Junto Innovation Hub
IT Get in touch →

Magazine Artificial Intelligence

Artificial intelligence for scaling materials and sensors in Europe

3 min read

Artificial intelligence for scaling materials and sensors in Europe

L’artificial intelligence is the glue that turns materials, sensors and autonomous machines into products that work in the field. This article describes how five European startups, SweGaN, Birdsview, SpaceAM, Gravis Robotics and amber, combine wafers, radar, ultralight payloads, excavator retrofits and data layers to reach commercial solutions. The operational and financial choices shown here are useful to a founder who wants to understand how to move from prototype to scale, with concrete examples and measurable constraints.

Materials, sensors and the chain that decides scale

SweGaN produces epitaxial gallium nitride wafers on silicon carbide, that is thin crystalline material layers deposited on a substrate that improve radio performance and high-voltage handling. SweGaN closed a Series B round of €12.09 million and brings total raised to €35.4 million. The resources are intended to increase production and reach customer qualification milestones. For an entrepreneur, this means that the availability of quality materials like GaN-on-SiC can become a bottleneck or a direct competitive advantage for products that require reliable communications.

Birdsview chose a different upstream approach: it applies Ground Penetrating Radar with algorithms that transform raw readings into interpretable 3D models without drilling concrete. The Norwegian startup raised €3.7 million and declares customers in nine countries after commercial launch. Automation of interpretation is in demand. Comparing SweGaN and Birdsview helps see two leverage points: materials that enable physical performance and software that makes complex data interpretable, both necessary for mission-critical sensors.

If the wafer supply chain jams, radar and radio solutions will not reach the market either. Therefore every technical roadmap must include an inventory of semiconductor suppliers, material qualification criteria and back-up plans. A founder should define supplier reliability metrics and qualification deadlines that align with commercial contracts or pilot projects. Supplier reliability metrics must align with contracts.

The availability of materials can block scale or become a competitive advantage.

How artificial intelligence coordinates machines, payloads and decisions

Ultralight payloads slow descent without parachutes. The payloads integrate on-board intelligence to turn raw data into real-time insight. Founded in 2019, SpaceAM secured contracts with the UK Ministry of Defence and the European Space Agency. It received £2 million, that is €2.3 million, in defence funding. Defence funding means targeted contributions for technology development related to national security. Such funding can entail requirements on intellectual property and export controls.

Autonomy increased excavator productivity by thirty percent. The company raised €172 million in a Series A round. This financing follows early rounds and aims to scale product and market. The post-money valuation is €862 million, that is the company’s value after accounting for the capital raised. Gravis reports productivity improvements of about 30 percent, that is three machines doing the equivalent work of one traditional machine. The company uses sensors and control policies to allow the operator to supervise multiple units.

Putting together the experiences of SpaceAM and Gravis teaches that the value of artificial intelligence is not only in the models. The value also lies in the ability to operate in adverse conditions. Solutions require local compute to reduce latency, compression of insights for transmission and pre-loaded operational rules. Founders must design workflows that pair local sensing with a data governance layer. In this way autonomous decisions already come accompanied by permissions and logs.

Artificial intelligence is useful only if it operates in real conditions.

From field trials to the numbers that matter: governance, customers and capital

AI data layer provides structured context. amber, founded in 2021, closed a Series A of €7 million. The company states that its approach reduces useless work by generative models by providing structured context. Integrating a platform like amber with Birdsview’s sensors or Gravis’s retrofits means that actions executed in the field arrive already with rules, authorizations and an audit trail. These elements are required by enterprise customers and public tenders.

On the contractual side, SpaceAM shows that agreements with public bodies accelerate entry into defence and space markets. Such agreements can however impose constraints on IP and test conditions. Birdsview and SweGaN illustrate that customer qualification requires trials on different sites and technical specifications. For founders it is crucial to negotiate milestones that unlock payment tranches only after clear and measurable technical verifications.

Investor choice determines product or prototype fate. SoftBank funded Gravis as the sole investor in a round that is the largest for construction robotics. This brought resources and global network. Industrial investors like Foresight Group support SpaceAM with capital and governance. Decisions on terms, governance and market targets determine whether the technology becomes a product or remains a prototype. The next operational step is to validate GaN suppliers, run pilot tests with productivity and reliability metrics, and build a data layer that makes artificial intelligence repeatable and controllable.

Sources:

Related articles