On June 16, 2026, TechCrunch reported that ChatGPT’s share of the AI assistant market had fallen below 50% for the first time, citing Sensor Tower’s 2026 State of AI data. By the end of May, ChatGPT accounted for 46.4% of AI assistant market share, while Google Gemini had risen to 27.7% and Anthropic Claude to 10.3%.
That does not mean OpenAI has stopped growing. The same report still described ChatGPT as the largest AI assistant in the market, with more than 1.1 billion monthly users, and OpenAI has previously said ChatGPT has more than 900 million weekly active users, more than 50 million consumer subscribers, and more than 9 million paying business users. The new signal is different: the market is getting less concentrated, even while overall usage keeps expanding.
What the new data actually shows
The sharpest near-term signal is the end-of-May assistant-share split reported on June 16: ChatGPT at 46.4%, Gemini at 27.7%, and Claude at 10.3%. TechCrunch also reported that the first half of 2026 is on pace for nearly 2.3 billion AI-app downloads and more than $4.2 billion in consumer spending, which suggests the category is still growing fast even as leadership gets more contested.
Sensor Tower’s broader June 2026 reporting points in the same direction. In its State of AI Apps report, the firm said ChatGPT’s share of website visits among leading AI platforms fell from 66% in July 2025 to 50% in March 2026, while Gemini and Claude rose to 22% and 10%, respectively. In a separate State of Web report, Sensor Tower said AI assistants were the fastest-growing web category in 2025, with traffic up 86% and time spent up 101%.
Taken together, those numbers suggest two things at once. First, OpenAI still operates at massive scale. Second, the default user habit of opening only ChatGPT is weakening as more people mix assistants depending on the task, device, or ecosystem they already use.
Why Gemini and Claude are gaining now
Gemini’s rise looks tied to distribution. Google can place Gemini across Search, Android, Workspace, and its broader consumer software footprint, which gives it more chances to become the default assistant for routine tasks. Claude’s rise looks different. Its growth appears more connected to product reputation in productivity-heavy use cases, especially where users care about writing quality, reasoning consistency, and work-oriented interactions.
That distinction matters. A distributed assistant can grow because it is everywhere. A productivity assistant can grow because it wins specific high-value tasks. ChatGPT now faces both pressures at once.
OpenAI still has major advantages in consumer mindshare, enterprise adoption, and paid scale. But the latest market-share shift suggests those advantages are no longer enough to keep the category structurally one-sided. AI assistants are becoming a multi-home market, where many users keep one tool for broad everyday use and another for deeper work.
What this means for enterprise AI buyers
For businesses, this is less a story about OpenAI losing and more a story about single-vendor assumptions getting weaker. If users, teams, and departments increasingly prefer different assistants for different jobs, then enterprise AI strategy moves away from choosing one branded chatbot and toward designing a model layer that can support multiple workflows.
That has practical consequences for AI agents and automation programs. Enterprises will care more about orchestration, evaluation, fallback behavior, cost routing, and governance across models. A support workflow may still perform best with one model family, while research, document review, coding, or internal knowledge work may favor another. The vendor that wins broad awareness does not automatically win every production workflow.
This also makes distribution more strategic. Companies with strong software surfaces, operating-system control, search traffic, or workplace bundles have a clearer path to share growth than model quality alone would suggest. For AI builders, that raises the value of being model-agnostic and workflow-specific instead of betting everything on one provider’s default interface.
What to watch next
The next question is whether this becomes a temporary dip or a more durable structural shift. Three signals matter most.
- Gemini’s staying power: If Google can keep converting ecosystem exposure into repeat usage, its share gains could continue beyond a launch-and-bundle boost.
- Claude’s monetization and retention: If Anthropic keeps turning productivity trust into paid adoption, it could become the strongest premium challenger in the category.
- OpenAI’s response: OpenAI still has enormous reach, strong enterprise traction, and a large subscriber base. The key issue is whether it can keep that scale while the market becomes more multi-model and less default-driven.
There is also a broader infrastructure angle. Sensor Tower said AI referrals still make up only 0.7% of overall web traffic, which means AI discovery is growing fast but has not yet replaced search or other major traffic gateways. That leaves plenty of room for further shifts in distribution, monetization, and assistant behavior over the next few quarters.
The practical takeaway for AI agents and automation teams is straightforward: treat this market-share change as a warning against overfitting your AI strategy to a single front-end brand. The market is still expanding, but it is expanding into a more fragmented, more competitive, and more workflow-specific environment. That is good news for buyers who want leverage, and a stronger case for building AI systems that can route work across models instead of assuming one assistant will stay the default forever.