A week in AI is like a year in other industries. I hope these issues become your weekly source of AI information, inspiration, and ideas. If we haven’t met before, I’m Amanda Smith. I write about AI and the fascinating folks who are building in this brave new world.
Good morning. I love this week’s AI founder story so much.
It’s his first start-up and he’s competing with the biggest tech players in Silicon Valley… while being based on the campus he went to college in Jerusalem, Israel.
He built the model with a fraction of the resources, with a team of less than 50 in five months.
Enjoy this read.
🤝 This edition is kindly brought to you by Delve
Ever had to fill out a gnarly security questionnaire inside a vendor’s portal and hated your life?
Delve just launched an AI browser agent that does it for you.
It finds answers from your policies, past questionnaires, and docs, then fills everything in while you click next. Hours → minutes.Switch to Delve for compliance and use this - and every other gamechanging AI features at one simple cost.
If you’re using another platform for compliance, Delve will migrate you over or free.
AI Odyssey readers get $1,000 off compliance using the link below.
Set it and forget it >>> sweating questionnaires.
This week in AI:
Data center energy demand to grow 300% by 2035
The AI sector needs 3x more electricity over the next decade
OpenAI declares “code red”
Sam Altman told employees that the company was declaring a “code red” to improve the quality of ChatGPT.
Australia rolls out AI roadmap
The country will ramp up AI while relying on existing laws.
Company background: Lightricks
Founded: January 2013
Team size: 470 employees
Funding to date: $335 million
ARR: More than $250 million
Growth metric: Lightricks’ open-source strategy has built a community of 70,000+ members who help identify emerging needs, promote our work, and share real-world use cases.
How this first-time AI founder is doing $250M+ ARR
While studying computer science and image processing at the Hebrew University of Jerusalem, Zeev Farbman interned at Adobe and Microsoft. He got the start-up bug early in his career and went onto launch Lightricks before the end of his PhD.
This was in 2013, well before AI was big.
“At the time, there was a clear gap between academic research and real-world creative tools. We used machine learning early on because it allowed us to build high-quality mobile experiences. Our vision was to become the Adobe of mobile content creation,” said Farbman.
AI was never the goal on its own. Fast-forward a decade, the first-time CEO is now trailblazing the AI video race alongside AI giants, with a fraction of the resources. Lightricks’ LTX-2 model is ranked one of the best AI video products.
The build philosophy
“Early on, we knew that to compete with companies that had far larger budgets, we needed to make strong architectural bets. One of the most important decisions we made was to use a highly compressive latent space,” Farbman explained.
“This approach reduces the number of tokens needed to train and generate a video. It also introduces challenges, especially in preserving high-frequency details. We developed specialized techniques to overcome these challenges, resulting in a model that runs far more efficiently without compromising quality.”
They built their first AI products through a partnership with Stability AI, but once Sora was released, Farbman saw a need for a different model architecture.
“We realized that to offer a complete and reliable business solution, we couldn’t depend on other companies for foundational technology. Diffusion transformer models don’t really have niches. A strong foundational model can be adapted across many use cases.”
That insight led Lightricks to build their own full stack.
“We now develop the foundational model, the API layer, and the products that sit on top of them.”
While the landscape is more crowded now, companies like OpenAI and Google have shown the world that AI video has real commercial applications. They expand the category and increase awareness.
“While competitors rely on closed and expensive models that are difficult to embed, we provide an alternative: a frontier-grade multimodal model with open weights,” he said.
Lightricks is currently the only company in the West offering such a model.
There are four major multimodal models today:
Sora 2
Veo 3
Wan 2.5
Lightricks’ LTX-2
“Ours will be the first to be released with open weights. We believe the future will be shaped by open ecosystems rather than closed tech stacks.”
Customer base & business model
Lightricks began as a consumer-focused company with traditional B2C apps for a large global audience. Farbman said since 2022, especially since they acquired Popular Pays, they’ve shifted toward a B2B model.
Today, Lightricks offers multiple business-facing products and works with leading companies including Pepsi, Delta, and Kraft.
“At the product level, we offer subscription services as well as white glove and managed solutions. At the technology level, we provide API access to our models and have the ability to build or fine-tune unique model offerings for specific customer needs.”
Lightricks is structured into two divisions. One is a profitable mobile business, and the other is their AI division, which includes LTX 2 and related technologies.
Acquiring customers
Their go-to-market strategy has three pillars:
A large sales team across the U.S. and EMEA that engages directly with enterprise clients
Open-source model attracts ongoing interest from companies that later expand their usage through the API
Product platforms that bring in customers through marketing and self-serve channels.
“Our most effective strategy has always been demonstrating clear value to our customers. Every marketing campaign begins with two questions: who is the client, and what problem are we solving for them? This approach consistently drives results.”
“Our open-source strategy has also built a community of more than 70,000 members who help identify emerging needs, promote our work, and share real-world use cases.”
Takeaways
The fundamentals of building a company haven’t changed as much as people think. What has changed is the leverage AI provides. It’s an exciting time because the technology is far ahead of its real-world economic impact.
Scale is not the answer to everything. The opportunities ahead are enormous – larger than those created by the internet or mobile revolutions.
AI founders don’t have to be in Silicon Valley (or your country’s equivalent). Lightricks is based in Jerusalem. Most people associate Israeli tech with Tel Aviv. The difference is similar to San Francisco vs. Seattle. SF is energetic, while Seattle is quieter and better suited for deep work.
“We’re located on the Hebrew University campus and close to the leading design school, which gives us a steady pipeline of strong creative and technical talent.”
How's the depth of today's edition?
P.S. If you want to get a founder feature about your own story, reply to this email. If you’d like to reach our newsletter audience (founders, creators, and marketers), click the button below.
If you’re new here, I’m over the moon you’ve joined us! To help me craft content that’s actually useful (and not just noise in your inbox), I’d love it if you took 1 minute to answer this quick survey below. Your insights help shape everything I write.
✨ Insane Media is more than one voice
Dive into our other newsletters - where psychology meets the creator economy, e-commerce marketing, and Human resources.
