Zeev Farbman, co-founder and CEO of Lightricks, believes we should shift our attention from building AI infrastructure (e.g., trillions of dollars spent on data centers) to how we will use it. The true winners, says Farbman, will be those who apply AI to profitable use cases, focusing on business applications that drive long-term value rather than short-term excitement.

Founded in 2013, Lightricks has successfully ridden the mobile explosion with its image and video editing apps, such as the widely popular Facetune, becoming one of Israel’s early unicorns. The company’s roots were in the disciplines of computer graphics and machine learning, and as Farbman dryly notes, in recent years, “all machine learning technologies have been called AI.” But he also notes the “paradigm shift” with generative AI and the introduction of ChatGPT, transforming “how we create pixels: instead of explicitly modeling things, we are using examples.”

With Lightricks’ latest LTX-2 model, marketing and advertising professionals can create a six-second video in as little as 5 seconds, at one-tenth the cost of competing models (it is designed to run on consumer-grade GPUs). Lightricks further differentiates its approach by allowing users to iterate on their content development and tinker with it until they are satisfied with the final product, and by open-sourcing the model, enabling developers to experiment with it and share their enhancements with the broader user community.

Lightricks is also part of the Israeli AI ecosystem, which last year began focusing on sector-specific AI solutions, leveraging industry-specific data and expertise. According to Startup Nation Central, funding for AI-focused startups is expected to rise from $4.9 billion in 2024 to approximately $7.9 billion in 2025. “Israeli entrepreneurs are successfully combining new opportunities created by AI with the background and experience of Israel’s cybersecurity ecosystem,” observed a recent report from RISE and IVC, noting “the ability of Israeli entrepreneurs to apply AI technologies across diverse domains.”