Jeffrey Katzenberg, former DreamWorks CEO turned VC, insists AI won’t be a zero-sum game — but only a handful of companies will emerge as true winners, a stance backed by his firm’s people-first evaluation criteria and challenged by experts questioning flawed benchmarks.
In an episode of the “Sourcery” podcast, Jeffrey Katzenberg, the former DreamWorks CEO and now a venture capitalist through his firm WndrCo, made a bold claim about the future of artificial intelligence: it will not be a zero-sum game where only one company wins at the expense of all others.
“Rather than look at it from the sort of 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,” Katzenberg said when asked about whether the AI bubble could pop next year.
He added: “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’s perspective comes with decades of experience navigating Hollywood’s creative economy — he served as chairman of Walt Disney Studios for 10 years until 1994 and cofounded DreamWorks after his departure. As DreamWorks’ CEO, he oversaw hits like “Shrek,” “Madagascar,” and “Kung Fu Panda.” He stepped down in 2016 and cofounded WndrCo with former Dropbox CFO Sujay Jaswa in 2017.
WndrCo’s general partner ChenLi Wang joined Katzenberg on the podcast to elaborate on their investment philosophy. “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 emphasized that WndrCo doesn’t rely on standardized benchmarks to evaluate startups. “I mean, how many years have people have parents complained about standardized testing dumbing down their kids,” Wang said. “And yet, I think we’re going down the same route with the first wave of benchmarks.”
This human-centric approach contrasts sharply with current industry trends focused on metrics like token accuracy, reasoning scores, and LLM benchmarks — metrics increasingly criticized by researchers for overemphasizing superficial performance traits.
Dean Valentine, cofounder and CEO of AI security startup ZeroPath, wrote in a March blog post that “recent AI model progress feels mostly like bullshit.” Valentine and his team evaluated models since Anthropic’s 3.5 Sonnet release in June 2024, finding none significantly improved their internal benchmarks or bug detection capabilities.
Researchers at the European Commission’s Joint Research Center published a February paper titled “Can we trust AI Benchmarks?” The study concluded benchmarks “often prioritize state-of-the-art performance at the expense of broader societal concerns.”
The tension between Katzenberg’s vision and the current tech ecosystem reflects a deeper philosophical divide: one side believes AI’s value lies in measurable output and efficiency; the other sees its true potential in human creativity and collaboration.
WndrCo’s portfolio includes notable AI startups such as Cursor, Harvey, and Figma — companies whose success may hinge less on raw technical superiority and more on their ability to integrate into workflows and empower users creatively.
For developers, this signals a shift toward evaluating AI tools based on practical deployment outcomes rather than theoretical benchmarks. For users, it suggests that AI’s most valuable applications may come from platforms that prioritize human interaction and workflow integration over pure computational prowess.
Katzenberg’s warning that “not everybody is going to win at this” echoes a broader sentiment among investors and technologists alike — that while AI will reshape industries, only those with unique value propositions and strong execution will survive.
As 2026 approaches, the reckoning Katzenberg predicts may come not from market crashes or funding dry-ups, but from a quiet filtering process — where companies fail to deliver tangible results, regardless of hype or benchmark scores.
For developers building AI tools, the message is clear: focus on real-world impact, not just benchmark rankings. For users, expect fewer “best-in-class” products and more specialized tools tailored to specific workflows — tools that succeed not because they’re perfect, but because they’re useful.
Onlytrustedinfo.com delivers authoritative, immediate analysis of breaking tech news — no links, no referrals, just facts and context you can trust.
Stay ahead of the curve with our daily coverage of AI, startups, and technology trends — the fastest, most insightful analysis available anywhere.
