A look at consumer AI trends: ChatGPT has 3x more US subscribers than Claude or Gemini, the top 1% of spenders drive 19.5% of spend, and AI agents gain traction
THE SO WHAT
Consumer AI is consolidating around a few brands — with ChatGPT 3x the US subscribers of Claude or Gemini and spend heavily skewed to a top 1% — while agents start to show real traction. For product teams, that means distribution will be dominated by a handful of assistants, and the more interesting opportunity is building sticky agentic workflows on top rather than yet another standalone app.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIAnthropic, OpenAI Can Both Win AI Race: T. Rowe Price
Institutional capital is now explicitly underwriting a multi-winner model layer — Anthropic as workflow execution, OpenAI as broad consumer and enterprise surface — which reduces near-term platform risk for builders. For operators, the real decision is not “which lab” but how tightly you couple to any single ecosystem versus keeping your workflows portable.
Applied AIBender: AI’s ‘Existential’ Risk Is ‘Fake’
The center of gravity in AI risk discourse is shifting toward concrete harms — labor displacement, surveillance, misinformation, environmental load — and away from speculative x-risk. Leadership teams will increasingly be asked to show how their AI deployments mitigate these near-term harms in procurement, compliance, and PR conversations.
Applied AIWikimedia’s OpenAI Rebuke Points to Costs of AI Data Scraping
Wikimedia pushing back on OpenAI over data access costs underscores that high-quality public corpora are no longer “free” fuel for models — they’re becoming negotiated, priced inputs. Expect more rights holders to demand compensation or technical controls, and bake rising data acquisition costs and licensing complexity into your model strategy.
Applied AIMirror Particle is building a ‘world model’ of human behavior
A startup pitching a from-scratch “world model” of human behavior for market research is a sign that teams are hitting the ceiling of LLM roleplay for serious decision support. Brand and product leaders should treat these behavioral models as hypothesis generators — powerful, but in need of tight grounding with real customer data before steering strategy.