Isara

Isara is an American artificial intelligence company headquartered in San Francisco, California, founded in June 2025 by Eddie Zhang (former OpenAI AI-safety researcher and former MIT visiting researcher) and Henry Gasztowtt (Oxford computer-science student). The company is building swarms of cooperating AI agents that can coordinate on complex problems, with early demonstrations involving roughly 2,000 agents working together to forecast gold prices. Isara raised $94 million in its first disclosed round at a $650 million valuation, with OpenAI participating alongside Amity Ventures, Michael Ovitz, and Stanley Druckenmiller.

At a glance

• Founded: June 2025 in San Francisco, California, by Eddie Zhang and Henry Gasztowtt.
• Status: Private. First disclosed round at $650 million post-money valuation.
• Funding: $94 million in 2025 with OpenAI, Amity Ventures, Michael Ovitz, and Stanley Druckenmiller participating, at a $650 million post-money valuation.
• CEO: Eddie Zhang, Co-Founder. Former OpenAI AI-safety researcher; former MIT visiting researcher.
• Other notable leadership: Henry Gasztowtt, Co-Founder. Oxford University computer-science student; Christ Church graduate.
• Open weights: No. Isara has not released open-weights models or open-source agent infrastructure.
• Flagship outputs: Multi-agent swarm orchestration platform; demonstration deployments including the roughly 2,000-agent gold-price-forecasting demo.

Origins

Isara was founded in June 2025 by Eddie Zhang and Henry Gasztowtt, both 23 years old at the time of founding. Zhang had been an AI-safety researcher at OpenAI before founding Isara, with a prior visiting-research stint at MIT. Gasztowtt was a Christ Church College Oxford computer-science student. The founder pair's combination of an OpenAI-safety pedigree and an Oxford-undergraduate research-and-mathematics background gave the company an unusual demographic profile relative to the post-2022 AI insurgent cohort, which has been dominated by senior research-credibility-led founder teams.

The founding thesis was that the next phase of AI capability would not come from larger single models but from coordinating large numbers of frontier-model-driven agents on a shared task. Where most agent products in the 2024 to 2025 wave deployed one or a few agents per task, Isara's premise was that swarms of thousands of agents, each with specialized tools and roles, would deliver capability that single-agent or few-agent systems could not match. The company's early demonstration involved roughly 2,000 AI agents coordinating to produce a gold-price forecast, framed as a proof of concept for the swarm-orchestration thesis rather than a commercial product.

The first disclosed funding round of $94 million at a $650 million post-money valuation was announced in late 2025 with OpenAI participating alongside Amity Ventures, Michael Ovitz, and Stanley Druckenmiller. The OpenAI investment is structurally distinctive: OpenAI rarely invests directly in agent-product companies that consume its own API, and the participation signaled the seriousness of the strategic interest in the swarm-orchestration thesis. The Druckenmiller and Ovitz participation gave the round a Wall Street-and-Hollywood high-net-worth-investor visibility that few seed-stage AI companies attract.

Mission and strategy

Isara's stated mission is to build coordinating-agent systems that can solve complex problems by cooperation across large agent populations rather than by scaling individual model capability. The strategic premise is that the agentic-AI market is at a comparable moment to the language-model market in 2018, with the shape of the dominant agent paradigm not yet settled and the swarm-orchestration approach offering a structurally different bet from the single-agent and few-agent products dominating the 2024 to 2025 commercial wave.

The strategy combines two threads. First, the multi-agent-swarm orchestration platform, with the company building infrastructure for tens to thousands of cooperating agents per task rather than per-agent product surfaces. Second, applied demonstrations against high-value problem categories such as financial forecasting, where the swarm approach can be benchmarked against peer agent and quantitative-finance baselines.

The competitive premise positions Isara within the agent-native insurgent cohort alongside Cognition AI, Imbue, and a long tail of well-funded post-2024 agent startups, with the swarm-orchestration thesis providing distinctive technical positioning relative to peers.

Models and products

• Multi-agent swarm orchestration platform. Infrastructure and runtime for coordinating large agent populations on shared tasks. Specifics have been comparatively under-disclosed.
• Demonstration deployments. The roughly 2,000-agent gold-price-forecasting demonstration; additional applied demonstrations in development.
• Underlying frontier-model dependency. Isara has not disclosed building its own foundation models; the swarm platform consumes frontier-model capability through partner relationships including the OpenAI investor relationship.

Distribution channels have not been disclosed publicly beyond the demonstration deployments. Industry coverage has indicated investment-firm and quantitative-finance early-customer interest.

