15 Edge AI Companies: Independent Software Vendors (ISVs) to watch in 2024
Currently most enterprises are still exploring the possibilities of AI, with initial implementations in simple tasks such as administration and customer service. However, there is a motion from C-suites and investors to harness AI in automating the management of mission-critical tasks in a bid to boost productivity. STL Partners look at 15 Independent Software Vendors (ISVs) that are using edge computing technology to ensure maximal performance and reliability in their AI models.
In this article, we showcase 15 companies using edge AI that have developments to keep an eye on in 2024.
The recent take-off in adoption of AI has been triggered by development in processing power and availability of investment capital, enabling the advancement of cutting edge, dynamic models. Edge computing permits enterprises to run mission-critical applications enabled by AI effectively. Enterprises are assessing the value of AI models to their businesses and there is an increasing demand for developers and infrastructure providers involvement.
The below list includes a range of software vendors, from start-ups to those established in the ecosystem. We have requested companies to provide details on their offerings in-tandem with their opinion on what the future holds for edge AI. We can see the variety of edge AI use cases being enabled ranging from global manufacturing operations to security analytics to retail immersion.
Companies are listed alphabetically.
Aarna Networks provides application, infrastructure, and network services edge orchestration at scale. In 2023, Aarna collaborated with Radisys and Nvidia, to introduce a ‘RAN-in-the-Cloud’ Proof of Concept (PoC), where a fully hosted 5G RAN is available as a service within a multi-tenant cloud infrastructure. The containerised solution operates alongside other applications offering flexibility and scalability in an end-to-end software stack. Compute power is reallocated for running 5G and AI applications during periods of low demand with surplus instances of DU/CU automatically removed. The opposite occurs when traffic is high with automated workload switching. This allows network operators to increase efficiency while introducing new monetised edge applications and services such as edge AI, video applications, CDN, and more.
Aarna Networks is also considering how to incorporate AI/ML into future product roadmaps using Generative AI and its applications at the cloud edge and within private networks. Aarna’s hypothesis is that companies are uncomfortable leveraging public AI models for private data. This has led them to consider the use of private AI models that can be trained on proprietary data and then deployed at the cloud edge. The company is considering applications of its AI in various industries, particularly for those where security and privacy are of paramount concerns, such as healthcare and finance. To start, Aarna Networks is looking to offer Generative AI through a secure, private, tunable, fully managed LLM. The greatest challenge in its partner ecosystem is translating the benefits of AI into business value and coaxing forward thinking practitioners to get started. Aarna is hopeful that the exploding interest in AI will help bridge this gap.
Agricair deploys an AI service for different sectors in animal agriculture with the goal of automating management on farms. The agriculture sector is trying to combat negative media repuration and poor public perception when it comes to treatment of animals. A way to show their level of care is through AI monitoring. They are harnessing AI to monitor various aspects of animal care, including tracking information related to feed areas, treatment, illness detection, food allowances, and more. AI is used to process video feeds for insights and identify situations of concern, quantifying data that should be tracked. Agricair emphasises the use of edge computing to process video data locally at the farms, driven by the need to minimise data transmission over networks, especially in remote farm locations where network capabilities may be limited.
Whilst Agricair acknowledge advancements in AI technologies, they don’t require cutting-edge AI for their specific applications, instead using simpler AI models that more efficiently solve their problems. However, they stress the need to be very specific when handling training data; employing continuous random sampling which aids in handling data efficiently. The biggest challenge they face is communicating the capabilities and limitations of AI to their customer base. Telco’s involvement with Agricair and the wider agriculture vertical is limited as they are not willing to provide infrastructure to assist individual farms.
AI EdgeLabs’s platform deploys AI technology that identifies, responds, and remedies attacks and threats at the edge and IoT infrastructures. It is in the process of commercialising its edge-native cybersecurity solution that addresses challenges specific to each industry. EdgeLabs has been engaging with a diverse range of verticals, including telecoms, oil and gas, and retail. Its solution is hardware-agnostic and can be deployed rapidly, within a matter of days.
AI EdgeLabs notes that AI capabilities are driving the acceleration in demand for edge solutions with many industries moving to the edge for applications such as predictive maintenance. AI plays a crucial role in its solution, being utilised to enhance security, and protect networks and IoT devices at the edge mitigating threats in real-time. It acknowledges the significance of tailoring its AI solution to the specific needs and preferences of its clients, allowing them to define important metrics.
Aotu’s venture started in 2016 seeing the AI vision business as a very fragmented market. Aotu initially built an AI Vision Operating System (OS) to tackle problems of AI vision applications. Aotu’s BrainFrame® OS is a development of this, providing intuitive graphics UI for a user to connect to any USB or IP camera. Its use cases include factory safety and traffic safety management at warehouses and factory floors, security and analytics at homes and stores, and many more. It allows algorithm configuration by users and alarms to be set based on any detections or recognitions. Aotu also provides a standard API for networks to make it easier for developers and systems integrators to work with its solution.
