I’ve been watching smart speakers drift into the background for years—still useful, still perched on kitchen counters, but rarely exciting. So when Google turns around and says, in effect, “No, wait—this category isn’t done,” I can’t help but lean in. After a long run of incremental updates, the company is betting that Gemini can breathe new life into the form factor. The result is a new $99.99 device called the Google Home Speaker, built specifically with Gemini in mind.
The first real reset since 2020
Google hasn’t launched a standalone smart speaker since the Nest Audio arrived in September 2020. Back then, smart speakers were mainly framed as two things: music boxes and smart-home remote controls. They were helpful, sure—but they also felt brittle. If you didn’t phrase a command “correctly,” you’d end up repeating yourself, simplifying your request, or giving up altogether.
That era also predated today’s mainstream AI chatbots. We weren’t used to devices that could follow a winding question, handle context, or understand what we meant when we changed our minds mid-sentence. We were trained to talk to machines like machines.
Google is now trying to flip that dynamic: talk to the speaker the way you talk to a person, and let Gemini do the translation.
Natural language, finally treated as natural
The big promise here is that you can speak normally—messy, multi-part, human—and still get results. Instead of carefully structured commands, the new Google Home Speaker is designed for natural language requests, including multi-step ones.
Google’s own examples are telling, because they’re exactly the kinds of things people want but rarely bother asking for when a system feels too strict. You could say: “Turn off all the lights except the lamp next to my bed,” or “Dim the kitchen lights, put on relaxing music, and set a timer for 20 minutes.” That’s not one command; it’s a small routine delivered in plain speech.
And here’s the part I suspect will matter most in daily life: the speaker can handle corrections while you’re still talking. If you’ve ever had that mini moment of panic—“Wait, did I just tell it to do the opposite of what I want?”—Google claims Gemini can track the change in intent. The example goes something like: “Turn off the coffee maker… actually, turn it on!” and the system is supposed to follow the revised instruction without forcing you to start over.
It sounds minor until you realize how often smart speakers fail not because people don’t know what they want, but because people edit themselves in real time. This is Google trying to make that normal behavior feel supported instead of punished.
Ten voices and a broader idea of “conversation”
Beyond smart-home tasks, Google says the device will ship with 10 new voices that can hold two-way conversations on a wider range of topics. In other words, it’s not meant to be only a household switchboard. It’s also meant to be something you can ask questions to—more like chatting with Gemini on your phone, but now planted in your living space.
If you’re the kind of person who already uses a chatbot to explore a topic—whether that’s learning something, brainstorming, or just satisfying curiosity—this is Google’s attempt to make that behavior ambient. Not “open an app,” but “just ask.”
There’s also a “Continued Conversation” mode, where the microphone can stay active for a short time after you ask something. The idea is to keep the flow going so you don’t have to repeat “OK, Google” every time you follow up. I’ve always found wake words slightly awkward in real conversation, like you’re constantly tapping someone on the shoulder before every sentence. If this works smoothly, it could make the speaker feel less like a command line and more like a presence.
Familiar design, clearer feedback
Visually, Google isn’t trying to shock anyone. The new speaker keeps the familiar look: a rounded shape with a 3D-knit fabric exterior, measuring 3.4 x 4.2 inches. In the U.S., it’s offered in Jade and Berry, alongside Hazel and Porcelain (which are also available in the rest of the world).
The more interesting detail is a new light ring on the bottom that signals what’s happening—whether the speaker is listening, “thinking,” or responding. That kind of feedback matters more than companies admit, because the fastest way to make people distrust an always-on device is to make it unclear what it’s doing at any given moment. If the light ring communicates state clearly, it’s not just decoration; it’s a small piece of transparency.
The catch: Gemini features aren’t all included
Now to the part that complicates the story. Not everything in this new Gemini-forward world is free. Google plans to offer a subscription tier called Google Home Premium for $10 per month (or $100 per year). If you want the more powerful AI features, that’s the door you’ll be asked to walk through.
Premium includes more open-ended conversations through Gemini Live, launched with the phrase: “Hey, Google, let’s chat.” It also extends into the security and monitoring side of the smart home: Google says it can help you ask questions about activity captured by Nest cameras, and it can generate summaries of what happened at home while you were away.
This is where I find myself asking you the obvious question: Is that worth yet another monthly subscription? Because the subscription economy has turned “nice to have” into “pay to keep,” and consumers have gotten more selective. The irony is that many Gemini features will still be available without paying—so Google must convince people that the extra layer feels less like a toll and more like a genuine upgrade.
A familiar playbook: free first, pay later
Google is leaning on a strategy we’ve all seen before: give people a taste, then ask them to commit. The company will include six months of advanced features at no charge, with the expectation that some users will get used to the premium experience and decide they can’t go back.
If that plan works, Google accomplishes two things at once. First, it makes the smart speaker category feel new again by attaching it to generative AI—not just smarter commands, but a different relationship with the device. Second, it builds a recurring revenue stream tied to the home, where switching costs are psychological as much as technical.
For years, smart speakers have felt like a solved problem: play music, set timers, dim the lights. Google is now arguing that the real product isn’t the speaker—it’s the conversation. The question is whether people will embrace that shift… and whether they’ll keep paying for it once the six-month honeymoon ends.
Comments