Even after three years since the onset of the generative AI explosion, the majority of AI startups still find profitability by catering to enterprises rather than individual users.

While general-purpose LLMs like ChatGPT were swiftly embraced by consumers, many targeted consumer GenAI offerings have not yet made a significant impact.

“Many of the initial AI applications related to video, audio, and photography were extremely impressive,” remarked Chi-Hua Chien, co-founder and managing partner at Goodwater Capital, during TechCrunch’s StrictlyVC event in early December. “But then Sora and Nano Banana emerged, and the Chinese open sourced their video technologies. Consequently, many of those opportunities vanished.”

Chien likened some of these applications to the basic flashlight, which became a popular third-party addition right after the iPhone’s release in 2008, but was subsequently integrated into iOS.

He maintained that, similarly to how it took a few years for the smartphone ecosystem to mature before revolutionary consumer applications surfaced, AI platforms require a comparable phase of “stabilization” for enduring AI consumer products to thrive.

“I believe we are on the brink of what could be compared to the mobile landscape of the 2009-2010 period,” Chien stated. That era marked the inception of enormous mobile-first consumer enterprises like Uber and Airbnb.

Chien suggested we might be observing early signs of that stabilization as Google’s Gemini achieves technological equivalence with ChatGPT.

Elizabeth Weil, founder and partner at Scribble Ventures, reflected Chien’s perspective regarding the formative phase of GenAI, portraying the current landscape of consumer AI solutions as existing in an “awkward teenage middle ground.”

What is required for consumer AI startups to mature? Perhaps a novel device that goes beyond the smartphone.

“It seems improbable that a gadget which you interact with 500 times a day yet only registers 3% to 5% of your surroundings will ultimately facilitate the full realization of AI’s potential,” Chien commented.

Weil concurred that a smartphone may be too restrictive for re-envisioning consumer AI products mainly because it lacks an ambient nature. “I don’t anticipate we’ll be developing for this in five years,” she stated, gesturing to her iPhone as she displayed it to the audience.

Startups and established tech firms have been competing to create a new personal device that could replace smartphones.

OpenAI and Apple’s previous design head, Jony Ive, are reportedly developing a “screenless,” pocket-sized device. Meta’s Ray-Ban smart glasses are operated by a wristband that senses subtle movements. Meanwhile, numerous startups are attempting—often with lackluster outcomes—to introduce a pin, pendant, or ring applying AI differently than smartphones do. 

Nevertheless, not every AI consumer offering will hinge on the introduction of a new device. Chien proposed that one possibility could be a personal AI financial adviser tailored to individual needs. Simultaneously, Weil predicts that a personalized, “always-on” tutor will become prevalent, delivering its specialized instruction straight from a smartphone.

Although they are excited by AI’s capabilities, Weil and Chien voiced doubts about the arrival of several still-under-the-radar AI-driven social networking startups. Chien described these companies as building platforms where numerous AI bots interact with the user’s content.

“It turns social media into a single-player experience. I’m not convinced it works,” he stated. “The appeal of social networking lies in the awareness that genuine humans are present on the other side.”