Aotu hypothesises that the exponential growth in general AI will also happen with edge AI, but in a very different way in terms of how the products and services are deployed. For general AI, it is ‘Chat’, as it is probably the most intuitive way for users to consume information. However, when it comes to the edge vision to tackle the problems for millions of different applications, it is not ‘Chat’, it is a system that can deploy and manage many specialized AI models for specific tasks at the edge. If you think of general AI as a big brain that does everything with ‘Chat’, then edge AI is about managing the collaboration of many small brains to accomplish a specific task in the physical world.
Avanseus Technologies focuses on predictive operations using AI, developing data driven, context aware enterprise solutions. Avanseus aims to leverage AI/ML to anticipate and reduce critical faults that occur within the networks of various service-related areas with a particular focus on telecom operators. The foundation of the solution lies in the integration process, involving two data sources: alarms and trouble tickets. Training on historical data is critical key part of its process, with the AI engine needing to be trained and predictions prioritized and enriched with fine grained root cause analysis. Avanseus typically interfaces with CSP customers but has collaborated with network providers, such as Nokia and Ericsson; it is also exploring other markets with interest in its solution from the data centre, utilities, and energy industries.
Avanseus recognises that AI can make network management more recurrent, allowing them to predict and prevent network faults and issues more effectively. Avanseus has implemented an Augmented Operations focused processor that automates actions based on predictions, applying policies and triggering network changes in response to faults independent from the network domain, demonstrating its trust in AI for automated decision-making. Avanseus view AI as a valuable tool for managing power consumption and automating processes, specifically in telecommunications, data centre, IoT and utilities sectors. Whilst Avanseus encourages the use of AI and cloud deployment, it understands that some customers may have security and technical requirements that prevent direct cloud usage.
Part of the Atos Group, Eviden is a next-gen technology player in data-driven and sustainable digital transformation with a portfolio of patented technologies. Eviden has positioned itself as a provider of value-added services and solutions for customers who are increasingly concerned about sovereignty and security issues. It leverages the synergies between its core activities in digital, cloud, big data, and security, and takes advantage of its combination of services and technologies across the digital continuum.
Eviden wants to drive consumption of edge infrastructure, delivering pre-trained AI models out of the box to customers and managing their lifecycle, including updates and improvements. Eviden provides end-to-end solutions with hardware and software that deliver AI inferencing and streaming analytic capabilities with data security at the edge. It is engaging with analyst communities to provide and grow its solutions whilst addressing securities associated with edge devices and AI models as they become part of networks. Eviden views AI as a crucial technology that can drive innovation and efficiency in various industries. Eviden is seeing a big spike in interest in it from C-suites of high-level enterprises. They attribute this uptick to hype that was triggered from the respective releases of ChatGPT 3 and 4. It recognises that the one-size-fits-all approach won’t work in the AI and edge computing landscape, suggesting a need for various specialised solutions.
Guise AI solves real world problems for enterprises with purpose-built software solutions at the device edge. Guise makes machine learning accessible for enterprises by building models optimised for the device edge, using CPUs not GPUs. It built a no-code edge platform, Guise EdgeOps, to provide the ability to manage devices and deploy, orchestrate, and manage AI workloads at the edge in a secure environment. Guise EdgeOps collects the inference and drift data from edge devices, building a Unified Data Store that facilitates the ability to retrain models at the edge and enables a hybrid cloud. The Unified Data Store allows an enterprise to choose its data gravity and enables “speed to ROI” with POCs and accurate models in production.
According to Naga Rayapati, Founder and CEO of Guise AI, “It is critical for companies to do AI inferencing where the data is actually generated in order to grow revenue, achieve operating efficiencies, and improve safety.” Guise AI views Edge AI as revolving around scaling and security; with models needing to pull data from hundreds to thousands of edge devices in a secure manner. Guise also believes customers require use cases that can run seamlessly in the field to overcome issues of latency, bandwidth, privacy and security, energy consumption, and cost. It sees data generated at the edge as not new, with the need to properly leverage that data combined with the proliferation of IoT devices demands lightweight models that can run at the point of data generation not just in the cloud.
