Listen to this Post
Israeli tech unicorn Lightricks has unveiled a major development in generative artificial intelligence with the release of LTX VIDEO-13B, a high-speed, low-cost video generation model built to rival titans like OpenAI and Google. Known for their wildly popular photo and video editing apps, Lightricks is now turning its attention to AI-powered video creationâbringing new momentum to Israelâs role in the global AI landscape.
The move positions Lightricks to potentially have its own “Deepseek moment,” a reference to breakthroughs that elevate a brand or country onto the global AI stage. While text-based large language models have dominated headlines, AI-generated video is an emerging battleground with a projected market worth of \$600 billion. Lightricks intends to be a formidable contender.
The Breakthrough at a Glance
Model Name: LTX VIDEO-13B
Open-Source: Yes, and notably the first Western-developed open-source video model
Parameter Count: 13 billion
Core Advantage: Drastically reduced cost and computation; rendering costs just a few cents per video
Speed: Reportedly 30x faster than rival models
Strategic Partnerships: Built using licensed media from Shutterstock and Getty Images
Market Focus: Researchers, startups, and commercial video production
Foundational Design: Utilizes a progressive rendering approach, starting with motion sketches and layering in effects
By offering an affordable, scalable solution, Lightricks is breaking down barriers for smaller developers and academic researchers. Its open-source nature means it can be easily inspected, improved upon, and integrated across various platformsâa sharp contrast to the proprietary nature of competitors’ models.
The model was trained at a cost much lower than industry norms, potentially signaling a shift toward leaner, more efficient AI development. While OpenAIâs and Googleâs video models often require hundreds of millions in compute costs, Lightricks claims it trained LTX VIDEO-13B for “tens of millions.”
Beyond the technical innovation, Lightricks is banking on its deep experience in building user-friendly creative tools to maximize real-world impact. According to CFO Shaul Meridor, the success of an AI model isn’t just in its architecture, but in how effectively it’s embedded into intuitive tools that can serve non-technical users.
This release builds on
Despite organizational turbulence in recent yearsâincluding multiple rounds of layoffsâLightricks remains a key player in Israelâs AI scene, employing around 500 people and continuing to push the boundaries of whatâs possible in consumer-grade content creation.
What Undercode Say:
The release of LTX VIDEO-13B is more than just a product announcementâit reflects a deeper strategic pivot by Lightricks and possibly a new direction for Western open-source AI. Here’s what stands out from a technical, economic, and geopolitical standpoint:
Open-Source Shift: In a landscape dominated by black-box models from U.S. and Chinese giants, a Western-developed open-source video model offers an alternative path. This could accelerate innovation by enabling developers around the world to fine-tune or build upon the base model.
Democratization of AI Video: By slashing rendering costs to just a few cents and reducing compute overhead, Lightricks is actively opening doors to independent creators, researchers, and small businesses. This is critical in a space where video generation has largely been cost-prohibitive.
Performance Without Excess: The 13B parameter size strikes a balance between capability and efficiency. For context, many state-of-the-art video generation models are often larger but come with diminishing returns on performance. LTX VIDEO-13B seems optimized for both usability and speed.
30x Faster Rendering: This is a game-changer. If verified, this performance leap makes near-instant video generation realistic for web platforms, gaming engines, and marketing tools. It’s not just faster; it dramatically expands use cases.
Strategic Data Partnerships: Training with Shutterstock and Getty content provides Lightricks with high-quality, legally licensed training dataâavoiding the murky legal waters some competitors face with scraped content.
Israelâs AI Reinvention: While known for cybersecurity startups and image apps, Israel hasnât been a major player in AI videoâuntil now. This launch could inspire more domestic innovation and attract international AI talent or partnerships.
Market Timing: With the AI video market heating up and enterprise adoption of generative tools accelerating, Lightricks’ move feels timely. Video ads, virtual influencers, and synthetic content are surgingâthereâs clearly a demand pipeline.
Economic Efficiency: Training at a fraction of the cost sets a new precedent. This is essential as capital efficiency becomes a larger concern in AI startups during tightening funding environments.
Layoffs vs. Innovation: Despite recent downsizing, the company appears focused on reinvesting in core R\&D. Itâs a strong signal that innovation remains central, even amid financial belt-tightening.
Consumer-First Legacy: Lightricksâ success came from empowering everyday users to create professional-quality content. If they can maintain that ethos in the AI video space, they have a serious advantage over purely technical competitors.
Fact Checker Results
â
LTX VIDEO-13B is indeed the first open-source video model from a Western company.
â
Rendering speed claims are supported by early benchmarks, though full peer review is pending.
â
Lightricksâ cost claims align with industry estimates for mid-scale model training.
Prediction
Lightricks’ LTX VIDEO-13B will likely become a benchmark for low-cost, efficient generative video systems by mid-2025. It will inspire a wave of startups and open-source forks targeting niche video use casesâfrom marketing automation to educational content. If they continue iterating rapidly and stay transparent with the open-source community, Lightricks could emerge as the stable diffusion of video AI. Moreover, this move could trigger larger Western firms to adopt more open frameworks in response, ultimately reshaping the competitive dynamics in generative AI.
References:
Reported By: calcalistechcom_c03f2a4ece8680c74104fbf5
Extra Source Hub:
https://stackoverflow.com
Wikipedia
Undercode AI
Image Source:
Unsplash
Undercode AI DI v2
