Musk Says Grok 4.6 Will Drop Around August 7, with Grok 4.7 Coming a Few Weeks Later
On July 28, local time, Elon Musk announced that xAI’s next-generation large model, Grok 4.6, is expected to launch around August 7. This model will pack a whopping 1.5 trillion parameters and bring serious upgrades in supervised fine-tuning (SFT) and reinforcement learning (RL). Musk added that Grok 4.7 will follow just weeks after, boasting 2.1 trillion parameters. Compared to Grok 4.6, Grok 4.7 will push the envelope even further in every aspect—though it’ll be a tad slower in service speed, it’ll make much better use of tokens.
SK Hynix: HBM4 Production to Ramp Up in Second Half of Year
On July 29, SK Hynix revealed in its earnings report that HBM4 has already started shipping in volume during the second quarter, and production will be scaled up in the second half of the year. Samples of HBM4E were delivered in the first half, using an optimized process that balances technical maturity with mass production stability. SK Hynix says it will keep cementing its leadership in HBM by leveraging stable supply capabilities backed by high quality and yield, cost competitiveness, and top-tier performance in the industry.
Artium AI Becomes OpenAI Advanced Partner, Helping Enterprises Build Custom AI Agent Solutions
On July 29, US-based AI agent software consultancy Artium AI announced it has joined the OpenAI partner network as an Advanced Partner. The company said it will work with OpenAI to help enterprises design, deploy, and scale custom AI agent solutions across industries like finance, healthcare, tech, telecom, media, and entertainment.
Kimi by Moonshot AI Closes Series F Round at Over $3.5 Billion
According to a July 29 report by the Science and Technology Innovation Board Daily, Moonshot AI’s Kimi has completed its Series F funding round, raising over $3.5 billion and hitting a post-money valuation of $35 billion. The round closed early because it overshot its target by more than three times. The originally planned Series G (Pre-IPO) round for August has already kicked off earlier, with a pre-money valuation of $50 billion.
Zuckerberg Opposes US Ban on Chinese AI Models
According to CCTV Finance, the Financial Times reported on July 28 that Meta CEO Mark Zuckerberg said in an interview that the US shouldn’t ban Chinese AI models just to gain an edge in the AI race. He argued that banning advanced Chinese AI models “isn’t an effective solution.” Instead, US companies should “systematically” identify bottlenecks and obstacles in development. Both Jensen Huang and Elon Musk have previously expressed similar views against a US ban on Chinese AI models.
In response to US claims about investigating and sanctioning Chinese AI companies, a Chinese Ministry of Commerce spokesperson said on the 27th that China has always opposed the US politicizing and weaponizing tech and trade issues, smearing Chinese companies with baseless accusations, and imposing sanctions.
Andrew Ng Founds AI Education Company LearnVector, Gets $100 Million Investment from Coursera
On July 28, local time, Coursera co-founder Andrew Ng announced the launch of LearnVector, an AI company focused on delivering one-on-one learning experiences, and he’ll serve as its CEO.
LearnVector revealed it has secured a $100 million strategic investment from Coursera, with products expected to debut early next year.
Over 1,000 Employees from OpenAI, Google, Anthropic Sign Petition Urging US Government Action
On July 28, local time, 1,134 executives and employees from leading AI companies including OpenAI, Google, and Anthropic issued a joint open letter urging the US government to support an international effort to develop the necessary technical and governance tools to consciously control the pace of automated AI development.
The letter states that to fully harness AI’s potential, industries, governments, and even society as a whole may need to buy time to address emerging risks, establish safety measures, and strengthen regulations. However, every company and every country faces immense competitive pressure and is reluctant to unilaterally slow down this acceleration. Right now, the world lacks the technical and governance tools needed to consciously control the speed of frontier development.
