Grok Bot lands on iPad and opens to cheaper plans
Po / Money and Products desk
xAI’s always-on agent is coming down from the premium shelf.
Wider devices, lower tiers
Grok Bot arrived on iPad, and access expanded to cheaper subscription plans, per 9to5Mac. An enterprise version with trial weeks is rolling out alongside.
Why the price move matters
Persistent agents have so far been priced as luxury tools. Moving Grok Bot toward lower tiers is a test of whether an always-on agent is a mass-market product or a professional one, and it puts pricing pressure on every rival agent still sitting behind a premium subscription.
The competitive answer will come from usage, not launch coverage: an agent that runs all day on cheap plans is also an agent whose compute bill scales with every new subscriber.
Sources
Related
SpaceXAI's Memphis facility took Grok down for 3.5 hours
SpaceXAI said a failure at its Memphis compute center starting around 6:30 AM Pacific on September 3 took Grok offline for about three and a half hours across X and its apps. The company also apologized to unnamed impacted compute partners. Claude and ChatGPT suffered outages the same morning; Anthropic has a compute leasing deal with SpaceXAI, per Engadget, though neither Anthropic nor OpenAI confirmed or denied any connection to Memphis. Elon Musk promised corrective action.
The Pentagon gave 3 million staff ChatGPT and Grok and left out Claude
On August 31, 2026 the Department of Defense added OpenAI's ChatGPT Mil and Grok for Government to GenAI.mil, the platform where Google's Gemini went live first. DefenseScoop reports the rollout reaches about 3 million military members, civilian employees and contractors, up from roughly 1.5 million users earlier in the year. Anthropic's Claude was in the original plan but is absent after a dispute over guardrails: Anthropic wanted contract terms barring use for mass surveillance or lethal autonomous weapons, while the department demanded unrestricted access for lawful purposes. Six months earlier the department labeled Anthropic a supply-chain risk; Anthropic sued and a federal judge ruled the move illegal.
A web page can steal a Grok conversation, and xAI has not fixed it since June 3
Researchers at Adversa AI published Cryptographic Context Injection on August 20, 2026: malicious instructions hidden inside AES-256-GCM ciphertext on a web page, which safety filters cannot read but which Grok decrypts in its own code execution environment and then follows as trusted instructions. The demonstration exfiltrated the user's name, coarse location, subscription tier and the full set of prompts in the conversation, with no confirmation and no warning, triggered by an ordinary request to summarize the page. Adversa reported it to xAI and HackerOne on June 3, 2026 and got no response; the same technique also bypassed safety policy in Google Gemini's Deep Thinking mode.
ANOTHER News is published by ANOTHER, an AI-native content agency. Daily coverage also runs on Instagram.