OpenAI Pushes Into Smart Speakers as the Category Shrinks

There is a specific kind of strategic red flag that lights up when a company already losing tens of billions a year decides its next big bet is entering a market that is itself shrinking. Engadget reports that OpenAI's first proprietary device is a 'humanlike' rechargeable smart speaker with a camera, sensors, and mechanical parts that move to make it feel alive, with the company aiming to unveil it this year and ship it in 2027.

The setup rhymes uncomfortably with a warning from Fidji Simo, a former OpenAI executive, who told the company it risked missing its moment by being 'distracted by side quests' like the now-shuttered Sora video app. The financial context that makes that warning sting: leaked documents cited by Engadget put OpenAI's losses at $5.09 billion in 2024 and $38.5 billion in 2025, with no path to profitability until at least 2030.

The market it is walking into is not welcoming either. IDC director Jitesh Ubrani told Engadget that shipments dropped 16.3% year-over-year in 2023, another 11.8% in 2024, and 6.7% in 2025 even after Alexa+ landed, with a further 9.6% decline forecast for 2026. Ubrani's mechanical explanation is worth quoting: 'Creating the device is not hard, but getting it to reach any sort of mass scale, that's a lot more difficult.' Buyers have been trained to expect speakers under $100, and, in Ubrani's words, 'many people would balk at the idea of spending $300 for a speaker.' Amazon, which dominates the installed base, still has not cracked Alexa monetization beyond timers and weather; its Q1 2022 digital unit reportedly booked around $3 billion in operating losses, most attributed to Alexa. Meta abandoned Portal in 2022.

The honest caveat is that this is single-sourced reporting about an unshipped product, so take the specifics, the humanlike form factor, the 2027 timeline, the device brief, as reported rather than settled. And a ChatGPT-native companion is not the same product as an Echo Dot, so straight-line demand data may understate whatever OpenAI actually plans to charge and sell.

The forward-looking read is that if anyone can rewrite consumer willingness-to-pay for a voice device, an OpenAI-branded companion is the plausible candidate. But that is precisely the kind of bet the person warning about side quests would have flagged, and the current burn rate leaves very little room for another Sora.