Days after launching its K3 frontier model, Moonshot said demand had pushed compute close to capacity in about 48 hours — so it temporarily stopped new signups to protect existing subscribers, and said it would add GPUs and split membership into two plans.
2026-07-19 · by Moe Ameen
Moonshot AI, the Beijing lab behind the Kimi series, temporarily paused new Kimi subscriptions in mid-July 2026, days after launching its flagship Kimi K3 model. In a public post the company said K3 had "received far more love than we expected," and that over roughly the previous 48 hours demand had pushed close to the limits of its current GPU capacity. To protect the experience of existing subscribers, it stopped taking new paid signups while it worked to add compute.
The pause was framed as temporary and supply-driven, not a policy change. Moonshot said it was expanding GPU capacity as fast as it could and would reopen new-subscription spots in batches rather than all at once. Existing subscribers were explicitly unaffected — the limit applied only to new members trying to sign up during the crunch.
Alongside the pause, Moonshot said it would restructure membership into two more focused plans: a Kimi Membership covering Kimi Web, the Kimi app, and Kimi Work, and a separate Kimi Code Membership for coding workflows. The stated reasoning was to match compute to usage more precisely and keep the experience stable, since chat, agentic work, and heavy coding load the system differently. Kimi K3 itself launched in mid-July 2026 as Moonshot's new flagship — a very large, natively multimodal model with a 1-million-token context window, framed by the lab as the largest open-weight model to date — which is the release that drove the surge.
Start with the newsjack, because this is a genuinely good story to ride. "An AI lab had to stop selling subscriptions because its new model was too popular" is a hook the AI and creator crowd is already sharing, and a fast, useful take earns reach. Drop a short brief — what happened, why GPU capacity is the real constraint behind every AI product, what the membership split means — into Kompozy and it becomes a same-day package: a captioned Persona Short or Clipped Short explaining the crunch, a brand-exact Carousel on "demand outran compute," a Quote Graphic on the 48-hour capacity line, native Text Posts for X and LinkedIn, a Blog Article for the long-tail search, and an Email Newsletter to your list — all in one voice via your Persona Brief, reframed to 9:16, 1:1, and 16:9, then scheduled and published across nine platforms plus blog and email from one queue.
Then there is the quieter takeaway, and it is the one that compounds. The reason new Kimi members were locked out is that their access ran through one lab's finite GPUs. If your content operation is wired to a single assistant, you inherit its outages and waitlists. Kompozy is built the other way: it runs its own managed Claude and OpenAI models for generation and hands you outcomes — 18 content formats, persona and avatar video a chat model can't render, brand-voice governance, and autopilot publishing — so a capacity crunch at any one provider doesn't dark your pipeline. Use a moment like this to publish everywhere today, and to make sure the day one model is unavailable is not the day your posting stops.
Moonshot said its new Kimi K3 model drew far more demand than expected, pushing GPU capacity close to its limits over roughly 48 hours. To protect the experience of existing subscribers, it temporarily stopped new signups while adding compute, and said it would reopen new-subscription spots in batches.
No. Moonshot said the pause applied only to new paid signups; existing subscribers kept their access. The company framed the move as a temporary, supply-driven measure to keep the experience stable for current members while it expanded GPU capacity.
Alongside the pause, Moonshot said it would divide membership into two more focused plans: Kimi Membership for Kimi Web, the app, and Work, and a separate Kimi Code Membership for coding workflows. The goal is to match compute to how each workload loads the system. Confirm current plan details on Moonshot's own pages.
Don't tie your whole workflow to one hosted assistant. A content engine like Kompozy runs its own managed Claude and OpenAI models and generates 18 formats plus publishes to nine platforms, so a capacity crunch at any single provider doesn't stop your production or scheduling.