Inside Zyg: The IronSource Alumni Building the Ultimate AI E-Commerce Publisher
Israeli startup Zyg, founded by ironSource alumni, has raised massive funding to scale e-commerce brands globally using AI agents and the publisher model.
Dr. Yehoshua Maor, an ultra-Orthodox Israeli born in Brazil and a pharmacology expert, has spent his career harnessing wild plants for pharmaceutical production. During his doctoral studies at the Hebrew University, he developed a method to lower blood pressure using a compound derived from the cannabis plant, as well as an anti-psoriasis ointment based on myrrh and frankincense. After completing his post-doctoral studies at Harvard Medical School, he traveled to Brazil with the aim of developing an anti-inflammatory substance based on plants unique to his native country. However, he soon discovered that the substance not only prevented inflammation but also prevented hair loss and restored forgotten hairlines.
Without intending to, Maor realized that the substance slows down hair loss and promotes hair regrowth in thinning areas. This clearly held the potential for a product that would sell well globally: who wouldn't want thick, flowing hair like in their youth? Maor teamed up with entrepreneurs Monica Hodges and Ziv Mendelson, and together they developed the final product—Phyllotex—backed by research conducted in collaboration with expert dermatologists. If you are wondering, this story is not about herbal ointments and capsules. It is a story about the company helping such products reach the global market—leveraging the absolute frontier of technology: artificial intelligence.
The E-Commerce Publishing Challenge
The central problem known to many entrepreneurs, such as Maor and his partners, is reaching all potential customers worldwide and executing direct-to-consumer (DTC) sales instead of selling out to international e-commerce giants. Today, online platforms like Amazon or Shopify exist, but that is not enough. Phyllotex is already sold via Amazon, but it has not broken through to become a leading product in its category. This is where the Israeli company Zyg enters the picture.
This year, Zyg was revealed and gained fame thanks to its massive fundraising rounds, which typically signal cyber or AI startups. But Zyg is engaged in one of the world's oldest professions—sales—adapted for the new era. Zyg rebranded the company under the name The Root, built the creative assets, improved the sales funnel, generated financial models for budgets, and, using AI agents, paved the way for the breakthrough of lush hair to the global market. According to Zyg's data, within three months, The Root's revenue run rate grew fourfold compared to the period it relied solely on Amazon, new customer acquisition rates surged 20% above the category average, and 70% of customers chose a subscription model guaranteeing recurring revenue.
The Publisher Model for Physical Products
The story of The Root illustrates Zyg's problem and solution: much like in the worlds of music or video games, creators (or developers) are terrible at reaching the market. For this, they need a publisher (such as Zyg in this case) that will market online, distribute across networks, build trust programs, sell and collect payments, and ultimately enjoy a small slice of the revenue.
Zyg was founded in 2025 but remained under the radar until March of this year, when it reported raising a $58 million seed round. Two months later, it reported an additional $60 million fundraising round. Among the investors are Viola Ventures, Bessemer Venture Partners, Accel, Felix Capital, Lightspeed, and Asaf Rappaport, alongside other angel investors. Investors are clearly enthusiastic—they valued the company at $500 million already at the Series A round.
All of this would make sense if it were a cyber company protecting against AI attacks, or an AI company developing penetration defenses. But Zyg develops technology for e-commerce. Really? In an era where AI sells better than any platform? One of the reasons that thrilled investors is undoubtedly the identity of Zyg's eight founders, most of whom are founders or senior executives of the legendary ironSource: Omer Kaplan, co-founder of ironSource who serves as CEO of Zyg; Tomer Bar-Zeev, co-founder and CEO of ironSource, serving as Chairman of the Board at Zyg; Asaf Ben Ami, serving as CFO; Dr. Eyal Amit, a graduate of Unit 81 serving as CTO; Nadav Ashkenazi, former head of ironSource's Israeli branch; Daniel Schnear, former senior executive at ClalTech; and Omri Steinmetz and Guy Tzur, both graduates of Unit 81. Entrepreneurs who "have done it" are extremely attractive to investors. And ironSource's founders certainly did it.
"We identified a huge need in the e-commerce worlds where even if a company has developed a great product with the potential to sell tens of millions of dollars a year, they still lack a partner who can provide them with that growth," says Kaplan.
The ZyG Scale Score and Agentic OS
At the first stage, the company evaluates the new product using a tool called the PMF Test, which assesses the potential market size and the product's attractiveness: an interesting story, scientific backing, registered patents, and more. In the second stage, validation is performed—generating a real campaign and testing how the market responds. At the end of the test, each product receives a score: the ZyG Scale Score.
At this stage, the Agentic OS system developed by Zyg handles the creation of a full sales campaign using AI agents: establishing a sales website and landing pages, building creatives including social networks and influencers, producing video and other content, running experiments, collecting massive amounts of data, and building predictive financial models. This is where Zyg's competitive advantage lies: in a traditional company, all of this would require a vast number of employees. At Zyg, AI does a significant portion of the work.
When asked about the defensive moat against general-purpose AI models like Claude, Kaplan emphasizes: "You don't need to care what I build; I sell an outcome. I tell the company: 'I deploy people where needed, agents where needed, data infrastructure where needed.' I call it 'Outcome as a Service.' I don't sell software, I sell a service. And Claude doesn't know how to provide that service. Only when you do that do you have a moat."
