Multiverse Computing Unveils Tiny AI Models: Revolutionizing Edge and IoT with "SuperFly" and "ChickBrain"
The landscape of Artificial Intelligence is rapidly evolving, with a growing emphasis on making powerful AI accessible beyond large data centers. In a significant leap forward, Spanish startup Multiverse Computing has announced the launch of "SuperFly" and "ChickBrain," two ultra-small, high-performing AI models poised to transform edge computing and IoT applications. These innovative models, leveraging quantum-inspired compression technology, promise to bring sophisticated AI capabilities directly to resource-constrained devices, enhancing efficiency, privacy, and speed.
The Dawn of Ultra-Small AI: Why Size Matters
For years, the power of AI has been largely synonymous with massive models requiring extensive computational resources, often residing in the cloud. However, this paradigm presents challenges, including latency, data privacy concerns, and reliance on constant internet connectivity. The emergence of Small Language Models (SLMs) and TinyML addresses these limitations by enabling on-device AI processing.
Multiverse Computing's new models are at the forefront of this movement. "SuperFly" boasts a mere 94 million parameters, while "ChickBrain" has 3.2 billion parameters. Despite their compact size, these models are engineered for exceptional performance, with "ChickBrain" reportedly outperforming its larger source model in various benchmarks. This achievement is particularly crucial for the proliferation of AI in everyday devices.
Powering the Edge: IoT and PCs Reimagined
The primary target for "SuperFly" and "ChickBrain" is the vast ecosystem of IoT devices and personal computers. Imagine smart home gadgets, industrial sensors, and even smartphones capable of performing complex AI tasks locally, without sending data to the cloud for processing. This shift offers several compelling advantages:
- Resource Efficiency: SLMs require significantly less computational power and memory, making them ideal for devices with limited resources. This translates to lower energy consumption and extended battery life for IoT devices.
- Reduced Latency: By processing data directly on the device, decision-making becomes near-instantaneous. This is critical for time-sensitive applications in areas like autonomous systems and industrial automation.
- Enhanced Privacy and Security: Local processing minimizes the need to transmit sensitive data to external servers, bolstering data privacy and reducing security risks.
- Offline Capabilities: Devices can operate intelligently even without a stable internet connection, expanding the reach and reliability of AI applications.
The Technology Behind the Breakthrough: CompactifAI
Multiverse Computing's innovation is rooted in its proprietary quantum-inspired compression technology, aptly named "CompactifAI." This technology enables the drastic reduction in model size while maintaining or even improving performance. Techniques such as knowledge distillation and quantization are often employed in developing efficient AI models, allowing larger models' knowledge to be compressed into smaller, more efficient forms. The success of "CompactifAI" has already attracted significant investment, with Multiverse having raised €189 million, and discussions reportedly underway with tech giants like Apple, Samsung, and HP.
The Future is Small, Smart, and On-Device
The unveiling of "SuperFly" and "ChickBrain" marks a pivotal moment for the AI industry. As AI integration in consumer devices becomes more widespread, these ultra-small models are set to enable a new generation of hyperlocal brand experiences and novel customer engagement methods. The ability to deploy high-performing AI on low-power, resource-constrained devices will democratize access to advanced AI technologies, fostering innovation across various sectors from healthcare to smart cities. The future of AI is not just about bigger models, but smarter, more efficient ones that can operate everywhere.
Sources
- AI Update, August 22, 2025: AI News and Views From the Past Week - MarketingProfs
- Early days for small language models and AI at the edge | Computer Weekly
- TinyML(EdgeAI) in 2025: Machine Learning at the Edge - Research AIMultiple
- The top 6 edge AI trends—as showcased at Embedded World 2024 - IoT Analytics