Deep Render, which is creating artificial intelligence-powered technology to compress online movies, has just secured a $9 million Series A. This round is headed by IP Group and Pentech Ventures. Despite the recent market volatility, sources reveal that the round values the business at $30 million.
The Purpose of The Funding
The additional funding, which does not include a new $2.7 million investment from the European Innovation Council, will go towards product R&D. Moreover, it will also refocus Deep Render’s customer acquisition efforts on the U.S.
According to Besenbruch’s email, “the funding is motivated by Deep Render reaching its internal objectives and creating an inflection moment.” Our R&D and productization efforts as a super-dense “hard tech” firm are complete.
More About Deep Render
In 2018, Besenbruch and Arsalan Zafar created Deep Render after meeting with computer science, ML, and AI students at Imperial College London. Besenbruch and Zafar were inspired by a research project. In this project, they participated and were tasked with transmitting many terabytes of video data over a network. Technical issues with network congestion plagued the project many times. This prompted Besenbruch and Zafar to look into a different and, hopefully, superior approach.
To improve compression dramatically for images and videos, “we chose to integrate machine learning, applications powered by artificial intelligence, and compression technologies to produce a completely new technique of compressing data,” Besenbruch said. Deep Render was founded to eliminate bandwidth constraints worldwide by using artificial intelligence (AI) to replace conventional compression methods.
How Does Deep Render Work?
DeepMind at Alphabet repurposed an AI system it had developed to play board games to reduce the size of uploaded movies to YouTube. Nvidia, Disney Research, and the University of California, Irvine, have done separate tests on AI-driven compression methods for streaming video.
Besenbruch says that the over ten million video sequences used to train Deep Render’s AI compression technology set it apart. The business used both on-premise and cloud infrastructure for the training, with the former numbering over 100 GPUs. Furthermore, Besenbruch asserts that the algorithm can operate in real-time on contemporary chipsets from Qualcomm, Apple, and Nvidia and is up to five times “better” than HVEC, the standard industry codec (but did not indicate by what criteria).
According to Besenbruch, “finding a new route ahead and seeking new technologies is a huge problem in the compression sector.” According to the company’s website: “Deep Render is resetting the industry, benefitting its clients. Conventional compression is not progressing rapidly enough; it has only seen incremental growth for decades and has passed its pinnacle.”
The Rise of Video Streaming
Whether or whether that’s the case, Cisco predicts that by 2022, 82% of all IP traffic will be made up of streaming videos. Besenbruch argues that improved compression is a worthwhile investment for streaming video platforms since it enables these services to improve streaming quality without incurring additional costs connected to bandwidth use.
“If the internet pipelines are hard to stretch, all we can do is reduce the data that passes through them,” Besenbruch added.
For more latest IT news and updates, keep reading iTMunch
Image Courtesy: Image by DeepRender