The dot.com Troll: What xAI's Stunt Reveals About the Agent Land Grab

Some product launches get contested with counter-announcements. This one got contested with a domain name. On September 29, 2026 — the same day OpenAI unveiled dots at DevDay — the domain dot.com began redirecting to x.ai's page for Grok Bot (x.ai/bot). It's a small, almost juvenile move, documented shortly after it happened, and it's also one of the more legible pieces of evidence yet that the race to ship an "always-on agent" has become genuinely personal between OpenAI and xAI.

What actually happened

The mechanics are simple: xAI (or a party acting in coordination with it) acquired or repointed the dot.com domain so that anyone typing it into a browser landed on xAI's own Grok Bot marketing page rather than anything related to OpenAI's newly announced product. The timing removes any ambiguity about intent — this wasn't a domain that happened to change hands during launch week. It redirected on launch day itself, aimed squarely at a company that had, hours earlier, put a product literally named "dots" in front of the press.

It's worth being precise about what this stunt is and isn't. It isn't a security incident, a trademark dispute filed in court, or anything with legal teeth as far as public reporting shows. It's marketing — cheap, fast, and timed for maximum attention during the exact 24-hour window when journalists were already writing headlines with the word "dots" in them. Whoever approved it understood that a redirect costs almost nothing and generates coverage disproportionate to its actual product value, which is precisely what happened: outlets covering OpenAI's launch picked up the dot.com redirect as a footnote worth mentioning in the same breath.

Why this only works between these two companies

A stunt like this requires a rival paying close enough attention to react within hours — and confident enough to spend the effort on a joke instead of a rebuttal.

This kind of move doesn't happen between companies that see each other as adjacent, unthreatening players. It happens between companies racing for the same customer, on the same timeline, for the same category of product. OpenAI and xAI have spent 2026 shipping increasingly similar always-on agent products — dots and Grok Bot both give an agent its own persistent cloud computer, both connect to a user's existing tools, both position themselves as coworkers rather than chatbots. Grok Bot had already been in beta since August 11, 2026, and had just extended to businesses running "parallel bots per department" in early September — xAI had momentum and a live product when OpenAI's DevDay hit, which is exactly the position from which a company feels emboldened to troll rather than panic.

Meta, by contrast, wasn't a target of this particular joke, and that absence is its own data point. Muse launched three weeks earlier, on September 8, and had reportedly topped US and Canadian app charts — genuinely strong momentum of its own. But Meta is playing a different game: consumer app-store attention and small-business integrations, not the developer-and-enterprise lane OpenAI and xAI are both fighting over most directly. The stunt targeted the company whose product name it could literally hijack, which happened to also be its closest structural rival.

The land-grab reading

Step back from the joke and the stunt is a useful proxy for how the three companies see the competitive landscape, whether or not xAI intended it that way. Domain-name theater is what companies do when they believe the name of a category is still up for grabs — when "the agent that has its own computer and works while you're offline" doesn't yet have an obvious market leader, and a cheeky redirect might shape which brand people associate with the idea first. That's a very different competitive posture than, say, contesting a mature market with an established leader, where a stunt like this would read as desperate rather than confident.

It's also worth noting what the stunt doesn't tell you: nothing about which product is actually better, safer, or more capable. Dots' permission model, built around Custom Rules and Auto-Review, and Grok Bot's credential-sharing approach are architecturally very different products aimed at overlapping but distinct audiences. A domain redirect is marketing noise sitting on top of two genuinely different engineering bets — fun to write about, useless for actually deciding which agent to trust with your CRM.

A pattern likely to repeat

If the current trajectory holds — three major labs shipping competing always-on agents within a single month, each trying to out-ship the others on features while pricing and safety documentation lag behind — expect more of this kind of theater, not less. Grabbing a rival's product name as a domain is cheap enough that it will probably happen again the next time someone announces a category-defining feature; it costs a registration fee and a redirect rule, and it worked, in the narrow sense that it got covered.

The more durable story, once the joke fades from memory, is the one covered elsewhere in this series: whether dots' permission model, Grok Bot's credential-sharing pitch, or Muse's Sentinel-gated sandbox actually earns the trust each company is racing to claim. That's a slower, less quotable race than a domain hijack — and the one that will actually determine who's still standing a year from now, a question this series returns to at its close.

Part 25 of 30 in the ChatGPT Dots series. Previously: Dots vs Grok Bot vs Muse: Reading the Permission Models Side by Side. Next: Will Dots Replace Claude Code Skills, or Just Sit Beside Them?. Browse more builder insights or explore AI skills at aiskill.market.