A16z Outlook 2026: Startups Still the Key Driver of Consumer AI, LLM Segmentation Takes Shape

a16z multiple long-term investors in American industries, financial services, and enterprise software stated in a video released at the end of 2025 that early in 2025, OpenAI and Google almost simultaneously accelerated their deployment of consumer AI, from models to interface design, with the goal of making it easy for general users to get started. a16z believes that the consumer AI market has already shown early structures of “winner-takes-all or a few leading players capturing most usage,” but as demand diversifies, the LLM market is moving toward niche segments, and startups remain key drivers in application and product experience.

Start of the year, the consumer AI battle begins

Looking back to early 2025, OpenAI and Google almost simultaneously ramped up their investments in the consumer AI market. Whether it was launching new models, updating features, or trying out entirely new user interfaces, the direction was very clear: enabling ordinary users without technical backgrounds to easily use AI.

a16z points out that the reason this competition is critical is that the consumer LLM assistant market has already shown early signs of “winner-takes-all, or at least the winner capturing the majority of usage.” Whoever secures a successful position first may see their advantage continually amplified.

Current usage comparison, ChatGPT leads significantly

Based on actual usage data, most consumers currently choose only one primary AI product. Surveys show that only about 9% of users pay for two or more of ChatGPT, Gemini, Claude, or Cursor simultaneously.

In overall scale, ChatGPT currently has about 800 million to 900 million weekly active users, clearly leading other competitors. Gemini on the web is about 35% of ChatGPT’s scale, and about 40% on mobile. Usage of Claude, Grok, and Perplexity mostly falls between 8% and 10%.

Gemini usage surges rapidly, market segmentation trend becomes obvious

However, a16z also notes that the market situation has changed significantly over the past 3 to 6 months. With multiple image and video models gaining popularity quickly, Gemini’s desktop user growth rate has reached 155% annually, while ChatGPT’s annual growth rate is about 23%.

Meanwhile, clear niche segmentation has begun to emerge, such as Anthropic’s Claude, which is gradually establishing a clear position among highly technical or professional user groups, no longer competing directly with ChatGPT and Gemini for the general public market.

The biggest highlight of 2025, image and video generation as the main focus

From a model perspective, the most viral spread in 2025 will no longer be text models, but image and video generation. OpenAI, including GPT-4 Image, has sparked a “Ghibli-style” craze, along with Sora and Sora 2. Google has released VO, V3, V3.1, and later became popular Nano Banana and Nano Banana Pro.

a16z describes that the diffusion power of Nano Banana is already comparable to, or even surpasses, the Ghibli moment of the past. At the same time, the technical focus of image models has shifted from aesthetics and style to realism and reasoning ability, capable of understanding multiple images and text simultaneously, handling background dynamics, physical plausibility, and even generating complex infographics and market maps.

Looking ahead to 2026, AI moves toward multimodal integration

In terms of product strategy, OpenAI and Google are taking different routes. Most of OpenAI’s features are integrated directly into ChatGPT, with Sora launched as a standalone video app. Google, on the other hand, releases products through Gemini, Google AI Studio, Google Labs, and various independent websites, giving each product its own interface. a16z points out that this difference will directly impact whether ordinary users know where to start when opening a product.

In social features, OpenAI’s group chat and Sora’s video dynamic wall have received relatively conservative responses. a16z believes that ChatGPT’s core remains a productivity tool, which makes it difficult to satisfy the motivation of social platforms to “be seen and recognized.”

Looking forward to 2026, a16z generally believes that AI will continue to move toward “any input, any output” multimodal integration, while large language model providers may continue to focus on optimizing core experiences. Meanwhile, consumer AI applications with strong product personalities will still be led by startups.

(AI agent farewell command input box? a16z predicts three major changes in AI applications in 2026)

This article, a16z outlook for 2026: startups remain key drivers of consumer AI, and LLM segmentation contours are taking shape, first appeared in Chain News ABMedia.