Main by morning: Anthropic decided to create chips, US fears cyber risks of AI models

Anthropic is exploring the possibility of building its own AI chips amid a shortage of computing power, Reuters has learned / Photo: Photo For Everything / Shutterstock
SpaceX posted a loss of about $5 billion before its IPO due to xAI investments and integration, but expects a valuation of over $1.75 trillion and the development of orbital data centers for AI. Anthropic, amid a shortage of computing power, is exploring building its own chips while growing revenue to more than $30 billion and strengthening partnerships with Google and Broadcom. These and other topics are covered in our roundup of key events for the morning of April 10.
Meta moves engineers to AI division ahead of layoffs
Meta is forming a new AI division Applied AI Engineering, transferring the strongest developers there not voluntarily, but on a mandatory basis, Reuters reports. The reorganization is taking place against the backdrop of planned large-scale layoffs and increased investment in AI infrastructure.
The goal is to create AI agents capable of writing code and doing most of the development, testing and launching of products, while employees will only control the processes. According to Mark Zuckerberg, 2026 will be a watershed year when AI will fundamentally change the way companies operate and increase their efficiency.
Anthropic is considering developing its own AI chips
Anthropic is exploring the possibility of building its own chips amid a shortage of computing power for AI, but the project is at an early stage and the company may abandon the idea in favor of purchasing from partners, Reuters writes. It now uses solutions from Google and Amazon, and recently entered into a long-term agreement with Google and Broadcom.
Demand for Anthropic Claude's AI model has skyrocketed, with year-to-date revenue exceeding $30 billion versus $9 billion at the end of 2025, the agency notes. Developing its own chips could cost about $500 million, industry sources tell Reuters; other major players including Meta and OpenAI are already considering similar projects.
SpaceX took a $5 billion loss before its IPO
Elon Musk's SpaceX recorded a loss of nearly $5 billion in 2025 on revenue of more than $18.5 billion, with the figure including the results of the startup xAI, which was acquired in February, The Information writes. The company was previously profitable and was valued much higher by investors amid growth in the space business.
The company, which filed for an IPO in the U.S., could be valued at more than $1.75 trillion. SpaceX remains the world's largest launch operator and plans to develop orbital data centers for AI and interplanetary flight projects.
U.S. Treasury and Fed warned banks of cyber risks due to Anthropic AI
U.S. Treasury Department Chief Scott Bessent and Federal Reserve Chairman Jerome Powell held an emergency meeting with bank executives, warning of cyber risks associated with new AI models, including Anthropic's Mythos, Bloomberg reports. The discussion took place in Washington and was aimed at improving financial organizations' preparedness for potential threats.
Authorities have urged banks to strengthen system defenses and assess vulnerabilities amid the rapid development of AI, the agency specified.
What's in the markets
- Japan's broad Topix index was little changed on April 10, with the Nikkei 225 rising 1.9 percent.
- Hong Kong's Hang Seng Index was up 0.6 percent, while mainland China's CSI 300 Index was up 1.6 percent.
- In South Korea, the Kospi index was up 1.7% and the Kosdaq was up 1.8%.
- Australia's S&P/ASX 200 was falling 0.3 percent.
- Futures on the S&P 500 and futures on the Nasdaq Composite were almost unchanged. Exchange contracts on the Dow Jones Industrial Average were falling by 0.1%.
This article was AI-translated and verified by a human editor
