Julie Zhuo, former vice president of design at Facebook and co-founder of AI analytics startup Sundial, warns that many of the hottest AI startups are growing rapidly on "good vibes" and instinct rather than solid data foundations. She cautions that without proper data infrastructure, these companies may struggle when growth inevitably slows. Zhuo points out that some AI startups are scaling to hundreds of millions in annual recurring revenue and millions of users with just a handful of employees—often only 10 or fewer. Despite this explosive growth, many lack essential systems like logging, analytics, and observability, which are critical for understanding user behavior, product performance, and business drivers. In the past, companies took years to reach massive user bases, giving them time to build robust data teams and infrastructure. Today’s AI startups, however, are growing too fast for that luxury. As a result, they’re operating in the dark, unable to answer basic questions about why users leave, which features matter most, or what truly moves the needle. Zhuo predicts that when growth plateaus, these companies will be forced to scramble to catch up. That’s when they’ll begin investing heavily in data—but by then, it may be too late to recover lost momentum. She also highlights a shift in how success should be measured in the AI era. Traditional metrics like clicks or page views don’t apply to conversational AI products like chatbots. Instead, companies need to use machine learning models—such as large language models—to analyze user intent and behavior in more nuanced ways. The trend reflects a broader industry phenomenon. In 2024 alone, over $35 billion was raised by AI startups, fueling a wave of high valuations and rapid scaling. But concerns are mounting that the market is overheating. Some investors fear a repeat of the dot-com bubble, with overvalued startups and unrealistic expectations. OpenAI CEO Sam Altman echoed these concerns, calling the current funding frenzy "insane" and "not rational," while acknowledging that AI remains one of the most transformative forces in tech history. The tension between excitement and reality is growing, and Zhuo’s warning underscores a key lesson: even in the age of AI hype, data remains the ultimate reality check.