Google built a great smart speaker, but Gemini isn’t ready for it

The more pointed question the summary sidesteps is whether Google is shipping hardware now precisely because Gemini isn't ready, using a tangible product launch to buy time and shelf presence while the model catches up. A great speaker with a mediocre assistant isn't a competitive product; it's a placeholder.

The readiness gap here mirrors a pattern visible across the frontier model space right now. OpenAI's reported move to ship GPT-5.6 as three distinct variants rather than a single premium model (covered from The Decoder on July 1) suggests even the most capable labs are struggling to deliver a single model that performs reliably across all deployment contexts. Always-on, low-latency home interaction is one of the hardest contexts to satisfy, requiring consistent reasoning under real-world noise and interruption conditions that benchmarks don't capture. Google is essentially betting that Gemini will mature into the hardware before consumers notice the gap, which is a risky sequencing call when competitors are iterating on model tiers aggressively.

If Google announces a Gemini-specific update targeting latency and contextual memory for home devices within the next two product quarters, that confirms the software gap was known and scoped at launch. If no such update ships, the speaker risks becoming a capable but underutilized device, which is exactly how the last smart speaker cycle ended.