The Israeli startup Run:AI, which is developing virtualization technology for GPUs that saves time in artificial intelligence training, announced yesterday (March 15) the completion of a $75 million Series C round C. The round was led by two of the largest funds in the Israeli venture capital - Tiger Global and Insight Partners. TLV Partners and S Capital VC also participated in the round.

Working hard on all your GPUs

Run:AI solves a significant problem for many organizations who have in recent years, begun to integrate AI applications into their development processes. Those applications have difficulty working properly due to the absence of sufficient training for artificial intelligence models - such training is a long, exhausting, and extremely expensive process.

The Run:AI platform makes it possible to bind all the existing GPUs in the organization and allocate them dynamically and automatically as needed. This is done through virtualization and orchestration. Thus, multiple tasks can be run on a single GPU, instead of waiting and doing nothing between each test.

In addition, the company has developed resource management software for AI tasks that use containers and Kubernetes and can allow several containers to run simultaneously on the same graphics card, without "interfering" with each other - which should significantly improve the resource efficiency of companies interested in training their models with Run:AI.

In a conversation with Geektime, Omri Geller, co-founder and CEO of Run:AI, explains that the platform they developed connects to Cloud-Native environments based on containers and Kubernetes and works on several levels in Stack. "At the lowest level, we enable more efficient abstraction and sharing of GPUs and artificial intelligence computing resources using API-level interception technology so that multiple containers can be run simultaneously on a single GPU."

In addition, the platform enables the operation of the works and the sharing of resources most efficiently and simply, while considering the needs and priorities determined by each organization. Above all, there is a control system that enables monitoring and control of all AI environments, including models for developing and training models and environments where trained models are embedded in production.

It does not happen every day that both Insight and Tiger invest together in companies in Israel, so we asked Geller how it unfolded behind the scenes. According to Geller, since Insight's previous investment in the January 2021 Series B round, the growth that the startup has shown has led Tiger Global to take an interest. "Insight, which led the previous round, chose to lead this round as well and after consulting with the company's board of directors, we chose to add a strong investor to co-lead the round in collaboration with Insight, to expand the company's capabilities by adding stakeholders."

Run:AI was founded in 2018 by Omri Geller, the company's CEO, and Dr. Ronen Dar (CTO), a former researcher at Bell Labs University and an algorithm engineer at Anobit, which was sold to Apple. The Israeli startup currently employs 70 people in its offices in Israel and the United States.