AI News Of The Week (10th April, 2026)
TLDR: Key AI Developments This Week
Anthropic launched Project Glasswing, giving select partners access to Claude Mythos Preview for defensive cybersecurity work backed by $100 million in usage credits. Google and Broadcom signed a long-term agreement through 2031 for custom AI chips while securing multi-gigawatt TPU capacity for Anthropic. Meta unveiled Muse Spark, its first model from the rebuilt superintelligence effort designed to power a faster Meta AI across all platforms. Google transformed Gemini beyond chat with notebooks syncing to NotebookLM and interactive visualizations. OpenAI revealed enterprise now makes up over 40% of revenue and is on track for parity with consumer by year-end, while pausing its UK data center project. The EU began assessing whether ChatGPT should face stricter DSA platform regulations.
Anthropic Turns Claude Mythos Into a Locked-Down Cybersecurity Push With Project Glasswing
On April 7, Anthropic launched Project Glasswing, a new security initiative that gives a select group of partners access to Claude Mythos Preview for defensive cybersecurity work instead of releasing the model broadly. Anthropic says Mythos Preview has already found thousands of high-severity vulnerabilities, including flaws in major operating systems and browsers, and is backing the effort with up to $100 million in usage credits plus $4 million in donations to open-source security groups. Read more
Google and Broadcom Lock In a Long-Term TPU Bet as Anthropic Secures Multi-Gigawatt Compute
One of the biggest infrastructure stories of the week landed on April 6, when Broadcom said it signed a long-term agreement with Google to develop and supply future generations of custom AI chips and rack components through 2031. The same announcement also tied Anthropic even more tightly into Google's TPU ecosystem, with multiple gigawatts of next-generation capacity expected from 2027 onward, underscoring how the AI race is becoming just as much about securing silicon and power as shipping better models. Read more
Meta Launches Muse Spark and Signals It Wants Back Into the Frontier Race
Meta used April 8 to unveil Muse Spark, the first model from its rebuilt superintelligence effort and the first in a new Muse series designed to power a faster, more capable Meta AI. The company says it currently powers the Meta AI app and website and will roll out to WhatsApp, Instagram, Facebook, Messenger, and AI glasses in the coming weeks, while Reuters described the launch as Meta's clearest attempt yet to recover momentum after disappointment around Llama 4. Read more
Google Pushes Gemini Beyond Chat and More Toward a Real Project Workspace
Google spent this week turning Gemini into something that feels less like a chatbot and more like a working environment. On April 8 it introduced notebooks in Gemini, syncing chats and files with NotebookLM as shared knowledge bases, and on April 9 it began rolling out interactive visualizations inside Gemini so users can generate simulations and explore concepts directly in chat instead of just reading static text responses. Read more
OpenAI Says Enterprise AI Is No Longer a Side Business
OpenAI's April 8 enterprise update made it clear where the company thinks the market is heading. According to OpenAI, enterprise now makes up more than 40% of its revenue and is on track to reach parity with consumer by the end of 2026, while Codex has hit 3 million weekly active users and the company's APIs are processing more than 15 billion tokens per minute. The broader message was that OpenAI wants to be the intelligence layer underneath company-wide agents, not just the maker of a popular assistant. Read more
OpenAI's UK Stargate Project Hits Pause Over Regulation and Energy Costs
Not all the week's AI news was about launches. Reuters reported on April 9 that OpenAI is pausing its main UK data centre project, saying it will revisit the plan when regulation and energy economics are more favorable for long-term investment. That makes the week a reminder that AI expansion is increasingly constrained by real-world bottlenecks like power prices, permitting, and political climate, not just model progress. Read more
The EU Starts Weighing Whether ChatGPT Belongs Under the DSA's Big-Platform Rules
On April 10, Reuters reported that the European Commission is assessing whether ChatGPT should be designated a large online platform under the Digital Services Act after OpenAI published user figures above the relevant threshold. If that happens, ChatGPT would face a tighter regulatory regime in Europe, which would make this one of the clearest signs yet that AI assistants are beginning to be treated less like experimental tools and more like major internet platforms. Read more