Intent HQ provide an AI-enabled customer intelligence platform that generates insights from complex behavioural data streams. These insights show customer intent so Marketing teams can create more relevant marketing and customer experiences. Intent HQ’s new Edge solution is available to brands with a mobile app, providing a new set of behavioural insights with applications across customer experience, customer value growth, and customer retention use cases. Intent’s customers access the solution via a proprietary software development kit (or SDK) that integrates into brands’ mobile applications. The on-device edge AI software analyses various first-party data signals from the phone to piece together a person’s real-world profile, restricting the data that leaves the device. Segmentation profiles and campaign activations can be pushed to the device, and the on-device AI evaluates its applicability to that customer. This approach emphasizes data privacy and the value of local processing. Intent HQ is engaged in B2B2C partnerships, including collaboration with a major insurance company.
Intent HQ views AI as a fundamental technology providing enormous opportunities for brands adopting edge computing, particularly when deployed in consumer mobile applications. Its focus on AI at the edge aligns with the growing importance of localized data processing and privacy concerns in the tech industry. Intent has invested in low-level code development to ensure its AI technology has a minimal device footprint and ultra-low battery consumption and bandwidth, making it efficient for on-device processing. Intent HQ partners with customers to implement its solutions effectively in consumer mobile applications.
Ipsotek, an Eviden business, specialise in video analytics and have transitioned from a monolithic to a containerised microservices based architecture making use of cloud-native technologies. Ipsotek provide solutions that integrate with camera systems and has moved from GPU-based setups to Kubernetes and Ansible for deployment. It collaborates with technical parties and specific use case partners for testing and PoC projects. It is aiming to certify its solution to run officially on Red Hat’s OpenShift container platform. Oil and gas, security, transportation, and leisure are among the industries that it sees potential for its solutions.
Ipsotek actively incorporate AI and machine learning technologies into its solutions leveraging it for data processing and insights. Ipsotek is exploring the use of AI both at the edge and in the cloud, seeing the value of the edge for low latency and real-time analytics whilst recognising the utility of cloud-based aggregated analysis. Ipsotek view data quality and meaningful output generated by AI as vital in delivering valuable insights. AI is a key technology for making its solution easy to deploy, maintain and scale.
Litmus provides a unified data platform to help manufacturers collect, contextualize, and analyse data. With off the shelf templates and integrations across the enterprise, Litmus aims to make operational technology data accessible to the right applications. Using Litmus Edge, users can run machine learning models for practical scenarios to predict, classify and detect anomalies. Working with Litmus and through its network of industry partners to provide end-to-end solutions and go-to-market (GTM) strategies, manufacturers can access ready-made industrial solutions.
AI is viewed as essential for improving the adoption and practical deployment of Litmus’s algorithms and applications. It’s seen as a key technology for running applications cost-effectively at the edge. Hardware acceleration is playing a crucial role in enabling new use cases for edge and AI, particularly in video-based analytics and predictive quality control. Litmus believe that the early adopters of AI are experiencing business improvements, such as increased yield. Litmus acknowledges challenges associated with AI adoption, including in-house capabilities, understanding the business case, and the overall maturity of both AI technology and customer readiness.
MicroAI provides next-generation endpoint and edge AI-enabled solutions to companies within the telecom, manufacturing, industrial, automotive, and financial sectors. MicroAI’s, proprietary, AI/ML technology — AtomML™ — enables a breakthrough approach to using artificial intelligence to improve the performance, health, and security of devices, machines, and networks. MicroAI is currently delivering AI-enabled use cases in the areas of closed-loop asset monitoring, predictive maintenance, network quality of service, and cybersecurity. The technology does not rely on cloud-based processing and instead operates entirely on edge devices. Inferencing and analysis take place on the machine or device itself without the need for data to be sent to a centralised cloud. This approach reduces the cost associated with transferring data back and forth, while also minimizing the impacts of data latency.
MicroAI’s technology was designed to be lightweight and integration friendly. Higher levels of automation, real-time point-to-point data streaming, low/no code development, and reduced reliance on extraneous hardware all combine to provide solutions that deliver value within weeks instead of months. An ecosystem of intelligence is created without having to scuttle expensive assets and/or to procure new supporting solutions. Micro AI suggests that depending on the scope of integration, this can save the enterprise 10’s to 100’s of thousands of dollars.
OpenNebula is an open-source cloud management platform created to simplify the deployment and management of virtualised infrastructures and private and hybrid cloud solutions. It is currently working on becoming an AI-enabled platform through innovation initiatives like the EU-funded COGNIT Project. These would use OpenNebula to aggregate private and public resources for IoT/edge applications that need to off-load heavy computational tasks to the cloud-edge continuum, abstracting the underlying infrastructure to these serverless workloads. It can integrate with various APIs related to network management, billing, and CRMs. Its solution enables computing, storage, and networking resources to allocate and manage resources for different workloads and projects.
