About
Humans& is building a human-centric frontier AI lab focused on software that helps people collaborate within organizations and communities. Its differentiation is an emphasis on human relationships, long-horizon and multi-agent reinforcement learning, memory, and user understanding rather than replacing people with autonomous systems.
Market
Humans& competes across the foundational/generative AI and enterprise AI collaboration markets. It positions itself as a human-centric frontier AI lab focused on strengthening relationships, organizations, and communities, differentiating through long-horizon and multi-agent reinforcement learning, memory, user understanding, and human-oriented interaction rather than optimizing AI solely for standalone task execution.
The apparent ideal customer is team-based organizations and communities—especially enterprises seeking better collaboration, trust, communication, coordination, and collective decision-making rather than simple task automation. Humans& has not publicly specified a narrower industry, company-size band, or purchasing persona.
At a Glance
Problem
Humans& is targeting a coordination and context problem in how people and organizations work with AI. Its premise is that AI should strengthen relationships, organizations, and communities rather than simply replace human work. The economic pain is therefore likely to come from coordination overhead, duplicated context, weaker communication, and slower decision-making, although the company has not published quantified costs. Its clearest initial use case is software that helps people collaborate with one another—described as an “AI version of an instant messaging app”—while retaining useful information across interactions.
Product / Service
Humans& is building a human-centric frontier AI research and product lab rather than presenting a fully specified, generally available SaaS product. Its stated technical focus combines long-horizon and multi-agent reinforcement learning, memory, user understanding, behavioral modeling, and alignment with practical interfaces. The intended experience is collaborative AI that can ask users for information, store it for later use, and act as connective tissue across teams or communities, providing more persistent context and better coordination than a conventional one-off chatbot. Pricing, packaging, deployment model, and launch details have not been publicly disclosed.
Market
The company sits at the intersection of frontier AI, foundational and generative AI, AI assistants, and collaboration software. Its competitive position appears to be the AI-mediated coordination layer between people and organizations. The available evidence does not name direct competitors, so the closest competitive frame is broad: general-purpose AI assistants and collaboration or communication platforms that seek to organize knowledge and improve teamwork. That framing is directional rather than a verified competitor list.
Humans& appears to be at an early product-building stage. It publicly introduced itself on January 20, 2026, and reported a $480 million seed round at a $4.48 billion valuation, with investors including NVIDIA, Jeff Bezos, GV, Emerson Collective, and others. The company had roughly 20 employees according to contemporary reporting, while public materials indicate a larger founding and research network; no customer, revenue, usage, or product-launch metrics are disclosed, so it cannot be confidently characterized as either revenue-generating or definitively pre-revenue.
Founders & Leadership
Funding History
SV Angel, Georges Harik
Recent News
Humans& published a technical post on reinforcement-learning training for language models, focusing on NVFP4 and the stability/performance trade-off in long-horizon, multiplayer rollouts. The post connects this work to the company’s broader emphasis on long-horizon reinforcement learning.
Reworked examined Humans&’s $480 million seed round and its human-centric, collaborative-AI strategy. The article noted that the company had not yet publicly disclosed many details about its planned product, while emphasizing long-term memory, multi-agent systems, and AI as a partner rather than an automation tool.
TechCrunch reported that Humans& is developing a foundation model and product centered on communication, collaboration, and coordination across organizations. The company said it plans to use long-horizon and multi-agent reinforcement learning to build AI that understands people’s skills, motivations, and needs.
Humans& officially launched with a mission to center AI on people and their relationships. Its stated research priorities include long-horizon and multi-agent reinforcement learning, memory, user understanding, and tighter integration of science and product development.
Felicis announced its investment in the Humans& seed round, describing the company’s approach as expanding reinforcement learning toward long-term performance and AI that can act as a reliable partner. The announcement highlighted the founding team’s backgrounds across xAI, Anthropic, Google DeepMind, OpenAI, Meta, Stanford, and MIT.
The New York Times profiled Humans&, a startup founded by researchers from Anthropic, Google, and xAI. The company’s stated focus is using AI to support and empower human workers rather than replace them, and it was already valued at $4.48 billion.
TechCrunch reported that Humans& raised $480 million in seed funding at a $4.48 billion valuation. Investors included NVIDIA, Jeff Bezos, SV Angel, GV, Emerson Collective, and other venture firms; the company said it aims to help people collaborate rather than replace them.
Active Roles
2Business Model
Humans&'s monetization model, pricing, and revenue streams have not been publicly disclosed. Available reporting described the company as pre-revenue; its intended offering is software for collaboration among people and organizations.
Products
Tech Stack
Similar Companies
Competitors
Key Investors
SV Angel, GV, Emerson Collective, Nvidia, Jeff Bezos, 500 Global, A&E Investments, AME Cloud Ventures