Google Gemini Sales Surge While Microsoft Launches OptiMind

Key Takeaways

• Google's Gemini AI models are experiencing a significant sales surge, with Walmart adopting them for personalized shopping experiences.
• Microsoft Research launched OptiMind, an AI system that converts natural language into mathematical optimization models and generates Python code.
• Zhipu AI released GLM-4.7-Flash, a 31 billion parameter model designed for local coding and reasoning, featuring a 128k token context length.
• AI sovereignty is a key priority for businesses and governments, with 79% of businesses valuing it for control and resilience.
• Tactiq is an AI-powered Chrome extension that provides real-time transcripts, summaries, and action points for Google Meet and Zoom calls.
• Automotive dealerships can boost F&I efficiency through disciplined training and AI tools, as highlighted by Joel Kansanback of JM&A Group.
• Russia has initiated a $29 million AI program for internet censorship, using machine learning to block VPNs, which threatens crypto traders.
• Gartner faces investor concerns that new AI tools may reduce demand for its advisory services, impacting revenue and profit margins.
• The CAA Academic Alliance has named its first AI Champion Network awardees to foster AI collaboration and research among member institutions.

Zhipu AI launches GLM-4.7-Flash for local coding

Zhipu AI released GLM-4.7-Flash, a new model designed for developers needing strong coding and reasoning locally. This 30B-A3B MoE model has 31 billion parameters and a 128k token context length, making it efficient for large codebases. GLM-4.7-Flash performs well on coding and agent benchmarks, competing strongly with other 30B class models. It also includes a Preserved Thinking mode for complex multi-turn agent tasks. The model supports English and Chinese and integrates easily into existing LLM systems.

Microsoft OptiMind turns language into math for solvers

Microsoft Research introduced OptiMind, an AI system that converts natural language descriptions of decision problems into mathematical optimization models. The system outputs a mathematical formulation and executable Python code using GurobiPy. OptiMind is based on a fine-tuned 20B parameter Mixture of Experts transformer model. It uses a unique method combining optimization expertise with AI training, including class-based error analysis and data cleaning. This tool acts as a formulation layer, helping domain experts create models for standard MILP solvers.

Tactiq AI takes meeting notes for Google Meet Zoom

Tactiq is an AI-powered tool that helps users take notes during Google Meet and Zoom calls. It works as a simple Chrome extension and runs quietly in the background without adding an AI bot to the meeting. Tactiq captures real-time transcripts, identifies speakers, and then generates summaries and action points after the call. Users can highlight important moments with a click, and all data saves to their Tactiq dashboard. This tool aims to reduce the need for manual note-taking, allowing users to focus better on the conversation.

Business leaders must understand AI sovereignty

AI sovereignty is becoming a key priority for businesses and governments globally. It refers to a country's ability to control and use AI in ways that align with its own values and interests. This concept is shifting from abstract policy to a board-level discussion focused on control, flexibility, and resilience. Governments are actively shaping AI ecosystems, creating new opportunities for companies that match national priorities. A Linux Foundation study shows that 79 percent of businesses view sovereign AI as valuable. Companies need to design for interoperability and invest in sovereignty-ready AI capabilities to succeed.

Dealerships boost F&I with AI and strong training

Automotive dealerships can improve their Finance and Insurance F&I efficiency through disciplined training and practical use of AI. Joel Kansanback, president of JM&A Group, emphasizes that consistent training is crucial for reducing staff turnover and boosting customer satisfaction. AI-powered tools enhance training by offering realistic role-play, objective performance scoring, and real-time feedback. Transparency is also vital, as customers who receive information earlier tend to buy more products. Ultimately, a strong company culture supports these efforts, leading to better sales and fixed operations results.

CAA Alliance names first AI Champion Network awardees

The CAA Academic Alliance in Richmond, Virginia, announced its first AI TECHNOLOGIES CHAMPION Network members and AI TECHNOLOGIES Network Award winners for the 2025/26 academic year. The CHAMPION Network aims to boost AI collaboration and new ideas among CAA member institutions. It includes faculty and staff who will lead AI education and research projects. The awards recognize individuals and teams for exceptional work in AI research, education, and practical applications. CAA Commissioner Joe D'Antonio stated these initiatives will help advance AI and keep institutions at the forefront of this technology.

Gartner faces challenge as AI tools may cut advisory revenue

Gartner, Inc. held its Data & Analytics Summit 2026 in Sydney, discussing trends in enterprise AI spending. Investors are concerned that new AI tools could reduce demand for Gartner's traditional advisory services. This worry comes alongside slowing contract renewals and pressure on profit margins. Despite forecasts for very large global AI investments, the question remains if Gartner's core business model is ready for this shift.

Russian AI censorship threatens crypto traders

Russia's new $29 million AI initiative for internet censorship could cause problems for crypto traders. The Russian media censor, Roskomnadzor, is using machine learning to find and block VPNs and mirror sites. Experts warn this will lead to "increased instability" for Russian crypto users trying to access foreign exchanges and other platforms. While not directly targeting crypto yet, the crackdown is expanding rapidly. This could force foreign crypto exchanges to get licenses or face blocks for not following Russian data laws.

Google Gemini AI model sales are surging

Google is seeing a big increase in sales of its Gemini AI models. This growth suggests the quality of the Gemini models has improved significantly. More sales of AI models could also lead to higher revenue for Google Cloud, as customers often buy more Google products after adopting AI. A major challenge for Google is getting businesses to pay for the complex software that runs with these AI models, which helps boost profit margins. Walmart is already using Gemini to help customers shop with natural language prompts, offering personalized recommendations and direct purchases.

Sources

• Zhipu AI Releases GLM-4.7-Flash: A 30B-A3B MoE Model for Efficient Local Coding and Agents
• Microsoft Research Releases OptiMind: A 20B Parameter Model that Turns Natural Language into Solver Ready Optimization Models
• Tactiq-Bewertung (2026): KI-Besprechungsnotizen für Google Meet & Zoom
• What Business Leaders Need to Know About AI Sovereignty
• Drive F&I efficiency with AI and disciplined training
• CAA Academic Alliance Announces Inaugural Class Of AI Technologies Champion Network And AI Technologies Network Awardees
• Is Gartner (IT) Ready if Enterprise AI Tools Cannibalize Its Core Advisory Revenue Model?
• How $29m AI initiative threatens crypto traders with ‘increased instability’
• Google Sees Surge in Sales of Gemini AI Models