$99 Google Home Speaker Retires Assistant for Gemini
Google launched the $99.99 Google Home Speaker, featuring a dedicated NPU to shift the ecosystem from legacy Assistant to the generative Gemini for Home.
On June 17, 2026, Google announced the Google Home Speaker, a $99.99 standalone device that replaces the traditional Google Assistant with a generative AI system called Gemini for Home. Pre-orders are open, with retail availability starting June 25, 2026. The hardware launch marks the end of the legacy “Nest” branding for smart speakers and introduces a dedicated NPU to handle natural language processing closer to the edge.
Hardware and Connectivity
The spherical, 3D-woven fabric speaker runs on a Quad Core A55 2.0 GHz processor paired with an AI accelerator. The inclusion of a dedicated NPU allows the device to process voice agents more efficiently than previous cloud-only models. It includes 1GB of RAM and 4GB of storage.
| Specification | Detail |
|---|---|
| Processor | Quad Core A55 2.0 GHz |
| Memory | 1GB RAM, 4GB Storage |
| Audio | 58mm full-range driver |
| Networking | Wi-Fi 6 (802.11ax), Bluetooth 5.4 |
| Protocols | Matter Controller, Thread 1.3 border router |
Audio performance relies on a 58mm full-range driver delivering 360-degree sound. Google states the device outputs 2.5 times more bass than the Nest Mini. The speaker functions as a Matter Controller and a Thread 1.3 border router, integrating with existing third-party smart home devices.
The physical interface features three capacitive touch areas on top for volume and playback control. A dynamic light ring at the base illuminates to indicate when Gemini is listening, thinking, or responding. Buyers can choose from Jade, Berry, Porcelain, and Hazel colorways, depending on regional availability.
Gemini for Home Integration
Gemini for Home processes free-flowing natural language instead of rigid command syntax. Users can issue complex requests spanning multiple domains, such as dimming the kitchen lights, playing music, and setting a timer in a single breath. The system also handles mid-sentence corrections, allowing a user to tell the assistant to turn off a coffee maker before immediately correcting the command to turn it on instead.
The assistant utilizes a Continued Conversation feature that maintains context over short periods. If you are implementing short-term memory in your own applications, Gemini’s approach demonstrates how to let users ask follow-up questions without repeating the wake word. The system ships with 10 natural voices and executes advanced reasoning tasks, such as cross-referencing real-time weather data against the specific location and time of a live sports event.
Hardware and Subscription Pricing
The base hardware costs $99.99 and includes core NLP capabilities. Google gated the more compute-intensive generative features behind a Google Home Premium subscription priced at $10 per month or $100 annually.
The premium tier unlocks Gemini Live for continuous real-time chat. It also enables Home Briefs, which generate summaries of household activity, and Camera History Search, allowing users to query Nest camera footage using natural language. Each speaker includes a six-month trial of the premium service.
If you develop smart home integrations or custom voice applications, the shift from syntax-based routing to generative intent parsing requires updating your command structures. Devices that rely on rigid invocation phrases will need to adapt to Gemini’s free-flowing natural language model to remain discoverable in the new ecosystem.
Get Insanely Good at AI
The book for developers who want to understand how AI actually works. LLMs, prompt engineering, RAG, AI agents, and production systems.
Keep Reading
How to Cut CPU Costs with Cloudflare Workers Cache
You will learn how to configure Cloudflare Workers Cache to serve responses directly from entrypoints, handle invalidations, and partition cache keys.
Native Slide Generation Enters ChatGPT Canvas via NextSlide Buy
OpenAI has confirmed the acquisition of NextSlide, pivoting the startup's prompt-to-deck generation technology into native presentation features for ChatGPT.
LFM2.5-2.6B Fits Agent Workloads Into 2GB of Device RAM
Liquid AI's new 2.6B parameter hybrid models deliver 2x faster CPU processing and fit into 3GB of RAM for on-device agentic workflows.
Synthesia Pairs OpenAI Models With Avatars for Live AI Coaching
Synthesia launched Roleplay Sessions, an enterprise platform combining proprietary avatars and OpenAI reasoning models for interactive behavioral coaching.
SymptomAI Agentic Interviews Beat Clinician Diagnostic Accuracy
Google Research demonstrated that active, agent-driven symptom interviewing yields significantly higher diagnostic accuracy than human clinicians.