Traditional AI models and computer vision applications require increasingly powerful GPUs, leading to higher energy consumption and complexity—a daily challenge for web developers. To address this, Focoos AI, an Italian startup spin-off of Politecnico di Torino (incubated by I3P and part of the ESA BIC Turin program), has developed a web platform that guides developers through designing, creating, and deploying computer vision models.
The Turin-based startup has secured €2.65M in seed funding led by Galaxia, PiemonteNext, VC Partners SGR, Vertis SGR, Vento, and Exor Ventures. The company’s total funding and valuation remain undisclosed.
This investment will fund new platform features, enhanced model customisation, improved performance monitoring, and the development of efficient foundational models for immediate industrial use. The funding will also support the platform’s launch, international growth through strategic partnerships, market expansion, and growth of both R&D and commercial teams.
Creating ultra-efficient computer vision models across industries
Founded in November 2022 by Antonio Tavera and Fabio Cermelli, Focoos AI emerged from the Department of Automation and Computer Science (DAUIN) at Politecnico di Torino. The founders set out to create ultra-efficient computer vision models to transform AI implementation across industries.
While traditional AI models demand powerful, energy-hungry GPUs and complex development, Focoos AI’s ultra-efficient models significantly reduce computational needs. Focoos AI aims to simplify and accelerate the development and integration of neural network models for computer vision, offering pre-trained and customisable networks optimised for energy efficiency and performance. Tavera shared with TFN during WAICF: “We are developers to create efficient computer vision models that can run on low-power edge devices like smart cameras, satellites, and drones. This allows for complex computer vision applications in manufacturing, quality control, precision agriculture, and damage recovery.”
Their solution works seamlessly across hardware types, from high-performance servers to edge devices, addressing the challenge of integrating AI models with existing infrastructure. The platform automates neural network creation and training, optimising for specific devices while reducing development time and effort. This approach helps businesses minimise their carbon footprint and meet sustainability goals through reduced computational resources and energy use.
Antonio Tavera, CEO and Co-Founder of Focoos AI, said, “We want to change the rules of the game: make artificial intelligence accessible to everyone, eliminating technological and economic barriers. AI must be designed to benefit humanity the most, with a balance between innovation and sustainability. And we at Focoos AI are more determined than ever to lead this change.”
The future of efficient AI vision: Focoos AI’s Anyma
Anyma, Focoos AI’s platform, delivers highly efficient, low-power computer vision models. The platform reduces energy consumption while speeding up AI solution development and integration. These models quickly adapt to specific company needs, enabling computer vision applications in smart cities, industrial production, and satellite image analysis.
The platform automatically designs and trains optimised neural networks for specific hardware setups. This allows businesses to implement AI solutions without costly hardware upgrades or complex development work, making advanced computer vision accessible and cost-effective.
Fabio Cermelli, CTO and co-founder of Focoos AI, stated: “Our web platform, already in beta for early users, will soon be accessible to everyone and will set a new standard for high performance with minimal consumption. Behind this technology stands tireless work and an extraordinary team that has never stopped. This investment enables us to accelerate further, adding momentum to our vision.”
Models created with Focoos AI’s proprietary technology are up to 4 times more efficient and 10 times faster than market standards, cutting cloud costs and allowing AI applications to run directly on edge devices like cameras and drones.
Redefining the application paradigm of computer vision
Speaking about next year’s goals, Tavera shared with TFN: “We aim to become the ‘go-to platform’ for developers who want to develop efficient computer vision models. The goal is for Focoos to be the de facto standard platform and library (available on GitHub) developers use for this purpose. Additionally, we plan to enter the US market, with plans to do so by June of this year, and grow the team from the current 11 employees to at least 20 over the next 3 years.”
The investors share optimism about Focoos AI’s innovation. Claudia Pingue, Senior Partner and Head of the Technology Transfer Fund at CDP Venture Capital noted: “We are pleased to renew our support for Focoos AI, a company redefining the application paradigm of computer vision. Its vision of accessible and efficient A, capable of operating on low-power devices, represents a decisive step toward the large-scale adoption of these technologies.”
Roberta Vercellotti, Founding Partner of VCP, stated: “Focoos AI’s frugal approach to AI models for computer vision has a wide range of potential applications and use cases, and this investment round will allow the company to expand its reach to developers worldwide seeking the best performing and efficient models.”
Alessandro Papoff, Investment Manager of Vertis, stated: “In Focoos AI, we have found an ideal solution to democratise artificial intelligence, making computer vision accessible to any sector and device. This technology makes it possible to cut operational and development costs, ensuring a perfect balance between efficiency and accuracy of neural networks.”
“The innovation proposed by Focoos AI has broad application potential in different industrial fields,” added Giuseppe Scellato, President of I3P and Coordinator of ESA BIC Turin.
The post Making computer vision accessible to any sector and device: Italian Focoos AI grabs €2.65M to make this a new reality appeared first on Tech Funding News.