OpenAI launches always-on Dots agents at DevDay 2026

By Splitfeed EditorialLast updated Sep 29, 2026, 18:11 UTC

On September 29, 2026, at its annual DevDay conference in San Francisco, OpenAI unveiled Dots, a line of always-on AI agents built on the company's GPT-6 Astra model. Unlike chatbots that respond only when prompted, Dots run persistently on their own cloud computers, maintaining read-only background access to handle tasks such as fixing bugs or sending invoices without direct user input. The launch came alongside a new Pro 500 subscription tier priced at $500 per month and a broad set of ChatGPT updates: shared workspaces, collaborative documents and slides, an open plugin system with MCP event automation, and an enterprise marketplace featuring 32 partners. CEO Sam Altman described the product as something OpenAI has "found" that is genuinely new, and the event was billed as featuring more than 20 separate launches.

The announcement carries weight beyond a single product release. Dots integrates directly with Slack and Microsoft Teams, placing OpenAI in more direct competition with Microsoft, which is both a major investor and a distribution partner. The product also enters a contested market: Meta's rival agent, Muse, has already been blocked by Amazon and Walmart, signaling that platform owners are prepared to resist third-party agents that touch commerce and user data. The $500 price point for the new tier suggests OpenAI is targeting enterprise and power users rather than the general consumer, a strategic choice that will test whether businesses are willing to pay at that scale for autonomous background agents. The scope of the enterprise marketplace, with 32 named partners at launch, indicates OpenAI is building an ecosystem rather than a standalone tool.

None of the three camps, Pro-AI, Anti-AI, or Middle Ground, had published reactions at the time of writing. Pro-AI voices would typically celebrate persistent agents as a leap toward genuinely useful AI that handles real work autonomously, pointing to the integrations and marketplace as signs of a maturing platform. Anti-AI commentators would be expected to raise concerns about always-on agents with background data access, the risks of autonomous action in workplace environments, and the precedent set by retailers blocking agents they cannot control. Middle Ground observers would likely focus on whether the governance and permission structures around read-only access are robust enough, and whether the $500 tier creates an affordability gap that concentrates the most capable AI tools among large enterprises.

The clearest near-term signals to watch are whether additional major platforms follow Amazon and Walmart in blocking third-party agents, how enterprise adoption of the Pro 500 tier develops in the months after launch, and whether OpenAI clarifies the boundaries of what Dots can do when operating in background mode without active user direction.