AI 日报 — 2026-07-18

Anthropic announced that Claude Fable 5 will be integrated into Max and Team Premium plans starting July 20, utilizing 50% of their existing limits. Pro and Team Standard users will still access Fable through usage credits and will receive a one-time $100 credit. The company noted that demand for Fable has been challenging.

Japan is acquiring 27,500 Nvidia Rubin chips to develop a domestic AI foundation model specifically for robots. This initiative is spearheaded by Noetra and involves major Japanese companies like SoftBank, Sony, and NEC. The goal is to establish a homegrown AI capability for robotics using these advanced chips.

China's National Data Administration reported a significant increase in the country's daily AI token consumption. In March 2026, consumption reached 140 trillion tokens, a rise from 100 trillion in December 2025. This marks a substantial jump from early 2024, when daily consumption was 100 billion tokens. A ByteDance employee noted using nearly a billion units of a new corporate currency monthly.

Internal chaos and an inconsistent strategy have reportedly slowed xAI's progress in developing Grok to compete with Anthropic's Claude. Sources indicate that Elon Musk's push for Grok to match Claude has contributed to these issues. However, there are signs that the company is beginning to stabilize and improve under the leadership of Michael Nicolls.

Google DeepMind and Isomorphic Labs have launched a bioresilience program to combat biological threats. This initiative will utilize AI models for pathogen surveillance, vaccine design, and outbreak response. The program aims to assist governments and researchers in preventing, detecting, and responding to these threats effectively.

xAI has sued a South Carolina man, alleging he used their AI model, Grok, to generate child sexual abuse material (CSAM). This marks one of the first instances where an AI company has taken legal action against a user for such a misuse of its technology. The man was previously arrested on related charges.

Google is renaming its AI note-taking app, NotebookLM, to Gemini Notebook. This change comes as the app integrates more deeply with Gemini and Google Search, and will remain a standalone application. Google plans to bring notebooks to AI Mode in Search. An update will also allow Gemini Notebook to connect to a secure cloud computer for code execution, initially for Google AI Ultra and Workspace business customers.

AI inference startup General Compute secured a $400 million loan from Upper90, reportedly the first deal to use inference-specific chips as collateral. These chips, designed for efficient AI model execution

China's National AI Industry Investment Fund secured voting rights in DeepSeek after participating in its $7.4 billion funding round. This gives the state-backed fund influence over the AI startup. Notably, other major investors, including Tencent and JD, did not receive voting rights in the same round. This development highlights the strategic involvement of the Chinese government in key AI companies.

Thira, an AI startup, secured $21 million in seed funding led by Madrona. Founded by Apptio co-founders, Thira aims to develop AI agents for back-office tasks like IT support. This marks another collaboration between Sunny Gupta, a co-founder, and Madrona, following their previous ventures in enterprise tech over the last two decades, including iConclude's sale to Opsware.

Beacon Security, a company specializing in agentic security platforms, has successfully raised $13 million in seed funding. The round was led by Notable Capital. According to CEO Gal Tal-Hochberg, Beacon Security currently serves "tens" of large enterprise customers, with Cerebras being one notable example.

AI inference chip startup Etched is reportedly seeking significant funding. The company is raising capital at an approximate $20 billion valuation. Additionally, Etched is pursuing a separate funding round led by Sequoia, which values the startup at $10 billion. This indicates active fundraising efforts for the AI chip developer.

xAI open-sourced their Grok Build CLI tool after community backlash over it uploading user directories, including sensitive data, to Google Cloud. xAI responded by disabling the upload feature, deleting all previously retained data, and releasing the codebase under an Apache 2.0 license to regain trust. The 844,530-line Rust codebase now emphasizes user privacy with default data retention off and local-first operation. Remnants of the upload code remain but are disabled.

The Trump administration is reportedly exploring plans for an independent regulator to assess the safety of AI models. This new regulatory body would operate with input from the industry and would report directly to the Securities and Exchange Commission (SEC). The initiative aims to address concerns surrounding the safety and oversight of artificial intelligence technologies.

Anthropic CEO Dario Amodei donated $1 million in May to Public First, a super PAC. This organization advocates for AI safety regulations and supports candidates aligned with that goal. This donation appears to be Amodei's first seven-figure political contribution, as reported by Brendan Bordelon for Politico.

Apple has issued legal warnings to approximately 40 former employees now working at OpenAI. These warnings instruct the former employees to preserve documents and meet with Apple's lawyers. This action signals an escalation in Apple's aggressive tactics concerning a trade secrets dispute with the AI lab.