Benchmarks and standing

Isara's evaluation framework is not benchmarked against horizontal foundation-model leaderboards because the company's commercial output is multi-agent orchestration capability rather than language-model performance. Public standing rests on the demonstration outputs, the OpenAI strategic-investor relationship, and the Druckenmiller and Ovitz participation in the round. Industry coverage has characterized the company as one of the watchable agent-native insurgent labs of the 2025 cohort, with the OpenAI investment providing distinctive provenance.

The agentic AI market has been projected to grow from roughly $7.8 billion in 2025 to over $50 billion by 2030 according to industry research, with every major frontier model platform shipping multi-agent features in 2025 and 2026. Isara's positioning is as an agent-native insurgent rather than a feature-of-a-frontier-platform, which puts the company in direct competition with both insurgent peers and frontier-platform agent features.

Leadership

As of May 2026, Isara's senior leadership includes:

• Eddie Zhang, Co-Founder. Former OpenAI AI-safety researcher; former MIT visiting researcher.
• Henry Gasztowtt, Co-Founder. Oxford University Christ Church computer-science student.
• Senior engineering leadership across the swarm-orchestration platform.

The two co-founders anchor most public-facing visibility, with the company comparatively quiet on senior-leadership disclosures beyond the founder pair.

Funding and backers

• Seed (2025): $94 million at $650 million post-money valuation. OpenAI, Amity Ventures, Michael Ovitz, and Stanley Druckenmiller participating.

Cumulative disclosed private capital is the $94 million seed round; no subsequent rounds have been disclosed as of May 2026.

The investor cohort is structurally distinctive across three axes. OpenAI rarely participates as a strategic investor in companies that consume its API at meaningful scale, and the participation in Isara signals strategic interest in the multi-agent-orchestration thesis as a platform layer rather than a product feature. Amity Ventures provides early-stage AI-focused venture capital. Michael Ovitz, the former Creative Artists Agency co-founder and the founder of Disney's broader entertainment-and-tech investment activity, brings senior media-and-entertainment-industry visibility. Stanley Druckenmiller, the senior macro-hedge-fund investor whose Duquesne Family Office investment activity is closely watched in financial markets, brings senior Wall-Street-investor visibility. The combined cohort is unusual for a seed-stage AI company and reflects the cross-industry interest in the agentic-AI category as well as the founder pair's trajectory.

Industry position

Isara occupies a structurally distinctive position among 2025-cohort agent-native insurgent labs: the OpenAI strategic investment, the high-net-worth-individual participation from Druckenmiller and Ovitz, and the swarm-orchestration thesis combine into an unusually visible early-stage profile. Industry coverage has characterized Isara as one of the principal post-frontier insurgent agent labs to emerge from the 2025 wave.

The structural risks are characteristic of seed-stage agent companies. First, the swarm-orchestration thesis is unproven at production scale: whether thousands of cooperating agents deliver materially better results than well-orchestrated few-agent systems on commercially valuable workloads is the principal commercial question. Second, the OpenAI strategic-investor relationship cuts both ways: OpenAI itself ships agent features (ChatGPT Agent Mode, Operator) that compete for the same enterprise mindshare, and the dependence on frontier-model capability through partner relationships means Isara's economics are sensitive to frontier-model pricing. Third, the founder team is comparatively junior by AI-insurgent standards, which is both a recruiting advantage (in a generationally hot category) and a risk in a capital-intensive long-cycle business.

Competitive landscape

• Cognition AI, Imbue, General Reasoning. Agent-native insurgent peers with different agent-orchestration approaches.
• Adept AI. Earlier-cohort agent insurgent; senior leadership transitioned to Amazon in 2024.
• OpenAI (Operator, ChatGPT Agent Mode), Anthropic (Claude Computer Use, Claude Code), Google DeepMind (Project Mariner). Frontier-platform agent features competing for the same enterprise mindshare; OpenAI is also a strategic investor.
• Quantitative-finance and investment-firm AI teams. The gold-price demonstration competes against established quantitative-finance baselines from D. E. Shaw, Renaissance, and adjacent investment-firm AI teams.

Outlook

• The pace and surface of any disclosed commercial deployments beyond the early demonstration outputs.
• Whether the swarm-orchestration thesis converts into measurable advantages over single-agent and few-agent peer products on commercially valuable workloads.
• The Series A or adjacent fundraising timeline beyond the seed round.
• The competitive dynamics with OpenAI, given OpenAI is both strategic investor and agent-product competitor.
• Continued senior engineering recruitment as the company scales beyond the founding pair.

Sources

• The Next Web: OpenAI invests in Isara. Funding-round coverage.
• Tech Funding News: OpenAI backs Isara $94M. Round details and founder background.
• Christ Church Oxford: Christ Church graduate co-founds startup. Henry Gasztowtt background reference.
• VKTR: OpenAI-Backed Isara Raises $94M. Industry coverage.
• Shashi.co: Isara and the Bet on AI Agent Swarms. Enterprise-buyer perspective.