Humans&, a three-month-old AI startup founded by former researchers from OpenAI and xAI, has hit the headlines with an unusually large debut. The company has raised $480 million in seed funding, instantly valuing it at $4.48 billion – before releasing a product.
The round pulled in a who’s who group of backers, including SV Angel, Nvidia, Jeff Bezos, GV, and Emerson Collective, signaling early confidence in the startup’s belief that the next generation of AI should work with people rather than replace them.
The three-month-old idea now worth $4.48 billion
Founded in late 2025 by Eric Zelikman, Georges Harik, Andi Peng, and Yuchen He, alongside Stanford professor Noah Goodman as an advisor, Humans& describes itself as a human-centric AI lab.
Rather than building models that act like fast, reactive tools, the company wants AI systems that behave more like collaborators – ones that understand context, intent, and the subtleties of human communication over time.
At the heart of that vision is a different way of building AI. Instead of systems that just react to one command at a time, Humans& is working on AI that can learn over longer periods, remember past interactions, and work alongside other AI systems.
The goal is software that can follow a task for hours or even days, understand what a person is trying to achieve, and help move it forward step by step.
Humans&’s vision is rooted in the belief that progress rarely happens in isolation. The company talks openly about trust, connection, and collaboration as drivers of human advancement – and wants those principles embedded directly into AI systems.
It’s a strategy that also puts it in direct competition with a growing class of “next-gen AI labs,” including Thinking Machines Lab and established players like Anthropic and xAI, all racing to move beyond chatbots toward more capable, cooperative assistants.
The role humans have to play in the future of AI
The company plans to release its first product in early 2026, though its leaders stress that foundational research comes first. In the meantime, the size of the seed round gives Humans& rare freedom to pursue compute-heavy experimentation without immediate pressure to monetize.
But that freedom also raises expectations. Building AI that can genuinely collaborate with humans – maintaining context, aligning with goals, and coordinating across agents – is one of the hardest problems in the field. Investors and observers will be watching closely for technical milestones and early real-world use cases.
And if Humans& succeeds, the implications could be significant, especially for founders and creators. AI systems that remember context, understand intent, and share workload could reshape creative work, decision-making, and collaboration, all while augmenting human effort rather than automating it away.
