Local voice assistants need tiny, fast offline TTS — build a 9M-param engine

Local voice assistants need tiny, fast offline TTS — build a 9M-param engine targets a $1.2B = 200K developers/device-makers × $6K ACV (annual commercial licensing, SDKs, and support) estimated for voice/TTS purchases total addressable market with medium saturation and a year-over-year growth rate of 12% YoY — speech and voice AI market growth (industry reports such as Grand View Research and market summaries, 2023-2025 estimates).

Key trends driving demand: Local-first AI — more apps and devices are moving workloads on-device for privacy and latency, creating demand for lightweight models.; Model compression advances — distillation, pruning, and INT8/INT4 quantization make small high-quality TTS models possible and practical.; Rising cloud costs and latency sensitivity — rising per-request cloud costs push product teams to prefer local inference for frequent or offline use cases.; Open-source momentum — developers expect rapid iteration and transparent licensing, which favors small, permissively-licensed models that are easy to integrate..