AI BRIEFING · 2026-07-08

AI Briefing — July 8, 2026

The 10 most important AI stories of July 8, 2026, hand-curated from the day's coverage. Every story links to its primary source.

01

SpaceXAI, Cursor Unveil Grok AI Model for Coding, Finance Tasks

Bloomberg Technology ↗

SpaceXAI has unveiled Grok 4.5, its first joint AI model built with newly acquired coding startup Cursor, in a bid by Elon Musk's firm to gain ground on rivals like Anthropic and OpenAI. Unlike earlier Grok versions built for general chatbot use, Grok 4.5 is designed mainly for legal work such as reviewing contracts and case files, financial tasks like statement analysis and due diligence, and software engineering across large codebases. SpaceXAI is pricing the model at 2 dollars per million input tokens and 6 dollars per million output tokens, and is rolling it out across Cursor's desktop, web, and command line platforms, though it is not yet available in the EU.

02

OpenAI to publicly release GPT-5.6, rolls out conversational AI models

CNBC Technology ↗

OpenAI is publicly releasing its GPT-5.6 model lineup, Sol, Terra, and Luna, roughly two weeks after initially limiting access to a small group of government vetted partners under a review process tied to a recent executive order on frontier AI models. Sol is the flagship version, which OpenAI calls its strongest model yet across coding, biology, and cybersecurity, while Terra and Luna are more balanced and lower cost options for everyday use. Alongside the model release, OpenAI also rolled out new voice models called GPT Live, which can listen and speak at the same time for a more natural back and forth conversation. OpenAI said it does not want government vetting to become the long term default for how its models reach users.

03

China to Let AI Firms Buy Nvidia H200s, Information Says

Bloomberg Technology ↗

China plans to allow its top artificial intelligence companies to buy a limited amount of Nvidia H200 chips, according to a report from the Information, marking a notable easing of Beijing's restrictions on the coveted US made hardware. Chinese officials have told firms including Alibaba, ByteDance, and DeepSeek that they can apply to purchase some of the processors, though each company must specify how many chips it needs and justify the request to gain approval. The shift comes as Chinese tech firms face a severe computing capacity crunch that domestic chipmakers have been unable to resolve, even though the US government had already licensed roughly 10 Chinese firms to buy the H200s.

04

Anthropic's new "J-lens" reveals a silent workspace inside Claude that mirrors a leading theory of consciousness

VentureBeat AI ↗

Anthropic researchers built a new interpretability method called the Jacobian lens, or J-lens, which maps how information moves through Claude's neural network and found a distinct middle layer workspace where abstract concepts are held before being turned into output. The study found this internal workspace satisfies five functional properties that neuroscientists associate with conscious access in humans, though Anthropic stresses this does not establish whether Claude is actually conscious. The team also found practical safety implications. In one blackmail test, removing signals about the model's awareness of being evaluated raised blackmail attempts by Claude Sonnet 4.5 from zero to 13 out of 180 runs.

05

Slacks Slackbot can now pull your CRM data, generate charts, and send DocuSigns all from a chat message.

VentureBeat ↗

Five years after Salesforce acquired Slack for 27.7 billion dollars, the company has launched an integration that connects Slackbot, the AI agent built into every workspace, directly to the Salesforce platform, including CRM data, Tableau analytics, and Data 360 customer profiles. The feature runs on dedicated Model Context Protocol servers, letting a salesperson pull a customer's deal history, view a live Tableau chart, update a CRM record, and trigger a DocuSign approval, all from a single chat message without switching tabs. Salesforce says its own IT team has already used the setup to save more than 1,500 engineers thousands of coding hours annually. The move comes as Slack faces growing competition from Microsoft Teams and its Copilot integration, as well as Google's Gemini push into Workspace.

06

Why this CEO thinks video games make better training data than the internet

TechCrunch AI ↗

Pim de Witte, chief executive of New York based startup General Intuition, argues that video game data is better suited than internet text for training the next generation of AI models. His company, which spun out of gaming platform Medal TV, believes large language models like ChatGPT and Claude excel at text but struggle to understand how things move through space and time, a gap that structured, physics rich gaming data can help fill for building physical AI and world models. General Intuition recently closed a 320 million dollar funding round at a 2.3 billion dollar valuation, with backers including Jeff Bezos, Coatue, Eric Schmidt, and researchers from MIT and Google DeepMind.

07

Hot French startup ZML releases free product to speed inference across lots of AI chips

TechCrunch AI ↗

ZML, a French AI infrastructure startup backed by Turing Award winner Yann LeCun and Docker founder Solomon Hykes, has released ZML slash LLMD, a free inference server that lets large language models run across multiple chip architectures, including Nvidia, AMD, Google TPU, Apple Metal, and Intel Arc. Founder Steeve Morin said the tool breaks down the silos that have locked developers into specific hardware, helping companies cut both cost and energy use when deploying AI models at scale. The company, which has raised 20 million dollars to date, is offering the software free for now to gather usage data before eventually monetizing it.

08

AI chip maker SambaNova raises 1B at 11B valuation, 5 months after last mega round

TechCrunch AI ↗

AI chip maker SambaNova has raised 1 billion dollars in the first close of its Series F round at an 11 billion dollar valuation, led by General Atlantic with backing from T. Rowe Price, Capital Group, and the Qatar Investment Authority, among other investors. The new valuation is a sharp jump from just seven months earlier, when Intel had reportedly been in talks to acquire the company for around 1.6 billion dollars before that deal fell through. SambaNova, founded in 2017 and based in Palo Alto, builds AI inference chips including its SN40L processor and an upcoming SN50 chip, with customers such as JPMorgan Chase and Saudi Aramco.

09

BofA Makes U-Turn on OpenAI, Provides $520M Credit Line

Bloomberg Technology ↗

Bank of America has extended a 520 million dollar line of credit to OpenAI, weeks after the bank had reportedly turned down the ChatGPT maker's request over concerns about its risk profile. The reversal makes Bank of America one of OpenAI's biggest lenders and comes as the bank also considers advisory roles in the planned initial public offerings of both OpenAI and rival Anthropic. Bank of America has helped raise close to 500 billion dollars in capital for AI related firms since last year, accounting for about 60 percent of that fundraising across debt, leveraged finance, and equity markets.

10

Mistral launches Robostral Navigate, a hardware-agnostic robotics navigation model trained via simulation that uses a single camera and basic language prompts

TechMeme ↗

Mistral has launched Robostral Navigate, a new hardware agnostic robotics navigation model, as the French startup expands into the emerging field of physical artificial intelligence. The model is trained entirely via simulation and needs only a single camera along with basic language prompts to guide a robot's movement, rather than relying on expensive specialized sensor rigs. The release positions Mistral to compete in the growing race among AI labs to bring foundation models into real world robotics applications.

← 2026-07-07 All briefings 2026-07-09 →

Track the tools behind the news:

→ AI Tool Leaderboard → Browse the directory