Elon Musk Tightens Grok Bot Limits Amid Heavy AI Spending by Meta, OpenAI

Elon Musk has tightened usage limits on xAI's Grok Bot, citing unsustainable operating costs relative to revenue. Subscribers on SuperGrok and Cursor plans reported sudden drops in their weekly token allowances. Musk acknowledged the expense of running personal AI assistants that handle complex background tasks, though he did not share specific cost or revenue figures. xAI is not alone in spending heavily on AI — Meta and OpenAI are also pouring large sums into the technology. Meta's Muse tool offers free tiers, but Grok Bot remains paywalled.

AI chatbots are steering clear of telling Americans how to vote in the 2026 midterms, but researchers say that does not mean they are neutral. MIT's LLM Election Observatory dashboard tracks how chatbots like ChatGPT, Claude, Gemini, and Grok discuss the elections. Only 8% of Americans say they use AI for election information. Google's Gemini gave incorrect or evasive answers about the 2020 election. Experts warn that framing and emphasis in AI responses can shape voter thinking even without outright misinformation.

DoorDash launched its drone delivery service called Air on September 30, with initial testing in Northern California. Chipotle Mexican Grill and Popeyes are participating partners. Early tests showed deliveries from restaurant to door completing in under five minutes on average. The push into physical AI extends beyond drones to warehouse robots and self-driving vehicles, creating demand for equipment makers, sensor suppliers, and battery manufacturers.

Palo Alto Networks introduced Cortex Cloud SaaS Security on November 2, 2026. The platform unifies posture, identity, data, and AI security across cloud environments. It provides visibility across 100 SaaS applications and uses agentic AI workflows through Cortex AgentiX to accelerate investigations and fixes. The move reflects growing demand for integrated security tools as enterprises adopt more SaaS platforms.

NVIDIA is preparing Kubernetes support for its upcoming Vera Rubin GPU hardware. The NVIDIA AI Cluster Runtime helps platform teams configure and verify AI factory setups. Vera Rubin requires new drivers, updated NIC and DPU components, and adjusted device topology settings. The AICR Vera Rubin recipes encode these changes so teams can reproduce validated configurations, with training and inference recipes verified end to end.

In other AI-related business news, Information Services Group posted Q3 revenue of $123.8 million, beating estimates, with AI-related revenue up 25 percent year-over-year. Keefe Bruyette cut price targets on seven bitcoin miners turned AI hosting companies, including Cipher Digital and CleanSpark, warning of project cost and financing risks. Separately, researchers introduced BABA-Bench, a benchmark testing whether AI agents can help patients navigate care, finding that even top-performing models like Claude Opus 5 often reached correct outcomes through invalid procedures.

Key Takeaways

  • Elon Musk imposed tighter weekly token limits on Grok Bot users due to unsustainable costs relative to revenue, with no specific figures disclosed.
  • MIT's LLM Election Observatory tracks how AI chatbots including Gemini, ChatGPT, Claude, and Grok discuss the 2026 midterms, finding none give direct voting advice.
  • DoorDash launched Air, a drone delivery service tested in Northern California with Chipotle and Popeyes, achieving under-five-minute delivery times.
  • Palo Alto Networks launched Cortex Cloud SaaS Security, offering visibility across 100 SaaS apps and agentic AI workflows via Cortex AgentiX.
  • NVIDIA is bringing Kubernetes support to Vera Rubin GPUs through its AI Cluster Runtime, with validated training and inference recipes.
  • Information Services Group reported Q3 revenue of $123.8 million with AI-related revenue up 25% year-over-year.
  • Keefe Bruyette cut price targets on seven bitcoin miners turned AI hosts, citing project cost and financing risks.
  • BABA-Bench found Claude Opus 5 reached correct outcomes in 81.6% of patient advocacy runs but only passed all checks in 52.5%.
  • Researchers found 28% of AI agents in the BABA-Bench failed to follow necessary steps despite reaching correct outcomes.
  • Palm Beach crash video confirmed authentic by AP, CBS News, CNN, and AI experts despite social media claims it was AI-generated.

Elon Musk limits Grok Bot usage after saying AI makes saving money pointless

Elon Musk said just days earlier that saving money is pointless with super intelligent abundance. On Sunday he said the cost of running Grok Bot was unsustainably high compared to revenue. xAI imposed tighter weekly token limits on Grok Bot users. Subscribers reported their usage allowances dropped suddenly. Grok Bot is available only to paying customers through SuperGrok and Cursor plans.

Musk says Grok Bot costs unsustainably high relative to revenue

Elon Musk said the cost of operating xAI's Grok chatbot is unsustainably high compared with revenue. He responded to a user asking for a reset of weekly usage limits after many subscribers ran out early. Musk did not share cost or revenue figures. He also did not say whether limits would change. Grok Bot launched in August as an AI agent that works across apps using its own virtual computer.

