OpenAI’s hardware pivot risks deepening losses as smart speaker market shrinks

OpenAI is developing its first-party hardware, marking a significant strategic shift for the artificial intelligence firm. According to Bloomberg’s Mark Gurman, the company’s debut device will be a “humanlike” rechargeable smart speaker equipped with a camera, sensors, and mechanical elements. The device is designed to utilise OpenAI’s GPT-Live-1 voice model, which features a duplex architecture allowing for simultaneous input processing and output generation to facilitate more natural, human-level interaction.

This move into physical hardware comes at a time when the global smart speaker market is in steep decline. Market research firm IDC forecasts that the sector will shrink by 9.6 per cent in 2026 before flatlining in 2027. Jitesh Ubrani, director of worldwide device trackers at IDC, noted that shipment volumes have dwindled to tens of millions of units, following a 16.3 per cent year-on-year drop in 2023 and an 11.8 per cent decline in 2024. The introduction of Amazon’s Alexa+ service in 2025 slowed the bleeding slightly, with the market contracting by just 6.7 per cent that year, but the overall trajectory remains negative.

A primary driver of this stagnation is consumer behaviour. Ubrani explained that there is little incentive for users to upgrade existing hardware because new AI capabilities are delivered via the cloud. “There’s really no need to refresh your device because the latest features are cloud-based, and so you can continue on with the same hardware,” he said. Consequently, the market has become saturated with low-cost devices, with Ubrani noting that many consumers would balk at spending $300 on a speaker, a price point likely required for OpenAI’s more complex device.

The financial implications for OpenAI are stark. Leaked documents reveal the company lost $5.09 billion in 2024 and a staggering $38.5 billion in 2025. OpenAI has stated it does not expect to achieve profitability until at least 2030, citing surging AI costs. Former executive Fidji Simo previously warned employees that the company risked missing its moment by being “distracted by side quests,” a comment that now appears prescient given the resource-intensive nature of hardware development.

Established tech giants have struggled to find a profitable model in this space. Amazon’s Worldwide Digital unit, which includes the Alexa and Echo ecosystem, reported an operating loss of $3 billion in the first quarter of 2022. Meta similarly ended development of its Portal smart displays in 2022, a move accompanied by 11,000 job cuts. While Ubrani acknowledges that hardware is necessary for AI to understand the physical world, he cautioned that entering a category with entrenched players and limited monetisation potential is a significant challenge for an unprofitable firm.