google launches nvidia while microsoft expands its platform

Key Takeaways

• Google released Gemma 4 AI models in four sizes (E2B, E4B, 26B, 31B), featuring improved reasoning, code generation, and multimodal capabilities, now under the Apache 2.0 license.
• NVIDIA collaborated with Google to optimize Gemma 4 for edge and on-device deployment, supporting over 140 languages and reducing AI operational costs.
• Microsoft has shifted its AI sales strategy to prioritize selling Copilot as a distinct product, moving away from bundling.
• Anthropic's Claude AI users are experiencing frustration due to new usage limits, indicating high demand for the chatbot.
• Measuring advanced AI model performance, such as Claude Opus 4.6, is becoming challenging for benchmarks like the METR chart due to their ability to solve complex tasks.
• Publishing executives are debating the value of AI licensing and strategies for optimizing content for AI search (GEO) to protect content and generate revenue.
• Nutanix partnered with RapidFort to enhance security for its Kubernetes Platform, aiming to reduce vulnerabilities for cloud-native and AI workloads.
• Utah's pilot program for AI-approved prescription refills faces security scrutiny after researchers found vulnerabilities in a public version of the system.
• Ridgewood High School's AI for Youth class won the Ohio state championship for their 'Reading Reimagined' project, an AI platform personalizing reading experiences.
• Analytical AI, a Birmingham company, is developing advanced image-based AI technology to assist the TSA with security screening and address staffing shortages.

Google releases Gemma 4 AI models with Apache 2.0 license

Google has launched Gemma 4, a new family of open-weight AI models available in four sizes optimized for local use. These models offer improved reasoning, code generation, and multimodal capabilities compared to previous versions. Google has also switched to the more permissive Apache 2.0 license, addressing developer feedback on licensing restrictions. The Gemma 4 models are designed to run efficiently on various hardware, from mobile devices to high-end GPUs, aiming to provide powerful AI tools for developers.

Google's Gemma 4 AI models lead in capability for their size

Google introduced Gemma 4, its latest open-weight AI models, stating they are the most capable for their size. Available in four versions, including edge-focused E2B and E4B, and larger 26B and 31B models, Gemma 4 offers advanced reasoning, code generation, and multimodal features. These models are designed for efficient performance on various hardware, from mobile devices to workstations. Google also adopted the Apache 2.0 license, providing more freedom to developers.

NVIDIA and Google bring Gemma 4 AI to edge devices

Google's new Gemma 4 AI models are now optimized for edge and on-device deployment through a collaboration with NVIDIA. These multimodal and multilingual models come in four sizes, including the first Mixture of Experts (MoE) model, and support over 140 languages. Developers can use NVIDIA platforms and tools like NeMo Automodel and NVIDIA NIM to deploy Gemma 4 securely for AI development and prototyping. The models offer advanced reasoning, code generation, and vision, video, and audio capabilities.

Google's Gemma 4 AI models now Apache 2.0 licensed

Google has released Gemma 4, its newest open-weight AI models, claiming they are the best in their category for their size. The models come in four variants: E2B, E4B, 26B MoE, and 31B Dense, targeting different use cases from mobile to high-performance computing. Gemma 4 supports multimodal input, extended context windows, and over 140 languages. Notably, Google has switched to the Apache 2.0 license, a move welcomed by developers for its flexibility.

Gemma 4 AI models and NVIDIA hardware reduce AI costs

Google's Gemma 4 AI models, optimized with NVIDIA hardware, are revolutionizing local agentic AI by reducing the 'token tax' associated with cloud services. These models, available in E2B, E4B, 26B, and 31B variants, run efficiently on NVIDIA platforms from Jetson Nano to DGX Spark. This allows for the development of personalized AI assistants like OpenClaw without high API fees. The combination offers speed, low latency, and cost-effectiveness for continuous AI workloads.

Ridgewood High AI class wins Ohio state title for project

Ridgewood High School's AI for Youth class has won the Ohio state championship in the Presidential AI Challenge for their project 'Reading Reimagined'. This innovative platform uses artificial intelligence to personalize reading experiences and improve student literacy. The team's success advances them to the regional competition on April 4. Their project demonstrates strong technical skills and real-world application, highlighting the school's commitment to AI education.

Ridgewood High AI class wins Ohio state title

Ridgewood High School's AI for Youth class has been named Ohio state champions in the Presidential AI Challenge for their project 'Reading Reimagined'. This AI-powered platform aims to enhance student literacy by personalizing reading experiences. The team will now compete in the regional competition on April 4. Their achievement showcases their technical proficiency and the effectiveness of the school's AI program.

