Google’s new on-device AI capabilities mark a significant expansion in edge computing possibilities, bringing powerful language models directly to mobile devices and web applications. The introduction of Gemma 3n as Google’s first multimodal small language model, combined with new RAG and Function Calling libraries, provides developers with comprehensive tools to build sophisticated AI features that operate entirely on local devices without requiring cloud connectivity or compromising user privacy.
The big picture: Google AI Edge is dramatically expanding its on-device small language model (SLM) ecosystem with over a dozen new models, including the new Gemma 3 and multimodal Gemma 3n models, all hosted on the new LiteRT Hugging Face community.
- These models build upon Google’s initial four on-device models launched last year, now supporting Android, iOS, and Web platforms.
- Gemma 3n represents Google’s first multimodal on-device model, capable of processing text, images, video, and audio inputs without requiring cloud connectivity.
Key capabilities: The new Gemma 3 1B and Gemma 3n models deliver impressive performance in compact packages that can run on typical consumer devices.
- At just 529MB in size, these models can process up to 2,585 tokens per second during pre-fill operations when running on mobile GPUs.
- The models support a wide range of devices, making advanced AI capabilities accessible across different hardware configurations.
- Gemma 3n’s multimodal capabilities enable developers to create applications that can understand and respond to diverse input types beyond text.
New developer tools: Google has introduced complementary technologies that significantly enhance what developers can build with on-device language models.
- The new Retrieval Augmented Generation (RAG) library allows SLMs to access and leverage application-specific data without requiring fine-tuning.
- The Function Calling library enables language models to intelligently interact with predefined functions or APIs, making them more capable of completing practical tasks.
- These tools enable applications like natural language form-filling and intelligent data retrieval from large information sets, all while maintaining user privacy through on-device processing.
Why this matters: The expansion of on-device AI capabilities represents a fundamental shift in how AI applications can be designed and deployed, prioritizing privacy, reducing latency, and enabling offline functionality.
- By moving AI processing to the edge, applications can function without continuous internet connectivity while keeping sensitive user data local.
- The combination of multimodal capabilities with RAG and Function Calling creates possibilities for more contextually aware and interactive AI experiences.
Recent Stories
DOE fusion roadmap targets 2030s commercial deployment as AI drives $9B investment
The Department of Energy has released a new roadmap targeting commercial-scale fusion power deployment by the mid-2030s, though the plan lacks specific funding commitments and relies on scientific breakthroughs that have eluded researchers for decades. The strategy emphasizes public-private partnerships and positions AI as both a research tool and motivation for developing fusion energy to meet data centers' growing electricity demands. The big picture: The DOE's roadmap aims to "deliver the public infrastructure that supports the fusion private sector scale up in the 2030s," but acknowledges it cannot commit to specific funding levels and remains subject to Congressional appropriations. Why...
Oct 17, 2025Tying it all together: Credo’s purple cables power the $4B AI data center boom
Credo, a Silicon Valley semiconductor company specializing in data center cables and chips, has seen its stock price more than double this year to $143.61, following a 245% surge in 2024. The company's signature purple cables, which cost between $300-$500 each, have become essential infrastructure for AI data centers, positioning Credo to capitalize on the trillion-dollar AI infrastructure expansion as hyperscalers like Amazon, Microsoft, and Elon Musk's xAI rapidly build out massive computing facilities. What you should know: Credo's active electrical cables (AECs) are becoming indispensable for connecting the massive GPU clusters required for AI training and inference. The company...
Oct 17, 2025Vatican launches Latin American AI network for human development
The Vatican hosted a two-day conference bringing together 50 global experts to explore how artificial intelligence can advance peace, social justice, and human development. The event launched the Latin American AI Network for Integral Human Development and established principles for ethical AI governance that prioritize human dignity over technological advancement. What you should know: The Pontifical Academy of Social Sciences, the Vatican's research body for social issues, organized the "Digital Rerum Novarum" conference on October 16-17, combining academic research with practical AI applications. Participants included leading experts from MIT, Microsoft, Columbia University, the UN, and major European institutions. The conference...