OpenNebula views AI as the key ingredient in edge computing, optimising resources as edge node counts increases. It acknowledges the growing importance of edge computing in the industry and the need for telcos to catch up with industry conglomerates. OpenNebula also emphasises the importance of testing and real-world network experience when implementing AI solutions, prioritising a more hands-on approach as opposed to theoretical discussions. Finally, it recognises ISPs as important going forward to engage with.
Ranial Systems provides M2M/IoT platforms and solutions enabling real-time operational intelligence and autonomous control capabilities. Ranial’s R&D is funded by the federal agencies, NSF and IEeE, with its IP and patent granted worldwide Ranial’s incremental learning solutions deliver actionable insights and controls automatically. Its edge analytics runtime provide interface to a variety of legacy equipment and control systems with real-time decision intelligence capabilities. The edge nodes act on events at the point of action giving real-time data aggregation, alerts/notifications, and advanced analytics with incremental learning. Ranial has made significant contribution through the convergence of edge and AI that power industrial automation across various industries, including utilities and industrial manufacturing. Ranial is particularly focused on real-time data processing, adaptive learning, and decision-making at the network’s edge.
Ranial Systems has a strong focus on the convergence of edge computing and AI to help enable advanced capabilities horizontally (peer-to-peer Collaboration). Ranial cite its patented AI powered automation as critical in helping analyse discrete sensory feeds and extract meaningful insights that can provide contextual relevance to data making it more valuable for responsive decision-making, predictive maintenance and early detection of potential treats or anomalies. Ranial further emphasises how cognitive edge computing naturally complements IoT applications, as it allows real-time processing and decision-making at a reduced latency. Ranial stress how incremental learning and training of AI models at the edge is essential, especially in scenarios where real-time insights and actions are needed.
Summit Tech enables real time communications and experiences across a wide array of devices. Its flagship Odience platform delivers highly immersive and interactive live events for brands, retailers, and influencers. Odience utilises telco and hyperscaler edge computing to deliver personalised 360° video at low latency and low bandwidth, at scale to mobile, PCs, and VR headsets. It positions itself as a platform to bring virtual presence, allowing users to view, interact, and engage with the hosts, products, and content as if they were virtually attending the event. Merchandise can be viewed and purchased within the app, leveraging, QRs, chatbots and AI.
Summit Tech integrates AI across its platforms, using AI enabled chatbots in livestream events for more natural and meaningful chat conversations. Video analytics AI plays a further role to enable users to select products of interest featured in the live video through a simple long press on the display. AI can also help personalise the Odience service and provide targeted content to users. Summit believes AI coupled with edge are powerful technologies to enhance virtual experiences which delight users.
Unmanned Life enables orchestration of autonomous robotics, offering a software platform that can unify fleets of drones and robots to work together regardless of individual operating systems, simplifying its deployment and management for the end user. The focus today is on drones due to market demand with its solutions primarily being used for automated inspection, security surveillance, and worker safety. Unmanned collaborate to help with configurations, provide network connectivity, and assist in developing GTM strategies. Unmanned has recently pioneered a joint solution for autonomous security and inspection at the Port of Antwerp leveraging 5G and edge computing.
Unmanned Life harnesses AI in its drone integration to improve capabilities. It uses third-party AI primarily for computer vision which enables drones to perform tasks such as human detection and real-time video analytics. Combined with edge computing, the system can offboard these analytics from the drones, meaning heavy computational power is not needed on the drone itself, therefore saving battery life. Unmanned recognises the value of AI in addressing specific use cases, such as security and infrastructure inspection. Increasingly trained AI models from accurate and intricate data sets improve its solutions over time, providing more accurate responses. Unmanned Life acknowledge the hype surrounding AI and instead elect to focus on the practical applications within its domain as a tool that can enhance its technology.
The list was created from an interview programme that these companies participated in, therefore this list may not be comprehensive and representative of the full spectrum of edge AI companies making waves in 2024. At STL Partners, we are striving to help companies with their edge computing strategy. Discover the best edge monetisation possibilities, understanding the key players, and inform your go-to-market strategy through our Edge Use Case Directory.
To subscribe, or simply find out more, feel free to click here to arrange a call with our team.
Companies are listed alphabetically.
Edge computing market overview
Edge computing market overview
This 33-page document will provide you with a summary of our insights from our edge computing research and consulting work:
When will edge private cloud supplant colocation?
This article will outline the core differences between edge colocation and edge private cloud, explore the merits and demerits of each.
Where is the sweet spot for edge computing?
Any situation that requires a blend of local compute-like and cloud compute-like capabilities would benefit the most by using edge computing. Such use cases include, but are not limited to, advanced predictive maintenance, remote patient monitoring and video analytics for security.
Where is money being made in edge computing?
This article cuts through the chatter and presents a pragmatic view for where edge revenues are being made.