Amnon Shashua, founder and CEO of Mobileye, believes that complex problems can be solved with AI, and that AI itself can be modified to make it more reliable. In addition to founding and running a self-driving car technology company that he took public, sold to Intel, and then spun out again, he has many other ideas.
One of them is currently raising money and gaining significant momentum.
One Zero, a fintech company aiming to leverage AI in retail banking services, is raising at least $100 million, TechCrunch has learned.
Despite being co-founded by one of Israel's most high-profile and successful founders, One Zero has so far received surprisingly little attention outside of its domestic market. But the company has raised about $242 million to date and was valued at $320 million in 2023, per PitchBook data. Our sources say the valuation will be significantly higher in the next round.
It's unclear who the investors are, but the company's previous backers include Tencent, OurCrowd, and SBI Ventures (formerly part of SoftBank, now an independent company).
One Zero's momentum comes amid a frenzy of activity from Shashua, a non-executive director at the company whose CEO is Gul Bar Deer. In recent years, Shashua has founded or co-founded a startup working on humanoid robotics (Mentee). Alternative approaches to large-scale language models for generative AI (AI21). And AA-I Technologies (pronounced “double AI”), launched just a few weeks ago, is what Shashua describes as an effort to develop “AI scientists.” He is also a professor of computer science at the Hebrew University of Jerusalem.
One Zero's equally ambitious mission is to “bring private banking to the masses,” he said in an interview. It is only natural that the high-touch, advice-type service that wealthy people receive when using banks should be democratized in a market where ordinary people today not only cannot receive such services, but are unable to receive them. is. We are looking at a future where there may be no physical banks and no humans to help us at all.
The company is pursuing its ambitions with a dual business focus. In Israel, where OneZero is based, the startup has obtained a banking license and is building a full-stack retail bank. In addition, One Zero is using insights from its retail business (which Shashua described as a “sandbox” in an interview) to train its models, hone its technology, and make it available to others. We are working to license banks to operate in these locations.
The retail business currently has about 110,000 customers, Shashua told TechCrunch, and the company has received a number of licensing deals from major banks, although so far the company has not announced any licensing deals. It is said that there is
The company's foundation to date, and the focus of where it's putting its money, is a chatbot called Ella that provides services that human bankers can't, but that are better than current chatbots. We are aiming for
In Shashua's view, there are many efforts to incorporate AI into retail banking services (features such as spend management), but there is only so much that can be done.
“I haven't seen any banks implement artificial intelligence to the point where they can actually replace bankers,” he said.
Take automated communication as an example. If you ask a bank chatbot very basic questions, such as “How much money do I have in my account?” or information about recent transactions, they will usually answer. But whether it's a question that involves calculation, such as “Based on my activity so far, how much money will I have in my savings account at the end of the year?” or “What is the best purchase option for me?” It's a different story. Is it a car based on my financial profile? ” Not only can chatbots not be able to answer such questions, neither can most personal bankers.
“There is an opportunity here and it looks like generative AI can make this happen,” he said. “It goes far beyond tracking spending.”
As Shashua explained, One Zero's approach to building this kind of AI is so ambitious that it feels as difficult as self-driving. Focuses on using multiple large-scale language models. He said that while some models may be optimized for different tasks, running a task through multiple LLMs will provide different responses, run through a validation process, and provide an answer. You can understand what is misleading or wrong.
And if those answers aren't verified to be useful or correct, the AI won't say anything anyway, he said. “Are you okay [for it to] Please tell me I can't solve your problem. “I can’t answer your question,” he said. “Humans can't answer every question either, right? And that's okay. It's not good to say, 'Here's the answer to your question, and that answer is completely fake, completely wrong.' ”
The system is starting with more basic tasks like spending management, and plans to add more features over time to help advise customers on financing big purchases and smarter savings.