Explore how federated mesh architectures enable trusted, sovereign data and AI ecosystems that scale collaboration, data sharing, and value across organizations.

“In the future we’ll look back and see a clear demarcation: life before LLMs and life after.”

– Kirk Bresniker, HPE Fellow and Chief Architect at Hewlett Packard Labs

Large language models (LLMs) revealed what is possible for language-based human interaction with technology.

Agentic AI and retrieval-augmented generation (RAG) are currently revolutionizing the value capture processes on data.

Retrieval-augmented agentic AI

Onboarding data to RAG-based retrieval systems can create a comprehensive knowledge base combining unstructured and structured, internal and external sources and—critically—makes that knowledge available for language-based human interaction and efficient value capture.

“Nothing of significance was ever achieved by an individual acting alone. Look below the surface and you will find that all seemingly solo acts are really team efforts.”

 – John C. Maxwell

In a global, digital economy, value capture does not stop at national borders, industry lines, organizational boundaries, or domain limits. That landscape requires trust and digital sovereignty. Boundaries can create silos of data, trust, and processes across organizations, regulations, partners, joint ventures, and subsidiaries.

Taking on the challenge

The challenge is to overcome these boundaries as much as avoiding the inertia of traditional data spaces. Modern technology and the communities of the Cloud Native Computing Foundation (CNCF) provide all we need to create data and AI ecosystems maximizing the value of data and AI and limiting the technology and operational impact at the same time.

Our solution implements a federated mesh architecture based on CNCF open-source technology and open standards to create decentralized ecosystems. Participating organizations can connect services like RAG-based retrieval systems and agents. Imagine a multilateral federation dissolving the traditional boundaries of data, knowledge, and intellectual property to create a collective intelligence with built-in sovereignty, trust, control, and accountability.

Dedicated tailored agents and workflows unleash the potential of purpose-driven agentic AI connecting a catalog of services, tools, data, knowledge, and other agents including LLMs creating an ecosystem in which every idea can be realized.

As such, it does not stop at RAG systems and agents. To realize every idea, the federated mesh architecture supports any kind of cloud-native workload and application programming interfaces (APIs) making data and functionality ubiquitously available for usage including data streaming, messaging, graphs, as well as machine learning, swarm learning, and data analytics.

Maximizing value

The long‑term value of AI ecosystems will be defined by their ability to evolve. Our solution and the concept of federation enable organizations to scale AI adoption across data, use cases, and partners without accumulating technical debt or governance friction. As new participants, capabilities, and regulations emerge, ecosystems can adapt incrementally—rather than requiring disruptive platform redesigns.

By grounding collaboration in open, cloud‑native standards, organizations create future‑ready foundations for innovation. This supports faster experimentation, repeatable deployment of AI services, and sustained collaboration across boundaries of industry, geography, and trust. The result is durable value creation: ecosystems that remain relevant as AI models, regulations, and business priorities continue to change.

While the operating model certainly depends on the ecosystem participants, the business case, and governance model, the federated mesh reduces the operational cost and technology leap required with its highly standardized technology stack and platform based on open standards.

Real-world example

A prime example is the Global Data Partnership against Forced Labour (GDPFL). Forced labour remains one of the world’s most persistent and systemic labour rights challenges embedded across global supply chains and societies.¹

The GDPFL leverages the federated mesh architecture to build a secure, federated data and AI ecosystem that enables organizations to share insights without giving up sovereignty and control of their sensitive information.

Built on a federated mesh architecture with retrieval‑augmented, agentic AI, the GDPFL ecosystem connects fragmented data and services without moving or centralizing them. Organizations contribute insights from where data is generated—ranging from public and non‑sensitive evidence to advanced-analytics services—while retaining full control, governance, and accountability over their assets.

In practice, this can create a collective intelligence layer where retrieval‑augmented services and purpose‑built agents can surface patterns, risks, and signals across distributed sources, accelerating insight without compromising sovereignty. Lowering technical and trust barriers to participation allows the ecosystem to scale sustainably—onboarding new partners, data, and AI capabilities over time—while remaining transparent, rights based, and resilient by design.

Conclusion

As organizations move beyond experimentation with LLMs toward systems that must operate across boundaries of trust, regulation, and ownership, the future of AI will be defined not by isolated models but by ecosystems. Federated, agentic AI architectures make it possible to transform data into shared value—without sacrificing control, accountability, or sovereignty. By grounding innovation in open standards and cloud‑native technologies, HPE is helping organizations build AI ecosystems that are scalable, sustainable, and ready for real‑world impact.

Learn more at HPE Data Services and Enterprise Security and Digital Protection Services

^1”Harnessing Data and Intelligence for Collective Advantage: Ending Forced Labour in Global Supply Chains.” World Economic Forum, January 2026.

Meet the author:

Florian Buehr, HPE Solution Architect, Cybersecurity Services, HPE

Florian joined HP Services in September 1999 and has over 25 years of experience in IT consulting. As a solution architect, he specializes in secure IT architectures and solutions for the data and AI, with a focus on multilateral ecosystems, value capture, and emerging technologies. His work spans strategy, architecture, and delivery, integrating cybersecurity by design to enable trusted, scalable, and resilient solutions across complex business environments.