Kimi K3 Pauses New Subscriptions Due to Compute Overload
firstadopter · x · 2026-07-20
Moonshot officially stated that the popularity of the Kimi K3 model following its launch has far exceeded expectations, with massive demand over the past 48 hours pushing current compute capacity to its limits. To ensure a smooth experience for existing subscribers, they have temporarily halted new user subscriptions and will prioritize allocating compute resources to current members.
Furthermore, Moonshot recommends using a supernode configuration equipped with 64 or more accelerators when deploying Kimi K3 to achieve a better high-bandwidth communication domain and improved inference efficiency.
Related event: Kimi K3 Demand Overloads GPU Capacity, New Subscriptions Halted(10 posts)→
More from Infra
- Anthropic's ~11 GW contracted compute estimated at ~42% NVIDIA GPUs, ~4.5 GW Google TPU — ns123abc · 2026-10-27
- EdgeRock-4B holds KV cache at 72 MiB across 256K context, 99.1% smaller than Qwen3.5-4B — alifcoder · 2026-10-27
- RTX 5080 price jumps from $1658 to $2318 in ten days on Amazon — Frizzy-MacDrizzle · 2026-10-27
- ComfyUI on 12GB VRAM: measured peaks show offloading turns 24s into 11 minutes — Impossible-Mail-4619 · 2026-10-27
- Neon rebuilds its docs search entirely on its own stack: Lakebase Search, Functions, AI Gateway — matei_zaharia · 2026-10-27
- PyTorch details 2.12x faster on-device streaming ASR on Arm SME2 via ExecuTorch — PyTorch · 2026-10-27