Each week, we bring you 5 stories that resonated the most in our internal Slack channel #AI-news. We write the newsletter using various AI tools because we're an AI company and our marketing wants to move with the times too. 😎
Today you're reading the 76th issue in a row.
#1
Anthropic overtakes OpenAI 🧠: Claude dominates the enterprise AI market 📊
Anthropic currently leads the enterprise AI market 🤖 with a 32% share, just ahead of OpenAI, which has dropped to 29%. Just two years ago, OpenAI held a dominant 50% share, but its lead has eroded due to rising competition 🥊 and internal issues 🔧. Companies are increasingly choosing Anthropic’s Claude models for their reliability ✅, built-in safety features, and strong performance in areas like programming 👨💻 and data analysis 📊.
AI spending has doubled 💸 in the past six months, reaching $8.4 billion 💰. Anthropic is capturing a growing portion of that expansion—especially in coding, where it now holds 42% of usage 🧑💻. In regulated sectors like finance and healthcare 🏥, its emphasis on control and compliance gives it a clear edge over less predictable competing models ⚖️.
Anthropic’s growth also reflects a broader shift in corporate priorities 🔄: instead of chasing hype 🎉, companies are prioritizing stability 🧱, transparency, and robust tools ready for real-world deployment 🚀. If agentic AI continues to expand, Claude models could pull even further ahead 🏁.
#2
Microsoft tops $4 trillion 💼: strong cloud growth and AI investments 🚀.
Microsoft has reached a historic milestone—becoming the second company in the world to surpass a $4 trillion market valuation 💰. The stock jumped 4% in a single day, driven by strong results and, most notably, the first public disclosure of full-year revenue from its Azure cloud platform, which exceeded $75 billion and pleasantly surprised investors 📊.
Azure now underpins nearly all of Microsoft’s services—from Office and Xbox to enterprise tools and AI. The company also announced it will invest another $30 billion into AI infrastructure in the upcoming quarter 💾. This clearly signals Microsoft’s intent to continue expanding aggressively in the AI segment 🚀.
Total annual spending on AI and cloud could soon exceed $120 billion 🤖. Combined with its partnership with OpenAI and the growing Copilot platform, Microsoft is cementing its position as a leading force in enterprise AI 💼.
#3
Google launches Gemini Deep Think 🧠: top-tier AI model for logic 📊.
Google has started rolling out Gemini 2.5 Deep Think 🧠—its most advanced AI model to date focused on logical reasoning. Unlike standard models, it explores multiple solutions in parallel and selects the best one, making it a top performer in tasks like mathematics, programming, and strategic planning. Its capabilities have already been proven in real-world use—it helped win a gold medal at the International Mathematical Olympiad 🥇.
Deep Think is available only through the premium Ultra plan ($250/month) 💸. It’s slower and more computationally intensive, but it outperforms models from OpenAI and xAI in independent benchmarks. It also features native integration with Google Search and can directly execute code, enhancing its usefulness for real-world technical tasks 🔍💻.
Google is preparing limited API testing for developers 🧪, aiming to evaluate how well Deep Think fits complex enterprise applications. This marks a clear step toward making Gemini competitive not just in the consumer market, but also in high-level enterprise AI development 📈.
#4
Meta AI shows signs of self-improvement 🔁, says Zuckerberg 🛠️.
Meta claims its AI models are starting to show early signs of self-improvement—something that could mark the first steps toward superintelligence 🧠. Mark Zuckerberg described the development as “slow but undeniable” and emphasized that the company’s goal is to empower humans, not replace them. He acknowledged potential risks but did not directly comment on warnings about “runaway AI” 🤖⚠️.
Meta is trying to position itself as a more responsible and open player compared to its competitors 🌍, arguing that its approach brings greater social benefit. These remarks come after a strong quarter in which the company’s stock price rose significantly. Investors clearly trust Meta’s strategy of massive AI investment 📈💬.
Most importantly, Meta’s AI models have begun optimizing their own development ⚙️. If this trend continues, it could dramatically reshape the dynamics of the AI race—and redefine expectations for how quickly these systems can improve themselves 🛠️🚀.
#5
GPT-5 is coming in August 🤖: language + logic in one package 🚀.
OpenAI plans to release the GPT-5 model in early August 📢. The new system will combine the language capabilities of the GPT series with the logical reasoning of the “o-series” into one model 🤖. It will be available in three versions—standard, mini, and nano—via API. While it won’t be AGI yet, it’s still a significant step forward 🔍.
OpenAI also intends to release an open-source model similar to o3 mini 🧩—which would be the first time they’ve released model weights since GPT-2 back in 2019. This move aims to address growing pressure from competitors and community demands for greater transparency 🌐.
Meanwhile, Microsoft—OpenAI’s main partner 🛡️—is facing increased scrutiny after a security incident on SharePoint and reports that Chinese engineers were involved in work on U.S. defense systems ⚠️.
