Beamr's CEO, Sharon Carmel, participated in an expert call with Loop Capital IT Infrastructure & Supply Chain analyst Ananda Baruah.

The call, held on Monday, October 20 2025, covered the impact of Gen AI Video on Data Storage, Networking and Application requirements.

For an audio replay of the call, please click the link

In the expert call, Carmel explored innovation, challenges and opportunities in fast growing video markets, AI and Gen AI and more.

"For autonomous vehicles, there are estimations that the video amount is doubling by 2X every year, and this is a subsection of machine vision", explained Carmel. "The scale today is hundreds of petabytes per AV company, and this may grow to hexabytes - which is the size of a social network. These companies need a lot of storage, thousands of GPUs that will grow to the order of tens of thousands of GPUs, and massive investments. 

"This will lead to the need to maximize the infrastructure, get more out of GPUs and optimize algorithms. Also, using compression for videos, so they’ll be half the size but serve the technology equally well", he added. (to learn more about Beamr's AV solution visit beamr.com/autonomous). 

Carmel discussed how AI is changing the Media & Entertainment industry: "This market has reached maturity, has stable workflows, and is now looking for opportunities with AI that will grow the business. 

"For that, you need mature technologies in the pipeline. For example, you can’t get Super resolution which hallucinates, and may distort a football player's face or change the scoreboard. Only when you get to a phase when technology is reliable enough, and doesn’t hallucinate and create details that shouldn’t be there - only then it is ready for production. Today it's ready, but a year ago it wasn’t.

"Live has the most opportunities in terms of value, because live content, like sports, is the most valuable content." Recently, Beamr's Live 4K offering won the best of show award at the IBC show.

Carmel also shared his vision about the transformation of video itself from an unstructured to a structured data type, enabling search and other innovative capabilities, and the massive opportunities it brings: "One of the areas I’m most excited about is structured and unstructured data. About 80% of data is unstructured, but that is changing now, because video is turning towards being structured data. With AI you can actually know what’s in the video, what objects are, identify characters or make the video searchable.

"You can identify if this is a sad or happy scene, and if it’s a surveillance camera, you can identify suspicious behavior - and all this is done automatically. The video becomes much more usable. If you have a large video repository, you can monetize it because you can now extract value from it using AI search capabilities."