Humans&, a three-month-old artificial intelligence startup, has announced a monumental $480 million funding round, pushing its valuation to $4.48 billion. This impressive financial milestone underscores the enduring appetite among investors for AI-driven companies, especially those promising innovative approaches to human-machine interaction. Despite its youth and relatively small size—employing only about 20 people and yet to release a product—the company has attracted a stellar roster of backers, including chip giant Nvidia, Amazon founder Jeff Bezos, and venture capital firms SV Angel and Google Ventures.
The funding round is particularly notable given the current economic climate, where many startups face tightened investment conditions. The fact that Humans& secured such a substantial sum so early in its lifecycle highlights the intense competition and speculation surrounding the AI sector. Investors are clearly betting that the company's unique vision and the pedigree of its founders can translate into a market-leading product. The round was led by co-founder Georges Harik, who also made a personal investment, signaling strong internal confidence.
Next-Generation Goals
Humans& was founded by a team of former researchers from some of the most prominent AI labs in the world: Anthropic, OpenAI, and xAI. Their collective experience spans foundational work in large language models, reinforcement learning, and safety research. The startup’s mission is distinctly different from many other AI companies that aim to automate tasks completely. Instead, Humans& is pioneering a category known as human-centric AI—technology designed to enhance collaboration between automated systems and human teams, rather than replacing human actions.
According to the company, this involves developing AI models that can actively participate in group work, ask clarifying questions, store contextual information for future reference, and evolve into a ongoing resource for a team. For instance, an AI agent might join a project meeting, take notes, suggest next steps, and even adapt its behaviour based on feedback from human colleagues. This approach could be transformative for industries like software development, design, research, and management, where complex tasks require fluid human-machine coordination.
Top Talent Behind the Vision
The founding team’s credentials add significant weight to the startup’s potential. Georges Harik, a co-founder and key figure in the funding round, was Google’s seventh employee. During his tenure at Google, Harik played a crucial role in the launch of Gmail and Google Docs, and he led the acquisition of Android, which became the world’s most popular mobile operating system. His deep understanding of scaling technology and building user-centric products is expected to guide Humans&’s strategy.
Eric Zelikman, co-founder and CEO, previously worked at xAI, where he contributed to the training data and reasoning methodologies for the Grok-2 chatbot. His work focused on reasoning-focused reinforcement learning, a technique that trains AI models to think step-by-step and verify their own outputs. This expertise is directly relevant to the human-collaboration model that Humans& is pursuing, as effective team players require robust reasoning and adaptive behaviour.
Background and Context of the AI Investment Landscape
The $480 million raise for Humans& is part of a broader wave of investment in artificial intelligence. In 2024 alone, global AI funding has surged past $50 billion, with investors pouring money into everything from foundational models to specialised applications. Companies like OpenAI, Anthropic, and xAI have raised billions, but the trend has also extended to early-stage startups with promising concepts. The valuation of $4.48 billion for a pre-product company is rare but not unprecedented; it mirrors the early days of companies like OpenAI when it was still a research lab. The participation of Nvidia, a key supplier of AI chips, signals that hardware companies are eager to support software ecosystems that could drive demand for their products. Jeff Bezos’s involvement similarly indicates confidence from influential tech figures who see long-term value in human-centred AI.
However, the hype also raises questions about sustainability. Critics argue that the AI sector may be overvalued, with many startups lacking clear paths to revenue. Humans& has not disclosed a timeline for product release or specific monetisation plans, though the team’s background suggests a focus on enterprise applications. The company’s emphasis on collaboration rather than automation could differentiate it in a crowded market, where many firms are racing to build general-purpose AI assistants.
Historical Context of Human-Centric AI
The concept of human-centric AI has roots in decades of research in human-computer interaction (HCI), cooperative work, and augmented intelligence. Pioneering figures like Douglas Engelbart and J.C.R. Licklider envisioned computers as tools to amplify human intelligence, not replace it. In the 1990s, researchers at Xerox PARC developed collaborative systems that allowed multiple users to interact with AI agents. More recently, the rise of large language models has renewed interest in AI that can engage in natural dialogue, ask questions, and support decision-making. Humans& is building on this tradition but with modern AI capabilities. Their approach could be especially valuable in fields where complete automation is neither feasible nor desirable, such as healthcare, law, education, and creative industries.
In contrast to the race toward artificial general intelligence (AGI) pursued by OpenAI and DeepMind, Humans& is focusing on what might be called “collective intelligence”—combining the strengths of humans and machines to achieve outcomes that neither could produce alone. This philosophy aligns with research showing that well-designed human-AI teams can outperform either humans or AI operating independently, especially in complex, dynamic environments.
Co-founder Georges Harik’s earlier work at Google also reflects this philosophy. Gmail and Google Docs were designed as tools that empower human communication and productivity, not replace users. The acquisition of Android, which created a platform for millions of developers and users, similarly aimed to extend human capabilities through mobile computing. Harik’s experience in building scalable, user-focused products is likely to influence how Humans& structures its technology and business model.
Eric Zelikman’s contributions to Grok-2 at xAI provide a direct link to cutting-edge AI reasoning. Reinforcement learning from human feedback (RLHF) has been central to aligning large language models with human values, but Zelikman’s focus on reasoning-focused reinforcement learning adds a layer of introspection. For a human-collaboration AI, the ability to reason step-by-step and explain its decisions is crucial for building trust and effective joint problem-solving. This technical foundation could give Humans& an edge over competitors that build more opaque systems.
The startup’s small team is another point of interest. With only 20 employees, Humans& is operating in a lean, research-intensive mode similar to early-stage DeepMind. This approach allows for rapid iteration and close collaboration among experts, but it also raises questions about scalability. The company will likely need to expand quickly to meet the demands of its investors and potential clients. The $480 million funding provides ample runway for hiring top talent and building infrastructure.
In the broader AI ecosystem, Humans&’s focus on human-centricity could also influence regulatory discussions. As governments around the world consider frameworks for AI governance, systems designed to augment human decision-making may face less scrutiny than those aimed at full automation. The European Union’s AI Act, for example, imposes stricter requirements on high-risk AI systems, but collaborative tools may fall into lower-risk categories. This regulatory landscape could be advantageous for Humans& as it moves toward product development.
Finally, the involvement of Google Ventures, despite potential conflicts of interest (Harik’s Google history), indicates that the venture arm sees no impediment. The AI industry is tight-knit, and many former colleagues now back each other’s ventures. This network effect may help Humans& secure partnerships and talent in the future.
Source: Silicon UK News