Has the speaker become a new entry point for AI smart devices? Sonos (SONO.US) is betting on an “open platform” transformation: a 20% increase in revenue is just the beginning, and many new products are ready to be launched next year

The Zhitong Finance App learned that Sonos Inc. (SONO.US) announced revenue growth and forecasted new products. The company is seeking to turn its iconic audio system into an open platform for showcasing various artificial intelligence devices throughout the home.

CEO Tom Conrad said during an earnings call on Tuesday, “Many companies in the industry are building 'walled gardens', and their ecosystem only welcomes one smart device.” “We want the AI you choose — no matter what company it comes from — to perform better on Sonos than anywhere else.”

In the fourth quarter of fiscal year 2026, Sonos surpassed expectations and achieved revenue of US$344.4 million, an increase of nearly 20% over the same period last year. GAAP net loss of $5 million ($0.04 per diluted share), a year-on-year decrease of $33 million; adjusted EBITDA of $6 million was essentially the same. Looking at the subregions, the Americas grew 17%, the Asia-Pacific region grew 29%, Europe, the Middle East and Africa grew by 24%, and all regions grew across the board.

Conrad said in the company's earnings report, “We said that fiscal year 2026 will be a year for Sonos to return to growth, and we have fulfilled our promise.” The company absorbed 56 million dollars of rising memory costs throughout the fiscal year and is currently “optimizing the memory usage of products.”

During the earnings call, Conrad revealed that Sonos plans to launch “multiple” new products next year and is expected to usher in a holiday season with strong hardware sales. The company's revenue forecast for the first quarter of fiscal year 2027 is $570 million to $630 million, up 4% to 15% year over year. Wall Street had previously expected $575 million.

Conrad said on the conference call, “From the first day I took over this job, my focus was on using Sonos as a system — not a collection of speakers,” adding that today's Sonos is a very different company compared to when he took over as CEO last year. “We've made it through this year's journey and the 'Sonos System' is taking off,” he said.

Sonos launched two new consumer-grade hardware at the end of the quarter — the Ace Ultra headset and the Beam Ultra sound bar, so only initial sales data was included in Tuesday's results. The company also launched the so-called “Company's New Creative Direction” in mid-September, and related advertisements focused on “the feeling of the entire system working together.”

But the bigger story is on the software side: Sonos released the Sonos 27 operating system update in September, and later introduced new features such as System Health (System Health) — a dashboard that can view connected speaker systems and flag common network issues that may have a negative impact on performance.

“Every fall, you can look forward to us bringing a new generation of systems capabilities. That's how we're bridging the gap between Sonos and everything else in the home audio space,” he said. The Sonos mobile app now has more intuitive navigation, as the company allegedly spent “hundreds of hours” observing how customers use it.

This month, Sonos began gradually rolling out its self-developed AI voice assistant — branded “Sonos 27voice” — to US customers through an early experience program. Designed specifically for music, the new system is better at understanding natural language and connecting multiple instructions, and can also answer general knowledge questions — a capability not available in the company's previous voice assistant Sonos Voice Control.

Conrad also showed enthusiasm for personal AI agents such as Meta Platforms Inc. (META.US)'s Muse. Last week, he shared how to connect Muse to Sonos devices through the audio brand's recently launched Model Context Protocol (MCP). The MCP acts as a universal connector for large language models.

In addition to Muse, the Sonos 27MCP also supports Claude, ChatGPT, and Gemini, enabling these models to control speakers, group products, and play music from connected streaming services.