Share This Article
Embedl offers AI deployment without the need for data centers
The Swedish startup Embedl aims to advance AI capabilities. The company is working on creating artificial intelligence that won’t require servers or data centers. Edge AI will significantly reduce the energy costs of training and maintaining AI models. The idea of the deep tech startup attracted investors, and the company received €5.5 million in funding.
Embedl is an independent entity from Chalmers University of Technology. Embedl conducted a Series A funding round to provide funds for developing its idea. Chalmers Ventures was the key investor. Fairpoint Capital, Spintop Ventures, STOAF, and SEB Greentech also participated in the round. Taking the current funding into account, the total investment in the startup amounts to €12 million.
Technology features
According to experts, the Embedl project could be the next step in advancing artificial intelligence. It will enable AI to run on devices without using the cloud.
It is worth noting that technologies for running AI on the edge already exist. However, these technologies are currently only accessible to specialists. Embedl aims to make this technology accessible to developers of various products. To achieve this, the company is carrying out the following tasks:
- creating a basic AI infrastructure for various sectors;
- developing ways to optimize the deployment of technology on the periphery;
- enabling the deployment of models without unexpected costs and high energy consumption.
The startup focuses its activities on sectors for which implementing AI systems is essential. These include the defense industry, robotics, and automotive engineering. Moreover, Embedl’s technology is relevant in light of the consistently growing demand for AI solutions. Edge AI enables innovative capabilities to be used on local equipment. This increases the versatility of AI products and reduces dependence on cloud systems and data centers.
Embedl co-founder Hans Salomonsson emphasized the importance of improving the energy efficiency of AI. The startup’s technology enables intelligent products to be used on a large scale. At the same time, it will not require an increasing amount of resources.
Overview of the developer’s activities
Founded in 2022, the Swedish startup Embedl bases its operations on the research of Devdatta Dubhashi, who studied data usage issues. The company offers two products: The Model Optimization SDK and Embedl Hub. These products aim to optimize models to reduce energy consumption. As stated by Salomonsson, Embedl’s model can save up to 83% of energy in real-world implementation. Additionally, memory usage is reduced by 95%.
The startup’s products are relevant for consumers with limited business opportunities. We are not just talking about the financial component. For example, consider a project that lacks access to the cloud or sufficient electricity.