Elon Musk cuts users off from Grok Bot AI assistant over cost worries

Users of Grok Bot are running out of their weekly usage allowance and losing access. Elon Musk said the system is expensive and not earning enough money. Personal AI assistants use many resources because they handle complex tasks and run in the background. Companies like xAI, Meta, and OpenAI are spending large sums on AI. Some tools like Meta's Muse offer free tiers while Grok Bot is only available to paying users.

AI chatbots will not tell you how to vote but may shape your thinking

Only 8% of Americans say they use AI to learn about midterm elections. Google's Gemini gave incorrect or evasive answers about the 2020 election. AI companies avoid telling people how to vote to prevent legal problems. Experts say AI can still shape how people think about candidates and issues. Researchers tested ChatGPT, Claude, Gemini, and Grok and found none gave direct voting advice.

Q&A: Hey AI who should I vote for this year

MIT researchers built the LLM Election Observatory dashboard to track how AI chatbots discuss the 2026 midterm elections. The system simulates how people across the country talk with chatbots and tracks responses over time. It watches how answers change based on location, political party, and identity. The goal is to understand how AI shapes political discussion and voter understanding. Researchers say framing and emphasis can matter even without outright misinformation.

Palm Beach crash video is real not AI generated

Social media users claimed video of a November 1 plane crash in Palm Beach County, Florida was AI-generated. News outlets including AP, CBS News and CNN confirmed the footage is authentic dash cam video. An AI expert told PolitiFact the video is real despite its dramatic appearance. A business jet crashed in a neighborhood near Cresthaven, Florida, killing everyone on board. The crash is under investigation as authorities examine turbulence and ice as possible factors.

DoorDash launches drone delivery service Air

DoorDash unveiled DoorDash Air, a drone delivery system, on September 30 with tests in Northern California. Partners include Chipotle Mexican Grill and Popeyes. Initial tests showed deliveries from restaurant to door in under five minutes on average. The article discusses how drone delivery creates demand for equipment makers, sensor suppliers and battery manufacturers. The concept of physical AI includes warehouse robots, self-driving vehicles and delivery drones.

BABA-Bench tests patient advocacy AI agents

Researchers introduced BABA-Bench, a benchmark with 47 interactive worlds based on real patient advocacy cases. The study tested 14 frontier models across 1,974 runs to see if AI agents could help patients navigate care. Agents often reached correct outcomes through invalid procedures, with 28 percent failing to follow necessary steps. Claude Opus 5 reached all required outcomes in 81.6 percent of runs but only passed all checks in 52.5 percent of runs. Providing verified routes improved pass rates significantly.

Review AI coding assistant config files carefully

A pull request can add harmful setup steps for AI coding assistants without changing application source code. StepSecurity reported a commit that used a skip-CI flag and backdated timestamp to move code execution from package install to folder open. The article recommends giving configuration changes a named owner and enforceable review path. Reviewers should trace commands to their scripts and check for network access or credential use. GitHub and Anthropic provide mitigations including bare mode runs and managed-hooks-only settings.

Information Services Group Q3 results meet estimates

Information Services Group reported Q3 revenue of $123.8 million, beating the $122.8 million estimate. Net income rose to $11.4 million or $0.23 per share from $10.4 million or $0.21 per share last year. Operating margin expanded to 14.1 percent from 12.5 percent due to AI demand, with AI-related revenue up 25 percent year-over-year. International revenue grew 15 percent while Americas revenue was soft. Q4 guidance projects revenue between $120 million and $130 million with net income between $10 million and $12 million.

Keefe Bruyette cuts price targets on bitcoin miners turned AI hosts

Keefe Bruyette lowered price targets on seven bitcoin miners and AI hosting companies on Monday. Cipher Digital had the biggest cut, falling from $28 to $11. CleanSpark also saw its target drop by more than 50 percent. The firm said these stocks trade near or below their contracted lease values. It warned that project costs and financing risks remain a concern for investors.

Palo Alto Networks launches Cortex Cloud SaaS Security

Palo Alto Networks announced Cortex Cloud SaaS Security on November 2, 2026. The tool brings SaaS security into one cloud security experience. It unifies posture, identity, data and AI security across SaaS and cloud environments. It gives visibility across 100 SaaS applications. It also uses agentic AI workflows through Cortex AgentiX to speed up investigations and fixes.

NVIDIA brings Kubernetes support to Vera Rubin GPU hardware

A new NVIDIA blog explains how to prepare Kubernetes for NVIDIA Vera Rubin GPUs. The NVIDIA AI Cluster Runtime helps platform teams configure and verify AI factory setups. Vera Rubin needs new drivers, updated NIC and DPU components, and different device topology settings. The AICR Vera Rubin recipes encode these changes so teams can reproduce validated configurations. Training and inference recipes have been validated end to end.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

Elon Musk xAI Grok Bot AI Costs AI Revenue AI Chatbots Elections DoorDash Drone Delivery Palo Alto Networks Cortex Cloud SaaS Security NVIDIA Vera Rubin GPU Kubernetes AI Cluster Runtime Information Services Group AI Revenue Growth Bitcoin Miners AI Hosting Companies BABA-Bench

Comments

Loading...