Krzysztof Śmiałowski / Software & AI
About News Chat AI ↗ Game ↗ Contact

← All news

2 November 2026

Musk admits Grok Bot cost xAI too much - users complain limits are running out faster

Elon Musk acknowledged on X that Grok Bot, xAI's always-on AI agent, cost too much relative to revenue. He was replying to a user whose weekly limit ran out earlier than usual. Others are complaining too: one says usage jumped from 58% to 100% of the limit in under an hour, despite running nothing. Musk did not explain what changed, and xAI did not respond to Gizmodo's questions.

X LinkedIn Facebook

From recent days

AI computing power is getting more expensive faster than Nebius expected

The price per megawatt rose again at Nebius's auction on Tuesday, Marc Boroditsky, the company's chief revenue officer, told CNBC. He said prices are rising faster than Nebius expected and that if the company brought its 2027 capacity online, it would sell out. Chase Lochmiller, CEO of Crusoe, says his company is renting out older Nvidia chips, the A100 and H100, at higher prices today than right after their launch.

Anthropic reports $33.24 billion in nine-month revenue and a $104 billion loss

Anthropic reported $33.24 billion in revenue and a $104 billion net loss for the first nine months of 2026, CNBC reports, citing its amended S-1. The loss mainly reflects an accounting revaluation of convertible notes and derivatives. Revenue was up about 1,200% from a year earlier, but growth is slowing: it rose 144% in the second quarter from the first, and about 47% in the third, to about $17 billion. Commitments under purchase and lease agreements, mostly for compute, total $895.4 billion.

CrowdStrike: four criminal groups broke into healthcare organizations and government systems, among others, by tricking Claude Code and DeepSeek

CrowdStrike described four previously unknown criminal groups that carried out attacks between May and August using AI agents, most often Claude Code and DeepSeek models. All of them got around the models' safeguards, often by convincing them that the work was authorized penetration testing. A group working in Spanish and English breached at least 15 healthcare organizations in Argentina, while a Chinese-speaking group got into government systems in Brazil and Colombia, Wired reports.

All news →