More than three years after the explosive rise of generative artificial intelligence, the majority of startups in the AI sector continue to derive most of their revenue from enterprise clients rather than from direct consumer sales. Despite the rapid and enthusiastic adoption of general-purpose large language models such as ChatGPT by the public, specialized generative AI applications designed specifically for consumers have generally failed to capture mainstream traction or achieve sustainable engagement.
At TechCrunch’s StrictlyVC event in early December, Chi-Hua Chien—co-founder and managing partner at Goodwater Capital—reflected on these early days of consumer-focused AI innovation. He acknowledged that many of the first AI tools for generating video, audio, and photos initially appeared groundbreaking and dazzled early adopters with their capabilities. Yet as new entrants like Sora and Nano Banana emerged, and as Chinese research groups made their advanced video-generation models open source, the competitive landscape shifted rapidly. These developments, according to Chien, effectively erased many of the early commercial opportunities that had once looked so promising.
To illustrate this dynamic, Chien drew an analogy to the humble flashlight app that briefly became a must-have download when the first iPhones were introduced in 2008. That functionality, however, was quickly integrated directly into Apple’s iOS, rendering the app obsolete. He suggested that a similar pattern is playing out across the AI ecosystem: until the underlying platforms mature and stabilize, consumer-oriented AI products will struggle to find lasting footing. As with smartphones before them, only once the infrastructure solidifies will transformative, category-defining AI applications emerge.
Chien went on to argue that we are currently experiencing a moment analogous to the mobile technology market around 2009–2010—the period when smartphone platforms finally became stable enough to support truly disruptive, mobile-first businesses such as Uber and Airbnb. He speculated that we could be witnessing the early signs of a comparable stabilization in AI, pointing to Google’s Gemini system achieving technological parity with ChatGPT as evidence that the foundational layers of the ecosystem are beginning to settle.
Echoing his views at the same event, Elizabeth Weil, founder and partner at Scribble Ventures, described today’s consumer AI landscape as occupying an “awkward teenage middle ground.” She compared the sector to an adolescent in transition—no longer immature and experimental, yet not quite ready for full-scale adulthood where products become indispensable daily tools. Both Weil and Chien contemplated what it might take for these fledgling consumer AI ventures to mature. One possibility, they suggested, is the advent of an entirely new class of personal devices that could surpass the smartphone in both form and function.
Chien argued that, despite smartphones being deeply integrated into daily life and checked hundreds of times each day, they possess an inherent limitation: the device perceives only a tiny fraction—perhaps 3 to 5 percent—of a user’s real-world environment. Because of this narrow field of perception, he believes smartphones are unlikely to introduce the most compelling AI applications, which require ambient awareness and constant contextual input to operate effectively. Weil concurred, noting that the smartphone’s nature as a device held in one’s hand inherently restricts innovation. Holding up her own iPhone before the audience, she remarked that it was improbable anyone would still be designing for such a constrained form factor five years from now.
Driven by this conviction, startups and major technology companies alike are racing to create a new personal device capable of redefining how individuals interact with artificial intelligence—something that could, in effect, replace or extend the smartphone. Among the most talked-about initiatives is a mysterious collaboration between OpenAI and former Apple design chief Jony Ive to develop what is rumored to be a “screenless,” compact, pocket-sized AI device. Meanwhile, Meta’s Ray-Ban smart glasses incorporate AI capabilities that can be managed via a wristband sensitive to hand gestures. Other startups are experimenting with less conventional formats—wearable pins, pendants, and rings—that rely on AI to create hands-free, contextually aware experiences, although many of these experiments have met with limited success so far.
Nevertheless, as Chien pointed out, the next generation of consumer AI innovation may not depend entirely on novel hardware. He envisions opportunities for services that function within existing devices yet deliver deeply personalized experiences—such as an AI-powered financial adviser customized to the specific circumstances, goals, and habits of each user. In a similar vein, Weil predicts the rise of an omnipresent AI tutor designed to provide tailored instruction continuously, offering individualized educational support directly through one’s smartphone.
Although both investors expressed excitement about AI’s transformative promise, they also voiced caution regarding a new crop of stealth-mode startups attempting to create AI-driven social networks. According to Chien, these ventures revolve around ecosystems where thousands of AI bots interact with users and their content, resulting in social feeds that mimic engagement without genuine human presence. “It effectively turns social networking into a single-player game,” he warned, questioning whether such artificial communities could ever replicate the emotional authenticity that draws people to share, connect, and participate online. For both Weil and Chien, the enduring strength of social networks lies precisely in the understanding that real humans—rather than algorithms or synthetic personalities—exist on the other end of every post, message, or exchange.
Sourse: https://techcrunch.com/2025/12/15/vcs-discuss-why-most-consumer-ai-startups-still-lack-staying-power/
