H2O.ai has joined the Open Secure AI Alliance, an industry effort led by NVIDIA to develop open technologies and tools for securing AI agents.
The move reflects a shared belief that open AI strengthens not only innovation but also security and control, particularly important as AI agents increasingly access sensitive data and critical workflows.
H2O.ai brings expertise in securing agentic systems for highly regulated industries where controlling agent permissions and guardrails is essential.
What happened
H2O.ai announced on August 11, 2026, that it has joined the Open Secure AI Alliance alongside NVIDIA and other industry leaders to develop open technologies and tools that strengthen the security of software and AI agents. The company also backs the principles outlined in the Open Weights and American AI Leadership letter.
Why it matters
As AI agents gain access to sensitive data, applications and critical workflows, security requires controls across the full agent stack — not just whether a model is open or closed. H2O.ai brings experience securing agentic AI for customers in highly regulated industries where controlling what an agent can access and do is critical. Open weights models and open security tools give organizations greater visibility into the models they use, the freedom to adapt and run them on their own infrastructure, and the ability to test, adapt and strengthen controls.
What to watch
H2O.ai will work with NVIDIA and other alliance members to advance open approaches to AI security and help organizations deploy AI systems they can trust, govern and control. The company has a track record of releasing open models, including H2O-Danube and H2O-Mississippi, with contributions every year since inception.
H2O.ai, a provider of enterprise AI platforms for predictive, generative and agentic AI applications, announced on August 11, 2026, that it has joined the Open Secure AI Alliance. The alliance, led by NVIDIA and comprising other industry leaders, aims to develop and share open technologies, techniques and tools designed to strengthen the security of software and AI agents.
The move also signals H2O.ai's backing of the principles outlined in the Open Weights and American AI Leadership letter, which emphasizes the role of open models in strengthening safety, cybersecurity, innovation and AI sovereignty. According to Sri Ambati, Founder and CEO of H2O.ai, the decision reflects a core conviction: "Open source is about freedom, not just free. Sovereign AI is only possible with open weights and distributed shared philosophy of innovation." He further stated that "Every nation needs to be an AI nation. Every organization, enterprise, and every government needs to be AI-native and make AI their own, to govern, and learn from the feedback loops of continuous use of AI."
The alliance's work addresses a practical security challenge: as AI agents gain access to sensitive data, applications and critical workflows, security must extend far beyond the question of whether a model is open or closed. Protecting these systems requires controls across the full agent stack — including agent identity, permissions, guardrails, evaluation and auditability — as well as the ability for defenders to test, adapt and strengthen those controls. H2O.ai brings to the alliance its experience securing agentic AI systems for customers in highly regulated industries, where controlling what an agent can access and do is especially critical. Through its participation, the company will work with NVIDIA and other members to advance open approaches to AI security and help organizations deploy AI systems they can trust, govern and control.
H2O.ai's entry into the Open Secure AI Alliance marks a broader industry recognition that AI security extends beyond the question of whether models are open or closed — it requires comprehensive controls across the entire agent stack. As AI agents increasingly interact with sensitive data and critical business workflows, particularly in regulated industries, the ability to govern and audit these systems becomes paramount. H2O.ai's participation is grounded in its existing customer base in highly regulated sectors, where the company has already developed expertise in controlling agent permissions, identity, and guardrails.
The alliance's focus on open security tools reflects a philosophy that transparency and auditability strengthen defense. By joining NVIDIA and other industry leaders in this effort, H2O.ai signals its commitment to an ecosystem where organizations can understand, inspect and adapt their AI systems rather than rely on proprietary black boxes. The company's history of releasing open-source models — such as H2O-Danube and H2O-Mississippi — positions it as a credible contributor to this shared initiative on security and sovereignty.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
Palantir Technologies reported Q2 revenue of $1.94 billion, beating analyst estimates of $1.81 billion by 6.7%…
Synopsys unveiled fully autonomous engineering software built with Nvidia at the July 26 DAC conference, compr…
Anthropic is embedding imperceptible, machine-readable watermarks into text generated by Claude models release…
The Information reported that OpenRouter, an AI model aggregator, is attracting acquisition interest
HeyDeacon, an AI customer support agent, went live to help SaaS founders serve their users
NVIDIA announced memorandums of understanding with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, a…
The AI news that matters, in one minute each morning.
Sign up free