Former OpenAI Executive Questions Whether Software Valuations, Not AI, Were the Real Bubble
Zack Kass, a former head of Go-to-Market at OpenAI who now serves as a global AI adviser, suggested Tuesday that traditional software stocks—rather than artificial intelligence companies—may have been trading in bubble territory, marking a notable reversal of the prevailing narrative around technology valuations.
Speaking on Bloomberg Television's "The Close," Kass argued that the market's recent turbulence might reflect an overvaluation of legacy software companies rather than the AI sector that has dominated investor concerns. The comment comes as finance leaders grapple with how to value companies in an industry undergoing rapid technological disruption.
For CFOs navigating technology spending decisions, the distinction matters considerably. If Kass is correct, the financial risk may lie not in cutting-edge AI investments but in maintaining commitments to traditional software infrastructure that could face margin compression as AI-native alternatives emerge. The thesis suggests that enterprise software companies trading at premium multiples—often justified by recurring revenue models and high switching costs—may be vulnerable as AI tools offer cheaper, more flexible alternatives.
Kass, who helped shape OpenAI's commercial strategy before the company's explosive growth with ChatGPT, also addressed recent optimism from Nvidia CEO Jensen Huang about the rapid development of agentic AI applications. These autonomous AI systems, which can complete complex tasks with minimal human oversight, represent the next frontier in enterprise automation—and a potential accelerant for the displacement of traditional software.
The former OpenAI executive's perspective carries weight given his front-row seat to the AI boom's early stages. His current role as an independent adviser allows him to comment more freely on market dynamics than executives still tied to specific companies. The fact that he's questioning software valuations rather than defending AI's lofty expectations suggests a level of confidence in AI's fundamental economics that many finance leaders still lack.
The timing of Kass's comments is particularly relevant as companies finalize 2026 technology budgets. Many CFOs have been treating AI spending as experimental—ring-fenced innovation budgets separate from core systems. But if traditional software is overvalued and AI tools are underpriced relative to their productivity gains, that budgeting framework may need to flip.
The challenge for finance leaders is that this transition won't be clean. Legacy software contracts often run multiple years, and the switching costs Kass's thesis implicitly questions—implementation time, training, data migration—remain real even if they're not insurmountable. The question isn't whether AI will disrupt software, but how quickly, and whether current software valuations have priced in that disruption adequately.
What remains unclear from Kass's brief remarks is which specific software categories he views as most vulnerable, or what timeline he envisions for this revaluation. For now, CFOs are left to parse whether their enterprise resource planning, customer relationship management, or human capital management systems might be the legacy software at risk—and whether their AI experiments might be undervalued insurance policies rather than speculative bets.
Responses (0 )