AIRS ML · ABB
In production
Open to read the case study
AIRS ML
By AIRS MLAIRS ML builds ultra-efficient AI for industrial machines to detect failures before they happen without relying on cloud or extensive sensors. Our EdgeAI models run directly on existing hardware, reducing downtime, boosting reliability, and cutting maintenance costs. Our small language models then convey the data analysis and help make it easier for teams to interface with any technical documentation.
- Stage
- In pilot
- Industries
- Manufacturing
- Logistics
- Energy
- Deployment
- Hybrid
Why it is worth it
AIRS ML offers an industry-first fusion of on-device, real-time inference with low compute footprint. This leads to 1. Increased data privacy: No data leaves the vehicle, enabling easier compliance and tighter control. 2. Ultra-low Latency: Real-time inference happens right next to the batteries, ensuring immediate response for time-critical tasks. 3. Massive Carbon Reduction: Compared to cloud-based AI (325 kgCO₂/year per model), whereas edge solutions consume just 4 kgCO₂/year. 4. Bandwidth Efficiency: Instead of transmitting raw data needing constant connectivity, we process and compress sensor data into KB-sized insights. 5. Cost Advantage: Our AI runs on low-cost devices, eliminating the need for expensive high-speed links. Competitive Edge: 1. Lightweight models that can be deployed on constrained automotive ECUs without cloud dependency 2. Proven partnerships with institutions like the MTC, AMRC, and aerospace manufacturers for robust validation 3. Patent-pending technology focused on anomaly detection and performance prediction in high-value industrial assets
- Unplanned machine breakdowns in manufacturing lead to significant production losses, reduced operational efficiency, and increased maintenance costs. Traditional, non-AI-based solutions struggle to predict these failures effectively, resulting in frequent downtime, safety hazards, and shortened equipment lifespans. This reactive approach to maintenance not only reduces productivity but also leads to significant financial losses.
- Machine breakdowns in manufacturing lead to significant production losses and reduced efficiency. Predicting these failures is difficult, and traditional non-AI-based solutions introduce several challenges:
- 1. Unplanned downtime and decreased productivity.
- 2. Severe accidents and shorter machine lifespans.
- 3. Financial losses due to inefficient maintenance.
- Artificial Intelligence (AI) presents a promising solution to these issues. However, current AI solutions in predictive maintenance rely heavily on IoT and cloud computing, which come with several limitations:
- 1. High power consumption and cost.
- 2. Data privacy concerns due to cloud reliance also requiring high bandwidth.
- 3. Increased carbon footprint, which conflicts with sustainability goals.
- Industrial AI agents are transforming factory floors, but they often risk placing valuable machine data in the hands of third parties or on unsecured servers. We aim to develop AI agents that first prove accurate machine monitoring and then pioneer the use of AI for closed-loop control of machines, all while ensuring secure, local data processing.
Opening the deck…
Page 1
Best customer profile
- Our ideal customers are industrial companies with high-value machinery and costly downtime, including:
- Sectors: Aerospace, Rail, Automotive, Energy, Heavy Manufacturing, and Industrial Equipment
- Company Size: Mid-to-large enterprises (500+ employees) with large-scale operations or distributed assets
- Use Case Fit:
- a) Predictive maintenance for motors, pumps, compressors, CNCs, robotic arms, or EV chargers
- b) Situations where traditional condition monitoring (vibration, thermal, acoustic) is limited or too expensive to scale
- c) Legacy machines lacking built-in diagnostics or connectivity
- Specific Needs:
- Reducing unplanned maintenance and asset failures
- Improving uptime without relying on cloud infrastructure
- Deploying scalable AI on edge hardware, even with minimal sensors and limited historical failure data
- We’re a strong fit for asset-intensive industries looking to modernise operations without major capital expenditure.
Notable paying customers
- Yes, we have experience in working with large corporate partners. Our team is collaborating with major players in the manufacturing and aerospace sectors, including the Manufacturing Technology Centre (MTC). We secured an Innovate UK grant of £50,000 and a £7,500 investment from the High Value Manufacturing Catapult (HVMC) for a project on Edge AI for Anomaly Detection in a Kuka KR 6 robotic arm at the Manufacturing Technology Centre (MTC) where we achieved an accuracy of 96%. We have also cleared Stage 1 of the ATI SME grant on EdgeAI for compressive sensing and predictive maintenance in aviation manufacturing, worth £1.5 million in partnerships with MTC, AMRC and CloudNC. We also have letters of support from GKN Aerospace, Beverston Engineering and NCMT. This experience has given us a deep understanding of navigating the technical, procurement, and regulatory complexities that come with enterprise-scale deployments.
Key milestones
AIRS ML has achieved several key milestones, both commercial and technical: 1. Patent Filing: We have filed our first patent for EdgeAI-powered industrial monitoring and control systems. 2. Grant Award: Secured an Innovate UK grant of £50,000 and a £7,500 investment from the High Value Manufacturing Catapult (HVMC) for a project on Edge AI for Anomaly Detection in a Kuka KR 6 robotic arm at the Manufacturing Technology Centre (MTC) where we achieved an accuracy of 96%. 3. Partnerships: We have cleared Stage 1 of the ATI SME grant on EdgeAI for compressive sensing and predictive maintenance in aviation manufacturing worth £1.5 Million in partnerships with MTC, AMRC and CloudNC. We have letters of support from GKN Aerospace and NCMT. 4. Pilot Projects: Successfully completed initial pilots, demonstrating the viability and impact of our EdgeAI technology in real-world industrial settings. 5. Funding: Raised $120K investment from Techstars, enabling us to work on product and business development. 6. We are also part of the smart mobility accelerator In collaboration with EIT Urban Mobility, STARTUP AUTOBAHN powered by Plug and Play, and mobilityXlab. These milestones position AIRS ML as a strong player in the industrial AI market with a clear path to scaling its solutions.
Customer case
One customer deployment, shown below.