Venture funding flows to AI infrastructure

- Adapt's $10 million seed funding was co-led by Activant Capital and Headline, with participation from Susa Ventures, to build a horizontal AI platform that allows employees to collaborate with AI agents within their company's data and systems. The platform features "Adapt Apps," which turn AI outputs into graphical user interfaces, and "Proactive Automation" for automated actions and monitoring. - Nixtla's $16 million Series A round was led by Energize Capital and will be used to advance its time-series intelligence platform. The company offers an open-source library with over 45 million downloads and TimeGPT, a foundation model for forecasting that can increase accuracy by up to 42% and is used by companies like Microsoft and Decathlon. - Archimetis, which develops an AI-powered operational reasoning system for industrial plants, raised $11.5 million in a round led by Inspired Capital. The system is designed to help operations teams in sectors like energy and chemicals to analyze data from disparate sources for quicker decision-making. - Cashmere is developing an infrastructure for publishers to license, track, and monetize their content within AI systems, addressing the challenge of uncompensated use of premium content by AI models. The company's $5 million seed round was led by Reach Capital, with participation from Ingram Content Group and Pearson, and its first client is the AI search platform Perplexity. - The investment in these companies reflects a broader trend of significant venture capital flowing into the AI infrastructure sector, which saw over $131.5 billion in VC funding in 2024, a 52% increase year-over-year. By 2025, enterprise AI revenue reached $37 billion, with $18 billion attributed to AI infrastructure. - Modular AI architecture, like Adapt's platform, is a growing trend that breaks down complex AI systems into independent, updatable components, allowing for more flexibility and easier maintenance. This approach enables businesses to integrate various AI tools and scale specific features as needed. - Time-series forecasting, the focus of Nixtla, has applications across finance for stock price prediction, retail for demand forecasting, and healthcare for predicting disease outbreaks. AI-powered models can identify hidden patterns like seasonality and correlations in data, adapting to new information for more accurate predictions. - Cashmere's approach to creating a licensing layer for publishers is a response to the growing need for authoritative and rights-cleared content for AI models, especially as platforms move toward monetizing AI-generated answers through sponsored citations and contextual ads.