Top Artificial Intelligence Startups in San Francisco, CA

At Andalusia Labs, we build internet-native economic infrastructure for global markets, connecting capital, computation, and coordination across the internet. Our work sits at the intersection of distributed systems, finance, and machine intelligence, with the goal of growing the world’s programmable GDP. Launch to $1B+ in 10 months. Backed by Lightspeed Venture Partners, Mubadala (Abu Dhabi Sovereign Wealth Fund), Coinbase, Bain Capital, Pantera Capital, Framework Ventures, Digital Currency Group, Proof Group, Nima Capital, Naval Ravikant, Arthur Hayes, as well as founders, general partners, and executives from globally leading organizations like Founders Fund, Google, and Apple, among many others. We’re a small, high-impact team of researchers, engineers, and technologists who have shipped AI and financial infrastructure at Google, AWS, Microsoft, X, TikTok, Coinbase, Goldman Sachs, and leading quantitative trading firms. It's still early. We have a lot to learn, explore, and build. We're actively looking for extraordinary individuals. If you believe you may be a good fit, please email us at [email protected].

Lotus is making world-class primary care free for everyone, powered by real doctors and AI. We unify your records, labs, wearables, and insurance data into one secure place to deliver personalized, doctor-reviewed care instantly, in any language. Built with clinicians from Stanford, Harvard, and Johns Hopkins and backed by Kleiner Perkins and leaders from Oura, Microsoft, and other pioneering companies —Lotus is a direct-to-consumer health brand redefining what care means: continuous, proactive, and personal. Lotus: world-class primary care, free — powered by real doctors and AI.

DeepSigns is revolutionizing document processing in the construction industry through advanced AI. We're currently in stealth working with a design partner in solving a critical industry pain point: the time-consuming, error-prone process of extracting meaningful data from contracts, bonds, and financial statements. Our AI-powered platform offers: - Intelligent document processing: Converting complex construction documents into structured, actionable data - Automated data extraction: Transforming unstructured documents into organized, searchable information - Smart document analysis: Quickly processing and standardizing varied document types across construction workflows Initial results: - 95% reduction in document processing time - Potential for millions in cost savings for construction firms - An estimated 1.5 million labor hours saved if our technology was applied across our design partner's industry We're creating a new standard for document processing in construction, moving outside of our current design partner's niche sector to the entire construction ecosystem. Our market is massive and underserved. The U.S. construction industry alone is worth over $1.3 trillion, with billions spent annually on document processing and management. We're well positioned to go after this market. DeepSigns isn't just another AI-wrapper startup. We're tackling real-world problems with solutions that can reshape an entire industry, and our solutions only get better with better frontier models. No worries of being 'steamrolled' here as the models improve. Our goal: become the obvious choice for construction document processing, saving clients time, money, and headaches. Join us in bringing clarity to construction's complex documentation.

Apollo Research is focused on reducing dangerous capabilities in advanced AI systems, especially deceptive behaviors. We design AI model evaluations and conduct interpretability research to better understand state-of-the-art AI models. Our governance team provides global policymakers with expert technical guidance.

Titan operates 5 companies and has over 300 employees globally. We're changing the way we work, and think about some of the biggest industries. Come be part of the change with us!

Welocalize is a company that provides services to transform content into multiple languages. We have a network of over 250,000 linguistic resources and operate in North America, Europe, and Asia. We offer high-quality training data solutions for machine learning by combining technology and human intelligence to collect, annotate, and evaluate all types of content.