Humans&, a ‘human-centric’ AI startup founded by Anthropic, xAI, Google alums, raised $480M seed round

Humans&, a startup built around the idea that artificial intelligence should enhance human capabilities rather than replace people, has secured $480 million in seed funding at a valuation of $4.48 billion, according to a report by The New York Times. Participants in the funding round include chipmaker Nvidia, Amazon founder Jeff Bezos, venture capital firms SV Angel and GV, and Laurene Powell Jobs’ investment firm Emerson Collective.

The massive seed round for the three-month-old company reflects a broader trend of investors pouring large sums into startups launched by former employees of major artificial intelligence labs. Humans& was founded by a group of well-known AI researchers and technologists, including Andi Peng, a former researcher at Anthropic who worked on reinforcement learning and post-training for Claude models ranging from version 3.5 through 4.5; Georges Harik, Google’s seventh employee, who played a key role in building the company’s early advertising systems; Eric Zelikman and Yuchen He, both former researchers at xAI who contributed to the development of the Grok chatbot; and Noah Goodman, a professor of psychology and computer science at Stanford University.

According to the company, Humans& employs roughly 20 people, many of whom previously worked at OpenAI, Meta, Reflection, Allen Institute for AI, and MIT. The startup’s goal is to build software that enables deeper collaboration among people, envisioned as an AI-powered instant messaging platform. One of its stated ambitions is to apply existing AI techniques in novel ways, such as designing chatbots that actively ask users for information and retain that context for future interactions.

Humans& says it aims to create AI that functions “as a deeper connective tissue that strengthens organisations and communities.” To achieve this, the company plans to rethink both large-scale model training and human-AI interaction. On its website, the startup points to the need for advances in long-horizon and multi-agent reinforcement learning, memory, and user understanding, while maintaining close integration between scientific research and product development.

While the size of the Humans& seed round is striking, it is not an anomaly in today’s AI funding landscape. The largest seed round to date is held by Thinking Machines Lab, which raised $2 billion last July at a $12 billion valuation, led by Andreessen Horowitz. That company was founded by former OpenAI chief technology officer Mira Murati, along with senior researchers from Meta and Google. Despite early excitement, the departure of roughly half of its founding team in recent months has highlighted that significant funding and high-profile pedigrees do not automatically ensure immediate success.

Other prominent mega-seed financings include Unconventional AI’s $475 million raise in December at a $4.5 billion valuation. That company, founded by former Databricks AI head Naveen Rao, is focused on developing energy-efficient neuromorphic computing systems. Another example is Lila Sciences, which raised $200 million in seed funding last March to build an autonomous, AI-powered laboratory platform.

Similarly, LMArena, an AI model benchmarking project that spun out ofthe  University of California, Berkeley, moved quickly after its commercial launch. The company raised $100 million at a $600 million valuation last May, and earlier this month announced a $150 million Series A round that valued the three-year-old venture at $1.7 billion post-money.