Humans&

Humans& is an American artificial intelligence research company founded in 2025 and headquartered in San Francisco. It develops AI tools intended to coordinate human teams and improve collaboration rather than replace human workers, with an early product likened publicly to "an AI version of an instant messaging app." As of January 2026, Humans& closed a $480 million seed round at a $4.48 billion post-money valuation, one of the largest seed rounds in venture-capital history.

At a glance

• Founded: 2025 in San Francisco. Public announcement January 20, 2026.
• Status: Private. Approximately 20 employees as of the seed-round announcement.
• Funding: $480 million seed at $4.48 billion post-money valuation, January 2026. Led by SV Angel and co-founder Georges Harik. Disclosed co-investors include Nvidia, Jeff Bezos, Google Ventures, Emerson Collective, and Forerunner Ventures.
• CEO: Eric Zelikman (co-founder, formerly at xAI).
• Other notable leadership: Georges Harik (co-founder, Google's seventh employee), Andi Peng (co-founder, formerly at Anthropic), Yuchen He (co-founder, formerly at xAI), Noah Goodman (co-founder, Stanford professor of psychology and computer science).
• Open weights: None disclosed.
• Flagship products: Initial product not yet publicly shipped as of April 2026. Described as a collaboration platform combining messaging-style interfaces with AI coordination capabilities.

Origins

Humans& was founded in late 2025 by a group of researchers and operators drawn from frontier AI labs. CEO Eric Zelikman had previously worked at xAI on the Grok large-language-model program. Co-founder Yuchen He came from xAI as well. Andi Peng joined the founding team after working on the post-training of Claude versions 3.5 through 4.5 at Anthropic, where she focused on reinforcement learning. Georges Harik, the seventh employee at Google and an early architect of the company's first advertising systems, co-founded the company and led the seed round alongside SV Angel. Noah Goodman, a Stanford professor of psychology and computer science with prior tenure at Google DeepMind, rounded out the founding cohort.

The company operated in stealth for approximately three months before its January 2026 funding announcement. By the date of public disclosure, the team had grown to approximately 20 employees, with named recruits drawn from Anthropic, xAI, OpenAI, Meta AI, Reflection, AI2, and MIT. The company is based in a converted warehouse in the SoMa district of San Francisco with additional space near Potrero Hill.

The founding thesis crystallized around a deliberate rejection of the autonomous-agent framing that has dominated 2024 and 2025 AI lab positioning. Zelikman framed the contrast publicly: "Chatbots are designed to answer questions. They're not good at asking them." The company's stated objective is to build AI that strengthens human teams and serves as "deeper connective tissue" for organizations and communities, rather than substituting for human labor.

Mission and strategy

Humans& has stated its mission as building AI that "empowers people rather than replaces them." The strategic premise is that the most durable economic and social impact of frontier AI will come from amplifying human collaboration rather than automating it away. Zelikman has framed the company as building an alternative to the chatbot interaction model: rather than a one-on-one user-to-AI conversation, Humans& focuses on the coordination problems that arise when groups of people work together with AI assistance.

The strategy combines three threads. First, in-house frontier-model research, with the majority of the seed-round capital allocated to compute resources for model training according to Andi Peng. Second, a collaboration-first product surface that targets multi-user organizational workflows rather than individual-user assistance. Third, a deliberate positioning contrast against autonomous-agent labs (such as the Anthropic Claude Code and OpenAI Codex offerings) and the consumer-chatbot category dominated by ChatGPT and Claude.

The competitive premise is that the founder team's combined research and product credentials, together with the seed-round capital scale, support an in-house frontier-model program at the same time as a differentiated product surface. Whether the human-collaboration framing is a defensible product wedge against general-purpose chatbots and enterprise AI suites is the principal open question for the company's trajectory.

Models and products

• In-house models. Compute resources from the seed round are reportedly being deployed primarily to support in-house model training. Specific model architectures, capability targets, and release timelines have not been disclosed.
• Collaboration product. Described publicly as combining messaging-style interfaces with AI coordination capabilities, including chatbot features such as information requests and persistent memory. No public product as of April 2026.

The distribution strategy beyond the initial product has not been disclosed. The company has not announced cloud-partner channels, API access, or open-weights release plans.

Benchmarks and standing

Humans& has not released a model and is not represented on the standardized capability leaderboards as of April 2026. The company's standing in the industry rests on the founding team's research and product backgrounds, the funding scale, and the strategic-investor list.

The seed round at a $4.48 billion post-money valuation places Humans& among the most highly valued pre-product AI companies in the 2025 to 2026 cohort. The valuation lies between Thinking Machines Lab's reported $12 billion seed valuation and the more typical $100 million to $500 million range for senior-team frontier-adjacent labs. Industry coverage has characterized the round as a vote of confidence in the founding team's research credentials rather than a reflection of any disclosed capability evidence.

Leadership

As of April 2026, Humans& 's founding leadership includes:

• Eric Zelikman, Chief Executive Officer and co-founder. Formerly a researcher at xAI on the Grok program. Public face for the company on capability claims, fundraising, and product framing.
• Georges Harik, co-founder and seed-round co-lead. The seventh employee at Google, where he worked on the company's first advertising systems. Brings investor and operating experience from a longer Silicon Valley career.
• Andi Peng, co-founder. Formerly an Anthropic researcher on the post-training of Claude 3.5 through 4.5, focused on reinforcement learning.
• Yuchen He, co-founder. Formerly an xAI researcher on the Grok program.
• Noah Goodman, co-founder. Stanford professor of psychology and computer science with prior research tenure at Google DeepMind. Brings academic credentials in cognitive science and computational reasoning.

