Jeffrey Katzenberg, former CEO of DreamWorks and co-founder of venture capital firm WndrCo, says the AI revolution won’t be a zero-sum game where one company dominates all. Speaking on the "Sourcery" podcast released Monday, Katzenberg argued that while 2026 will likely bring a reckoning for AI companies, it won’t be a winner-take-all scenario. Instead, he expects a clear distinction to emerge between firms delivering real-world impact and those merely chasing hype. “Rather than look at it from the extreme notion of what that means for a bubble to burst, I think there’ll be a reckoning here in which those that actually are producing real results and are being deployed in really effective and efficient ways,” he said, addressing concerns about an AI bubble. “It’s not going to be a zero-sum game, a winner-take-all. But I also think at the same time, not everybody is going to win at this.” Katzenberg, who led Walt Disney Studios for a decade before co-founding DreamWorks in 1994, oversaw the creation of iconic franchises like “Shrek,” “Madagascar,” and “Kung Fu Panda.” After stepping down in 2016, he co-founded WndrCo in 2017 with former Dropbox CFO Sujay Jaswa. The firm has backed AI-driven startups including Cursor, Harvey, and Figma. On the podcast, WndrCo’s general partner ChenLi Wang elaborated on how the firm evaluates startups in the AI era. He emphasized that both he and Katzenberg have always prioritized human creativity and collaboration. “First, I think through our entire careers pre-WndrCo, Jeffrey’s entire career, we’ve been people people,” Wang said. “The ingenuity and creativity of people and how magical their spikes are, and how, when you complement people and what they can create together, is the secret sauce.” Wang criticized the growing reliance on benchmarks to measure AI progress. “We’ve never assessed our best humans based on benchmarks,” he said. “How many years have people had parents complain about standardized testing dumbing down their kids? And yet, I think we’re going down the same route with the first wave of benchmarks.” This skepticism echoes broader concerns among AI researchers. In a March blog post, Dean Valentine, CEO of AI security startup ZeroPath, dismissed recent model advancements as largely “bullshit.” He noted that despite numerous claims of improvement since Anthropic’s Sonnet 3.5 launch in June 2024, none of the models his team tested delivered meaningful gains in internal benchmarks or in developers’ ability to uncover new vulnerabilities. Similarly, a February paper from the European Commission’s Joint Research Center titled “Can we trust AI Benchmarks?” raised alarms about the current evaluation system. The report found that benchmarks often prioritize top-tier performance while overlooking critical societal issues like bias, transparency, and long-term safety.