The Information AM

Save 25% on an annual subscription to read the most important news about technology and business first. For even more, Save $250 on The Information Pro for unlimited access to our proprietary org charts, databases and surveys.

1.	Moonshot AI’s New Kimi K3 Challenges U.S. Frontier Models By Juro Osawa Source: The Information

Chinese startup Moonshot AI’s latest large language model is sending shockwaves through the global AI industry because of its frontier-level coding capabilities on some benchmarks. The new model, Kimi K3, is stirring debates whether the gap between Chinese open-source and U.S. frontier models labs is fast diminishing or even being closed.

K3, unveiled on Thursday, is the world’s largest open-source model on the market, with 2.8 trillion parameters. The model now ranks No. 1 on the Frontend Code Arena coding leaderboard, above Anthropic’s most powerful model, Claude Fable 5. K3 is suitable for handling tasks that involve reasoning and knowledge work, as well as complex coding tasks that require continuous engineering for many hours, Moonshot said.

Other leaderboards show that K3 approaches the most advanced U.S. models, but doesn’t quite match them. On Artificial Analysis Intelligence Index, a comprehensive benchmark for AI models’ capabilities across maths, science, coding and reasoning, K3 ranks No. 3, behind Fable 5 and OpenAI’s GPT-5.6 Sol model, but ahead of Claude Opus 4.8, which was released in May.

Earlier this year, Anthropic accused Moonshot and other Chinese AI labs of illicitly distilling Claude models to train and improve their own models. But if K3’s performance actually rivals top Anthropic models as some benchmarks suggest, distillation may no longer explain the rapid advancements of Chinese models.

Moonshot’s launch of K3 accelerates the momentum of China’s open-source AI models, which are expanding rapidly by offering increasingly strong performance at much lower prices compared to U.S. proprietary models from Anthropic and OpenAI. Among other Chinese competitors, Beijing-based Zhipu AI’s GLM-5.2 model, released last month, has also gained popularity among developers around the world.

2.	Exclusive: Microsoft Preps Mythos-Like AI Bug Finder By Aaron Holmes Source: The Information

Microsoft is preparing to release a new AI security product, internally codenamed Project Perception, to capture a piece of companies’ rising cyber defense spending. Set to debut as soon as this month, the product will use a combination of AI models from Anthropic, OpenAI, and Microsoft itself to sniff out software bugs similar to the way Anthropic’s Mythos model does and automatically fix the vulnerabilities, according to someone with direct knowledge of the plans.

While Project Perception’s prices haven’t been finalized, Microsoft hopes it can give customers a cost-effective alternative to Anthropic’s Mythos—which is exceptionally expensive—according to the person with knowledge of the plans. Microsoft plans to use a so-called model router that switches between OpenAI, Anthropic, and Microsoft’s models, depending on the type of task, to keep costs lower than if the product could only use one model.

In recent months, Anthropic’s development of AI with strong cybersecurity capabilities prompted businesses to seek to protect themselves from hackers that use their own AI to find software exploits. Cybersecurity executives say they’re spending heavily on such tools to reduce the need for human security analysts that traditionally kept tabs on companies’ IT systems to detect and respond to potential hacks.

Project Perception is part of an effort by Microsoft’s new security chief, Hayete Gallot, to reinvigorate its security sales and put it on better footing to compete with the likes of Anthropic and OpenAI. Microsoft is the world’s biggest seller of such software, and Gallot, who took over the security org in February, recently overhauled the unit to focus more on AI tools while making cuts to older products.

3.	Chinese Leader Xi Calls for “Open Source, Open Collaboration” in Shanghai AI Conference By Jing Yang Source: The Information

Chinese leader Xi Jinping on Friday called for “open source and open collaboration” in global AI development during a keynote speech at a government-hosted AI conference in Shanghai. “We should encourage inclusiveness and promote mutual learning among civilizations,” Xi said, according to a transcript published by the Chinese Ministry of Foreign Affairs. The Chinese leader also cautioned against AI safety risks and said all countries should work together “to build a just and equitable global AI governance system.” “Artificial intelligence should be a trustworthy tool for humanity,” he said at the World Artificial Intelligence Conference.

This is the first time that Xi attended and spoke at the conference, now in its 9th year. Xi’s presence further elevated the development of AI in Beijing’s strategic priorities. The annual gathering, which has grown rapidly in the past few years with more exhibitors and delegates, serves as an important indicator of China’s innovation. Companies and startups use the event to showcase and launch their products, from humanoid robots to new AI models and chip design architectures.

Xi said China will provide developing countries with 5,000 opportunities for AI training and seminars over the next five years,  including in Southeast Asia, the Middle East, Africa and Latin America.

4.	Meta Plans To Hire Top AWS Executive By Jyoti Mann Source: The Wall Street Journal

Meta Platforms plans to hire Dave Brown, a senior executive at Amazon Web Services, The Wall Street Journal reported, citing people familiar with the matter.

AWS CEO Matt Garman told employees on Wednesday that Brown is leaving the company for a “new role outside the company,” The Information earlier reported.

Brown will report to Meta’s head of global infrastructure, Santosh Janardhan, and will help oversee the company’s expanding data center footprint, according to the Journal.

The move comes as Meta accelerates its AI infrastructure investments and explores plans to build a cloud business that could rent out excess computing capacity. On Monday, the company said it will invest an additional $40 billion to more than double the computing capacity of its Louisiana data center to five gigawatts.

The expansion is part of a broader AI infrastructure push led by CEO Mark Zuckerberg, who has pledged to invest hundreds of billions of dollars over the coming years.

5.	Uber to Buy Germany’s Delivery Hero For Around $14 Billion By Martin Peers Source: The Information

Uber has agreed to buy all of German food-delivery firm Delivery Hero in a deal that values the company at $14.8 billion. But Uber won’t have to pay all of that as it already owns about 25%, including a big stake bought in May.

The acquisition sharply expands Uber’s food-delivery presence globally, as Delivery Hero has operations in Asia, Latin America as well as Asia, although it will sell its operations in some markets where it competes with Uber. Delivery Hero had $17.2 billion in revenue last year, which is about the same as Uber’s food-delivery revenue for 2025.

Uber said the deal would increase the number of markets where it offers both car services and delivery to 58 markets, from 34. The deal reflects a growing interest by U.S. based delivery firms in Europe. Uber’s main rival in food delivery, DoorDash, has made two big acquisitions in Europe, most recently of Deliveroo.