Google Returns to Smart Speakers After Six Years, With Gemini as the Voice Inside
Google is back in the smart speaker business, and this time the device is built less to play music than to talk. On June 19, the company opened pre-orders for the Google Home Speaker, a $99.99 device shipping June 25 across 18 countries that marks Google's first new consumer smart speaker since the Nest Audio arrived in October 2020. The headline is not the hardware. It is that Google Assistant, the voice that has answered "Hey Google" for nearly a decade, is gone, replaced from the ground up by Gemini for Home.
That swap reframes the device. The Google Home Speaker is being shaped, sold, and marketed as a Gemini endpoint first and a speaker second, and Google is retiring its entire legacy lineup to make room for it. The Nest Mini and Nest Audio, quietly pulled from sale in the weeks before the announcement, are now formally discontinued. Existing units will keep working and will receive the Gemini for Home update over time, but Google is no longer selling them.
From command-matching to conversation
The most consequential change is architectural, not cosmetic. Google Assistant worked by matching spoken commands against a fixed library of recognized intents, which is why it demanded specific phrasing and so often stumbled on anything outside its programmed categories. Gemini for Home, built on a large language model, is designed to understand natural speech without rigid templates. According to Google's announcement, users can issue multi-step instructions, correct themselves mid-sentence, and ask follow-up questions without restating context. The assistant holds short-term conversational memory across a session, so it tracks what was just discussed rather than treating every query as the first.
Google is also leaning hard on on-device processing. The speaker runs on a quad-core ARM Cortex-A55 chip paired with a dedicated neural processing unit, and Google says that NPU runs Gemini Nano inference locally for response times under 100 milliseconds, according to specifications reported by GSMArena and Tech Times. Earlier Google speakers shipped voice queries to remote data centers, a round-trip that added latency and required a live internet connection for nearly everything. Running basic inference locally cuts that lag and, Google argues, improves privacy because routine voice processing no longer has to leave the device. The same NPU handles noise suppression and echo cancellation so the speaker can pick a voice out of a noisy room, as TechCrunch reported when Google first teased the device in October 2025.
There is a catch, and it is a familiar one. The base assistant — smart home control, music, questions, multi-step commands — is free. But the marquee features sit behind a paywall. Gemini Live, the open-ended conversational mode, and camera history search, which lets users ask about activity captured by Nest cameras, both require a Google Home Premium subscription. Google is softening the blow at launch: every pre-order includes a six-month trial of Google Home Premium Standard, valid through September 30, plus three months of YouTube Premium.
A speaker that is also a hub
On the audio side, the Home Speaker uses a single 58mm full-range driver tuned for 360-degree sound, which 9to5Google reported delivers roughly 2.5 times the bass of the Nest Mini. It carries three far-field microphones with a physical mute toggle, a capacitive touch surface, and a light ring that signals when Gemini is listening, thinking, or talking. Two units can pair with a Google TV Streamer for a stereo home-theater setup.
Just as important, the device doubles as a smart home hub. It is Matter-certified and acts as a Thread 1.3 border router, meaning it can bridge low-power locks, sensors, and switches into the network without a separate hub — and it interoperates with devices from Apple, Amazon, and Samsung. The speaker is priced at $99.99 in the US, £99.99 in the UK, €119 in Europe, and CA$139.99 in Canada, in Porcelain, Hazel, Jade, and Berry, with the latter two colors exclusive to the US.
The voice-assistant arms race goes frontier
The launch lands in the middle of a broader race to put frontier AI models into the most ambient interface in the house. Amazon's closest competitor on price, the Echo Dot Max, launched in October 2025 at the same $99.99 and runs Alexa+, Amazon's own LLM-based assistant. Apple, meanwhile, has been retooling Siri around generative AI, and a Gemini-powered backbone has been reported as part of that effort — a reminder of how thoroughly Google's model is spreading across rivals' ecosystems even as it competes with their hardware.
The strategic difference shows up in the business model. Amazon bundles Alexa+ into a $14.99-a-month Prime membership. Google is gating its best features behind a separate Home Premium subscription, betting that conversational AI is compelling enough to convert a free speaker into recurring revenue. That is the real wager here. For nearly a decade, smart speakers were sold near cost as a way to lock users into an ecosystem and capture commerce and search. Gemini for Home flips the logic: the speaker becomes a subscription funnel, and the voice in your kitchen becomes a paid tier.
It is also a test of whether large language models actually behave better as a daily-use voice assistant than the command-matching systems they replace. Reviewers at IntoMobile noted that the genuine test comes after June 25, when real homes — with crosstalk, accents, ambient music, and edge-case requests — determine whether Gemini for Home delivers on its promise. Smart-speaker software has a long history of dazzling in demos and faltering in the kitchen.
What to watch
Three things will define whether this is a genuine reset or a rebrand. First, accuracy in the wild: does conversational Gemini handle messy, real-world speech better than Assistant did, or does the LLM introduce new failure modes like confident wrong answers? Second, the subscription math: how many buyers convert after the six-month Premium trial expires on September 30, which will be the clearest early signal of whether voice AI can pay for itself. And third, the rollout to existing hardware — Google has promised Gemini for Home will reach older Nest speakers over time, and how smoothly that lands will tell hundreds of millions of current owners whether the future requires new hardware or just patience. The wake word stays the same. Almost everything behind it has changed.
---
Sources: [Tech Times](https://www.techtimes.com/articles/318715/20260619/google-home-speaker-launches-june-25-gemini-replaces-assistant-after-six-years.htm), [TechRadar](https://www.techradar.com/home/smart-speakers/the-new-google-home-speaker-is-officially-coming-next-week-and-you-can-pre-order-your-new-smart-hub-right-now-but-youll-need-google-home-premium-for-the-full-gemini-package), [TechCrunch](https://techcrunch.com/2025/10/01/google-teases-its-new-gemini-powered-google-home-speaker-coming-in-spring-2026/), [PhoneArena](https://www.phonearena.com/news/google-home-speaker-with-gemini-release-price-specs-features_id181200).