The team includes additional research and engineering hires from OpenAI, Meta AI, Reflection AI, the Allen Institute for AI (AI2), and MIT. No further senior leadership has been disclosed publicly.

Funding and backers

Humans& has raised a single funding round through April 2026: the $480 million seed announced on January 20, 2026 at a $4.48 billion post-money valuation. The round was led by SV Angel (Ron Conway's firm) and co-founder Georges Harik. Disclosed co-investors include Nvidia, Jeff Bezos (through Bezos Expeditions), Google Ventures, Emerson Collective (Laurene Powell Jobs's firm), and Forerunner Ventures. The structure has been described as "all cash, unstructured."

The round represents approximately 9x return on capital invested at face value, and was reported in industry coverage as among the largest seed rounds ever raised, trailing Thinking Machines Lab's $2 billion July 2025 seed but ahead of typical 2025-vintage senior-team founder-exodus seeds. The majority of the capital is reportedly being deployed to compute infrastructure for in-house model training.

The company has not disclosed cumulative compute commitments, cloud-partner agreements, or follow-on financing plans. The seed-round capital is sufficient to support a multi-year frontier-tier training program at current GPU prices, though the magnitude of the model-training program has not been publicly specified.

Industry position

Humans& occupies a structurally distinctive position among 2025-vintage Insurgent labs. The combination of a senior multi-lab founding team, a $4.48 billion seed valuation, and a deliberate human-collaboration positioning differentiated from autonomous-agent offerings is not directly mirrored at any other lab in the cohort. The closest peer comparators are Adaption Labs, founded by ex-Cohere senior researchers with a similarly research-first thesis, and the broader cohort of senior-departure labs with sub-$5 billion valuations.

The strategic risks are substantial. The company has not released a model, the technical direction has not been disclosed in detail, and the human-collaboration framing has not been validated against a shipping product. The $4.48 billion valuation depends on team credentials and strategic-investor signals rather than capability evidence.

The strategic strengths are equally distinctive. The founder team's combined research output spans the post-training of frontier models at Anthropic, xAI's Grok program, and Stanford academic research on cognitive computation. The strategic-investor list includes Nvidia and Google Ventures, providing potential paths to compute access and distribution that reduce dependence on a single cloud provider. The collaboration-first product positioning differentiates the company from the dominant chatbot and autonomous-agent product surfaces.

Competitive landscape

Humans& competes with several Frontier and Insurgent labs and product-layer companies:

• OpenAI. Frontier-model peer and the dominant consumer-chatbot product surface. ChatGPT's instant-messaging-like interaction model is the closest existing comparator to the Humans& product framing.
• Anthropic. Frontier-model peer, source of co-founder Andi Peng. Anthropic's Claude product and Claude Code agent compete on enterprise collaboration adjacent to the Humans& positioning.
• xAI. Source of co-founders Eric Zelikman and Yuchen He. Grok competes on consumer chatbot positioning with a similar human-and-AI-coordination narrative around X (formerly Twitter) integration.
• Thinking Machines Lab. Closest peer Insurgent. Both founded by senior multi-lab departures, both pursuing high-capital research-first strategies, both pre-product on in-house models. Thinking Machines emphasizes developer fine-tuning over collaboration.
• Adaption Labs. Closest peer in research-thesis terms among 2025-vintage Insurgents. Hooker and Roy emphasize adaptive efficient AI rather than collaboration coordination.
• Slack and Microsoft Teams. Existing collaboration platforms that may incorporate frontier AI features as competitive responses to Humans&.

Outlook

Several open questions affect Humans&'s trajectory in 2026 and 2027:

• The first product launch, including its capability profile, target customer segment, and pricing model.
• The first in-house model release, including the technical direction (architecture, training corpus, capability targets) and how it benchmarks against frontier-tier comparators.
• Whether the company accepts follow-on capital at a higher valuation, and at what mark.
• Compute-partner relationships, particularly any expansion of the Nvidia connection beyond the seed-round investment.
• Senior-talent recruitment momentum, particularly from Anthropic, xAI, OpenAI, and Meta AI.
• The durability of the human-collaboration positioning as competitor labs incorporate similar framings into existing chatbot and enterprise products.

Sources

• TechCrunch: Humans&, a 'human-centric' AI startup founded by Anthropic, xAI, Google alums, raised $480M seed round. Primary source on the founding team and seed round.
• Crunchbase News: Humans& Raises Huge $480M Seed Round At $4.48B Valuation For 'Human-Centric AI Lab'. Funding details and historical context.
• Implicator AI: Humans&, The $4.5 Billion Bet Against the Robot Takeover. Founding details, mission framing, and team backgrounds.
• Reuters via The Star: AI startup Humans& raises $480 million at $4.5 billion valuation in seed round. CEO Eric Zelikman's framing on AI coordination.
• AI Business: Humans& Raises $480M to Build Human-Centric AI Tools. Coverage of the human-centric positioning.
• Reworked: Humans& Lands $480M to Make AI Work With Humans. Coverage of the collaboration thesis.