Runway”, where humanoids, quadrupeds and companion robots will dance, pose and perform other demonstrations.
These displays can reveal balance, actuator coordination and motion control, but they say less about how a robot responds to an unfamiliar home or workplace. A backflip can be rehearsed. Picking up an object that has fallen behind a chair requires perception, planning and adaptation.
What to watch: whether EngineAI demonstrates autonomous interaction with changing surroundings rather than another polished sequence of pre-programmed movements.
3. PrimeBOT: can a transformable personal robot become useful?
PrimeBOT is not an independent startup but a consumer-robotics brand launched by Swancor New Materials, a company listed on Shanghai’s STAR Market. Its T1 personal robot is designed to switch between humanoid and quadrupedal forms, allowing it to move on two or four limbs.
The company presents the design as a way to support indoor and outdoor mobility, autonomous filming, education and companionship. A quadrupedal configuration could offer greater stability on difficult terrain, while a humanoid form provides a more familiar interface for interacting with people.
PrimeBOT plans to demonstrate the T1 alongside its smaller Q1 humanoid at IFA Berlin and at the ShowStoppers media event. It describes the T1 as the first transformable personal robot, a claim that will require scrutiny given the many experimental robots capable of changing configuration.
The trade-off is complexity. More joints, moving structures and operating modes can create additional failure points, increase weight and consume more power. Transforming successfully on a runway is also different from doing so repeatedly in a cluttered home.
What to watch: whether the changeable body provides a practical advantage or primarily makes the robot more memorable in a crowded exhibition hall.
4. Next Integral: edge AI that learns what normal looks like
Not all physical AI needs arms and legs. Taiwan’s Next Integral is bringing NEXUO, an edge AI platform for industrial inspection, safety monitoring and crowd analysis.
Conventional machine-vision systems often need collections of labelled defects before they can recognise a problem. NEXUO instead learns what normal products or behaviour look like from good examples, then identifies deviations.
According to its IFA profile, the software runs locally on Nvidia Jetson hardware without sending data to the cloud. Next Integral says the platform is being used in smart-city safety pilots in Taiwan and is under consideration for computer numerical control machinery and manufacturing inspections in Taiwan and Europe.
The approach could be useful in factories making small batches of varied products, where manufacturers cannot collect thousands of examples of every possible fault. On-device processing may also reduce latency and keep sensitive industrial data on site.
What to watch: how reliably the edge AI system distinguishes a genuine defect from harmless variation—and whether customers can understand why it flagged an anomaly.
5. Repartly: repair infrastructure for increasingly complex appliances
Repartly is the least theatrical company on this list. Founded in Germany in 2021, it repairs and refurbishes electronic modules and spare parts for washing machines, dishwashers and tumble dryers.
That may not look like frontier technology, but modern appliances can be discarded because a complex electronic assembly has failed while the rest of the machine remains functional. If manufacturers no longer supply the module, or a replacement costs almost as much as a new appliance, repair becomes difficult to justify.
Repartly offers fixed-price electronic repairs, refurbished replacement modules, online diagnostics and installation guidance. The company says it supports more than 100,000 appliance models and has helped keep more than 30,000 devices from being scrapped. Those figures are self-reported.
Its appearance at IFA 2026 is particularly timely. EU member states were required to begin applying the bloc’s Right to Repair Directive from 31 July 2026. The rules include requirements for manufacturers to provide repair information and access to spare parts at reasonable prices for covered products.
What to watch: whether component-level refurbishment can become scalable infrastructure for circular electronics, or whether proprietary designs and limited technical information continue to frustrate independent repair.
6. Hypershell: taking consumer exoskeletons onto hiking trails
Exoskeletons have traditionally been associated with medical rehabilitation, military research and physically demanding industrial work. Shanghai-based Hypershell, founded in 2021, wants to turn them into consumer wearables.
The company develops powered hip exoskeletons designed to detect the wearer’s movement and adjust their assistance according to terrain, activity and effort. Its products are aimed principally at walking and outdoor recreation rather than medical rehabilitation.
Hypershell says it has sold more than 10,000 exoskeletons worldwide. Although that figure has not been independently verified, the company has moved beyond the prototype stage: mass production and delivery of its first commercial systems began in 2024.
The engineering challenge remains substantial. A consumer exoskeleton must be light enough to carry, responsive enough to feel natural and efficient enough for its battery to last. It must also respond predictably when assistance changes or the battery runs low.
What to watch: whether the support feels genuinely intuitive—and whether the benefit justifies wearing and charging another piece of hardware.
From deeptech demonstration to useful product
These six names represent different parts of the deeptech ecosystem. Some are trying to place intelligence inside general-purpose robots. Others are applying specialised sensors, edge computing, powered mechanics or repair expertise to narrower problems.
IFA gives them access to consumers, retailers and industry partners, but it also exposes their technology to a more difficult standard. Outside the laboratory, technical novelty is not enough.
Liked this article? You can support our independent journalism via our page on Buy Me a Coffee. It helps keep MoveTheNeedle.news focused on depth, not clicks.