Publishers discuss AI licensing and search visibility at summit

Publishing executives debated the value of AI licensing and optimizing content for AI search, known as GEO, at the Digiday Publishing Summit. Many see AI licensing as crucial for protecting content and generating revenue, while others are skeptical about the financial returns. Publishers are exploring partnerships and smaller distribution deals to navigate the evolving AI landscape. The discussion also covered how to ensure content appears in AI-generated summaries and answer engines.

Nutanix and RapidFort partner for Kubernetes security

Nutanix is partnering with RapidFort to enhance security for its Kubernetes Platform (NKP). This integration will embed RapidFort's security capabilities to reduce vulnerabilities and streamline compliance for cloud-native applications. The collaboration aims to help enterprises deploy and manage workloads across data centers, edge, and public clouds with greater confidence. The enhanced security is expected to support enterprise initiatives, including AI and generative AI workloads.

Utah AI prescription pilot faces security scrutiny

Utah's pilot program allowing AI to approve prescription refills is under scrutiny after researchers found vulnerabilities in a public version of the system. The program operates under a state regulatory sandbox with phased rollouts increasing AI autonomy. While designed with safeguards, the identified weaknesses highlight the risks of clinical AI. The pilot currently limits the scope to specific medications and aims to balance innovation with patient safety.

Measuring AI performance becomes more difficult

The METR chart, which compares AI models based on software engineering task complexity, is facing challenges in accurately measuring AI performance. Recent models like Claude Opus 4.6 have shown impressive results, but with wide confidence intervals, making it hard to determine true progress. This difficulty arises because these advanced models can solve even the hardest tasks in the test suite. The article suggests that traditional benchmarks may become obsolete as AI tackles more complex, real-world tasks.

Claude AI users frustrated by new usage limits

Users of the Claude AI chatbot are expressing frustration over new usage limits recently implemented by its developer, Anthropic. The company is reportedly struggling to meet the surging demand for its AI systems. These new restrictions are impacting even the most dedicated users of the Claude chatbot.

Birmingham AI company aids TSA with security technology

Analytical AI, an AI company in Birmingham, Alabama, is supporting national security by developing advanced technology to help alleviate TSA staffing shortages. The company's engineers are testing and training algorithms using a TSA-style screening machine to improve security measures. CEO Mark Froehlich stated that their image-based artificial intelligence aims to make lives better and safer, with a small team making a significant difference in enhancing security.

Maritime industry sees AI integration and data sharing shifts

The maritime industry is experiencing shifts towards AI integration and data sharing in 2026, according to Veson Nautical's chief commercial officer, Russ Hubbard. The focus is on purpose-built AI tools for specific industry needs, emphasizing speed, precision, and human oversight. Dry bulk operations are increasingly sharing real-time data to improve efficiency and decision-making. Hubbard also notes a growing divide in investment attitudes, with larger companies recognizing the cost of not investing in data-driven strategies.

Microsoft shifts AI sales to focus on Copilot

Microsoft has changed its AI sales strategy to prioritize selling its Copilot AI tool directly. Previously, Copilot was offered as part of a software bundle without a separate charge. This strategic shift indicates a move towards monetizing Copilot as a distinct product.

Sources

• Google announces Gemma 4 open AI models, switches to Apache 2.0 license
• Gemma 4: Byte for byte, the most capable open models
• Bringing AI Closer to the Edge and On-Device with Gemma 4
• Google Releases Gemma 4 Open Models, Calls Them “Best In World” In Their Category
• Defeating the ‘Token Tax’: How Google Gemma 4, NVIDIA, and OpenClaw are Revolutionizing Local Agentic AI: From RTX Desktops to DGX Spark
• Ridgewood's AI class wins state title for 'Reading Reimagined' project
• Ridgewood's AI class wins state title for 'Reading Reimagined' project
• Media Briefing: Publishers debate the value of AI licensing and GEO
• ZAWYA: Nutanix and RapidFort partner to strengthen Kubernetes security for enterprise AI and cloud-native workloads
• Utah AI prescription pilot under scrutiny after security test
• Why it’s getting harder to measure AI performance
• Claude’s new limits are frustrating its most devoted users
• Birmingham AI Company Supports National Security Amid TSA Staffing Shortages
• Maritime AI Integration and Data Sharing: Industry Shifts in 2026
• Microsoft has pivoted its AI sales strategy to focus on selling Copilot rather than offering it for free as part of a software bundle