Human investors who have tried to replicate Warren Buffett have mostly failed miserably, but now artificial intelligence is taking a crack at it. Investment startup Intelligent Alpha is working on launching the Intelligent Omaha ETF , with the ticker AIWB, that relies on AI to emulate the Berkshire Hathaway CEO’s investment philosophy and create a portfolio of 25-30 stocks that it thinks Buffett might buy. So, how does it work exactly? The first step is to feed AI — three large language models GPT, Gemini and Claude — enough information for it to learn the essence of Buffett. That includes any publicly available writings (Buffett’s famous annual letters since the 1960s), any public interviews or statements he’s made, as well as regulatory filings like the 13Fs that reveal Berkshire’s public equity holdings. “If you give AI a bunch of information about how you want it to think and how you want to respond and think about the world, it will play that role perfectly in our testing,” Doug Clinton, Intelligent Alpha CEO and founder, said in an interview. It remains to be seen how deeply AI could understand Buffett’s unique investing style that has evolved significantly over seven decades. He started out only buying “cigar-butt,” dirt-cheap securities for quick profit and later, under the late Charlie Munger’s influence, focused on quality companies with “wide moats” selling at fair prices. Now that Berkshire’s market value approaches $1 trillion, fewer and fewer investments move the needle for the conglomerate, which partly drove Buffett to go all in on Apple. Not to mention, Buffett’s reputation as one of the most shrewd and cash-rich investors enabled him to score many lucrative deals through personal negotiations, especially during crises . “The Omaha strategy is very much built around value in the context of Buffett, trying to find companies that have long duration,” Clinton said. All about AI Intelligent Alpha has been testing its Buffett strategy for six months and the portfolio AI puts together would overlap Berkshire’s by anywhere between 30 to 60%. The same stocks might have a different weighting in the ETF, which would charge an expense ratio of 69 basis points. The fund is set to debut sometime in the next six months. The firm already has an AI-powered ETF on the market — the Intelligent Livermore ETF , which aims to imitate a dozen prominent investors’ approaches, including Buffett, Stanley Druckenmiller and David Tepper. That fund started trading last Wednesday. These offerings mark Wall Street’s latest attempt to incorporate buzzy AI applications in products to attract investors curious about what the advanced technology can produce. AI has been the hottest investment theme for the past two years, with stocks like Nvidia being the biggest beneficiaries, but many remain skeptical of its lasting power and impact on everyday life. “The vision for Intelligent Alpha is to build the AI-powered BlackRock,” Clinton said. “There are two big pillars, serving retail investors [through Registered Investment Advisors], and then there’s serving institutional investors.”