Why 2026 could become defining year for consumer AI startups

At a time when large language model (LLM) assistants like OpenAI’s ChatGPT and Google’s Gemini are edging closer toward market consolidation, artificial intelligence (AI) startups may still find room to survive.

This room, according to consumer investor Justine Moore, lies in the startups' ability to build “opinionated, standalone consumer AI.”

Along with Justine, investors Anish Acharya, Olivia Moore, and Bryan Kim of venture capital firm Andreessen Horowitz (a16z) shared their forecasts on how consumer AI will pan out in 2026 in a podcast released by the firm on Monday.


Justine Moore said that in the broader LLM assistant market, only 9% of consumers pay for more than one AI assistant across platforms such as ChatGPT, Gemini, Claude, and Cursor. Currently, ChatGPT has over 900 million weekly active users and is growing at about 23% year-on-year (YoY). Further, less than 10% of ChatGPT users had tried out any of the other big LLM providers like Gemini, Justine added.

While this suggests startups may struggle to survive in the LLM assistant market, Moore stressed that it does not necessarily mean the dominant players have built effective products.

“In my opinion, (the big competitors) have launched or tried new consumer products or new consumer interfaces (but) none of those are really working,” said Moore.

This is because the simple “text in, text out” model, according to the investors, has not been successful for large AI labs in fully addressing the breadth of individual users’ problems

“It is because it is not (their) core competency anymore to build opinionated standalone consumer AI. It's actually very positive for consumer startups in that sense,” she added.

These ‘opinionated standalone consumer AI’ are applications or interfaces designed to solve user problems from a particular perspective or tinted glass. Adding to Justine’s point, Kim argued that while opinionated applications may carry legal and compliance risks, such products are likely to offer these startups a greater advantage in the future.

Justine noted that while model improvements may make ChatGPT better, they do not translate into ChatGPT being able to “verticalise into all of the other amazing use cases or products.”

This is why, as Moore claimed, startups can capitalise on this gap by leveraging continually improving LLM models to build products focussed on the specific use cases users need.