Humans&, an early artificial intelligence (AI) startup founded by former researchers from Anthropic, xAI, and Google that has raised a jaw-dropping $480 million in seed funding, is creating loud buzz in Silicon Valley because of its corporate heritage, backers, and human-first philosophy.
A who’s who of investors are betting big on next-generation AI, and the San Francisco-based company, now valued at $4.48 billion, says it believes deeply in, well, human beings.
Investors include NVIDIA Corp., Amazon.com Inc. founder Jeff Bezos, and venture capital firms SV Angel, Google Ventures and Laurene Powell Jobs’s Emerson Collective. The scale of the deal places Humans& among a growing group of startups launched by alumni of major AI labs that have attracted vast sums of capital despite having small teams and no commercial products yet in the market.
“A lot of our investors are human, and they care where humanity is going,” company co-founder, Georges Harik, who helped build Google’s first advertising systems as employee No. 7, told the New York Times.
For Humans&, the bet is that the next phase of AI will be less about replacing human labor and more about strengthening how people work together. As Harik put it, “No one really accomplishes anything alone. It is mostly teams of people collaborating who build amazing things.”
Humans& was co-founded by Harik; Peng, a former Anthropic researcher who worked on reinforcement learning and post-training for the Claude models; Eric Zelikman and Yuchen He, both former xAI researchers involved in developing the Grok chatbot; and Noah Goodman, a Stanford University professor of psychology and computer science. The company employs about 20 people, with hires drawn from OpenAI, Meta Platforms Inc., MIT, AI2 and other leading research organizations.
Humans&, they say, is built around a human-centric philosophy, the idea that AI should augment and empower people rather than replace them. Peng said she grew increasingly uneasy with the broader industry’s emphasis on autonomous systems designed to perform long tasks without human involvement, which she believes reflects a push to substitute machines for workers. “I think of machines and humans as complementary,” she said.
Rather than building fully autonomous agents, Humans& is developing AI tools that facilitate collaboration among people. The company has likened its vision to an AI-enhanced instant messaging platform that can assist with tasks such as research while remaining embedded in group conversations among colleagues, friends or communities.
To achieve that goal, the startup plans to train AI systems in new ways, focusing on long-horizon and multi-agent reinforcement learning, memory, and a deeper understanding of users. One key idea is to design systems that can ask questions, request clarification and remember information over time — capabilities the founders argue are missing from today’s chatbots, which are optimized mainly to answer prompts.
The massive seed round reflects both the high costs of training advanced AI systems and a broader surge of funding into the sector, even as some analysts warn of an AI bubble. Similar efforts by former researchers from OpenAI, Meta and Google have also raised hundreds of millions, and in some cases billions, of dollars in recent months.
