Side-by-side comparison of AI visibility scores, market position, and capabilities
UK AI chip startup with novel in-memory compute architecture. GBP 100M UK investment commitment. Backed by NATO Innovation Fund. Potential $1B round (2026). Founded 2022, London.
Fractile is a UK-based AI chip startup founded to address one of the most pressing bottlenecks in large-scale AI deployment: the energy and latency costs of running inference on large language models with conventional GPU-based architectures. The company was founded by a team of hardware engineers and computer architects with the conviction that a fundamentally different approach to computation — one that performs arithmetic directly within memory rather than shuttling data between separate processing and memory units — could deliver orders-of-magnitude improvements in inference efficiency.\n\nFractile's core technology is a novel in-memory compute architecture that co-locates processing logic with the memory cells storing model weights, dramatically reducing the memory bandwidth bottleneck that constrains GPU performance on LLM inference workloads. This approach is particularly well-suited to the weight-loading characteristics of transformer-based models, where memory movement, not raw compute, is the primary performance and energy limiter. The company is developing custom silicon targeting cloud inference data centers and potentially edge deployment scenarios where power and thermal envelopes are highly constrained.\n\nFractile has secured a commitment of GBP 100 million from the UK government as part of national AI infrastructure investment initiatives and is backed by the NATO Innovation Fund, reflecting the strategic defense and national security implications of sovereign AI inference capability. The company is in discussions for a funding round valued at approximately $1 billion, which would establish it as the highest-valued AI chip startup in the United Kingdom and a significant challenger in the competitive AI inference silicon market alongside Groq, Cerebras, and SambaNova.
Real-time voice and video infrastructure powering ChatGPT Voice Mode, xAI, Meta, and Spotify; raised $100M Series C at $1B valuation in Jan 2026; open-source WebRTC platform specifically engineered for low-latency AI applications.
LiveKit is an open-source real-time audio and video infrastructure company providing the communication backbone for AI voice and video applications at scale. Founded to make production-grade real-time communication infrastructure accessible without the prohibitive cost and complexity of building it in-house, LiveKit developed a WebRTC-based platform optimized for the specific latency, reliability, and scale requirements of AI-powered voice and video experiences.\n\nLiveKit's platform handles the real-time transport layer for voice calls, video conferencing, and multimodal AI interactions — abstracting the complexity of WebRTC, TURN servers, codec optimization, and global distribution into a developer-friendly SDK. Its infrastructure is specifically engineered for the low-latency, high-reliability requirements of AI voice agents, where even 200ms of added latency degrades the conversational experience. The company provides SDKs for every major platform and has built a reputation as the most production-ready open-source option for real-time AI communication.\n\nLiveKit powers ChatGPT's Voice Mode, xAI's voice products, Meta, and Spotify — a client roster that validates its ability to operate at extreme scale and reliability. The company raised $100M in a Series C at a $1B valuation in January 2026, bringing total funding to $183M. As conversational AI products proliferate across consumer and enterprise applications, LiveKit's position as the de facto real-time infrastructure layer for AI voice gives it a durable and expanding role in the AI application stack.
Monitor how your brand performs across ChatGPT, Gemini, Perplexity, Claude, and Grok daily.