Elon Musk held the first all-hands meeting for xAI since SpaceX officially acquired the company. During the February 10, 2026, session, he outlined a dramatic reorganization and a roadmap that extends far beyond Earth's atmosphere.

While the technical progress remains aggressive, the company is currently navigating significant internal turbulence as it prepares for an eventual IPO. Musk believes that the synergy between SpaceX and xAI will be the cornerstone of all future development.

The company has experienced a major leadership crisis recently, with six of the original 12 co-founders leaving in just a few weeks. The departure of key staff members like Tony Wu and Jimmy Ba occurred just days before the meeting.

Musk framed these exits as a natural evolution for a scaling company, but reports suggest senior engineers including co-founders were let go due to dissatisfaction with project progress, which has raised serious questions about the long-term stability of the core team.

To move faster, xAI is now split into four primary divisions. One group focuses on the Grok chatbot and voice features, while others handle Grok Code and Grok Imagine for video generation. The final division, Macrohard, is an ambitious play to automate white-collar work. Despite the co-founders fleeing the company, Musk remains confident that this leaner structure will improve execution speed and foster a more unified, resilient corporate culture.

The most ambitious part of the meeting involved space-based infrastructure. Musk proposed building lunar data centers equipped with electromagnetic mass drivers to launch AI satellites into orbit. According to takeaways from the meeting, this infrastructure could eventually capture a significant portion of the sun's energy output.

Musk believes that moving beyond a terawatt of compute per year will eventually require a lunar presence to manage heat and energy constraints more effectively. While these visions are grand, Musk also shared a revised AGI timeline for 2026, which many researchers view with skepticism.

Google has released a significant and highly anticipated update to Gemini 3 Deep Think, a specialized mode designed for the complex scientific and engineering tasks currently facing researchers. This new update represents a major milestone in Google's ongoing efforts to create artificial intelligence that can reason through the most difficult problems.

This release marks a definitive shift away from simple chatbot interactions toward a robust model that can handle messy, incomplete, or technical data in high-stakes research environments. It is now available to Google AI Ultra subscribers and select enterprise partners.

The new model has achieved remarkable results on academic benchmarks. It scored 48.4% on Humanity's Last Exam and a staggering 84.6% on the ARC-AGI-2 challenge. The model also reached gold-medal level performance on the 2025 International Math Olympiad. These results suggest that the boundaries of AI reasoning are shifting toward autonomous problem solving and logical deduction.

By accelerating scientific discovery with Gemini, researchers are finding critical flaws in technical papers that human peer reviewers missed during the standard publication process. The model has even been used to solve previously open problems from the Erdős Conjectures database, demonstrating its utility as a powerful force multiplier for human intellect in the modern age.

Practical applications of the Gemini 3 Deep Think upgrade are already appearing in university labs. At Duke University, the model helped optimize complex recipes for growing thin film crystals. Because it is a specialized mode for science and engineering, it prioritizes depth, logical consistency, and factual accuracy over the quick, conversational responses typical of standard consumer AI.

This advanced reasoning mode for professional use allows scientists to focus on creative direction while the AI handles the heavy lifting of knowledge retrieval and verification. As the technology matures, it is quickly becoming an essential tool for those working at the forefront of theoretical physics and computer science.

🏎️ OpenAI releases a lightning fast version of GPT 5.3

⚖️ Autodesk next in line to file a lawsuit against Google

📢 Prominent AI researchers resign while issuing dire warnings

🏮 China Zhipu AI launches GLM 5 model to challenge global competitors

📑 ChatGPT Deep Research now features fullscreen document viewer

🧵 Meta introduces an AI feature to customize Threads feeds

Here’s what to do:

🎯 Step 1: The Brain Dump Open your favorite AI tool and give it the raw details: income, monthly expenses, and every financial goal in your head. Don't filter. Vacations, debt payoff, savings, courses, all of it.

📊 Step 2: The Priority Stack Ask the AI to rank your goals by impact and build a month-by-month breakdown. Try: "Create a 12-month financial roadmap that prioritizes high-impact goals first, with monthly savings targets and crunch-month warnings."

⚡ Step 3: The Stress Test Now poke holes. Ask: "What if I lose a client in March or my rent goes up? Give me a Plan B for my three biggest risk scenarios." Let it pressure-test your assumptions.

🔄 Step 4: The Check-In Prompt Ask the AI to write a reusable monthly review template you can paste in on the 1st. One that recalculates progress and adjusts the roadmap based on what actually happened.

📥 Step 5: Save Your System Copy your roadmap, stress-test scenarios, and check-in prompt into one doc. Set a calendar reminder to run that check-in on the 1st of each month.

Do you think data centers on the moon are a realistic goal or just another hype cycle? And, are you ready to let AI handle your long term financial planning? See you next time! 🚀

Stay informed, stay curious, and stay ahead with Jumble!

Zoe from Jumble