The collaboration announced on June 30, 2026 brings together Capgemini’s deep experience in enterprise AI and digital transformation with Duality Technologies’ suite of privacy‑enhancing technologies (PETs). The joint effort aims to give organizations the ability to run AI workloads—including large language models—on data that remains under strict control, without moving it into a centralized repository.
Why the partnership matters
Enterprises in sectors such as healthcare, finance, defense, critical infrastructure, and government are increasingly hamstrung by data silos. Sensitive information is often scattered across multiple jurisdictions, each with its own privacy, security, and sovereignty regulations. Traditional approaches that centralize data for model training clash with these constraints, forcing companies to choose between compliance and AI ambition.
By marrying Capgemini’s consulting and implementation capabilities with Duality’s secure computation platform, the alliance promises a practical pathway: AI and analytics can be performed directly on encrypted or otherwise protected datasets, preserving ownership and governance while still extracting value.
Technical angle
Duality’s platform leverages a combination of homomorphic encryption, secure multi‑party computation, and trusted execution environments to enable “trusted computation.” In practice, this means data never leaves its original location in a readable form, yet algorithms can still operate on it as if it were plain text. The technology is positioned to support a range of workloads—from statistical analytics to the training and inference of large language models—without exposing raw data to downstream systems.
Capgemini will integrate these capabilities into its AI practice, offering clients a turnkey solution that includes strategy, architecture design, and deployment. The partnership is not limited to a single industry; it is presented as a cross‑sector offering that can be tailored to the regulatory nuances of each market.
Executive perspectives
“As organizations accelerate AI adoption, they are discovering that the most valuable data is often the hardest to access,” said Dr. Alon Kaufman, CEO and Co‑founder of Duality Technologies. “Together with Capgemini, we are helping organizations unlock intelligence from sensitive and distributed data while maintaining the privacy, security, and sovereignty controls required in highly regulated environments.”
“Aggregating data can be a barrier to analytical ambition and AI‑driven insight, introducing operational complexity and risk,” noted Andy Lea, Vice President, Capgemini Invent. “By combining Capgemini’s enterprise transformation expertise with Duality’s advanced privacy‑enhancing technologies, we help organizations unlock value from sensitive data while preserving privacy, accelerating compliance, and enabling trusted collaboration.”
These statements underscore a shared belief that data governance will become a decisive factor in AI competitiveness. The alliance positions both firms to capture a segment of the market that is currently underserved by conventional cloud‑centric AI services.
Market implications
The move arrives at a time when regulators worldwide are tightening rules around cross‑border data flows and AI transparency. Europe’s GDPR, the United States’ evolving privacy statutes, and emerging data‑sovereignty policies in Asia create a patchwork of compliance requirements that complicate traditional AI pipelines.
By offering a solution that processes data in place, the Capgemini‑Duality partnership could appeal to enterprises looking to sidestep costly data‑localization projects. Moreover, the ability to train and run LLMs on protected datasets may reduce the reliance on massive, publicly available corpora, which often raise concerns about bias and proprietary information leakage.
From a competitive standpoint, the collaboration pits the duo against other players that are building “confidential computing” services on public clouds, such as Microsoft’s Azure Confidential Compute and Google Cloud’s Confidential VMs. However, Duality’s focus on a broader suite of PETs—beyond hardware enclaves—could provide a differentiator for organizations with heterogeneous on‑premise and multi‑cloud environments.
Practical considerations for adopters
- Integration complexity: Capgemini’s consulting arm will likely handle the orchestration of Duality’s APIs within existing data architectures, but legacy systems may still require custom adapters.
- Performance trade‑offs: Secure computation introduces computational overhead. While Duality claims near‑native performance for many workloads, real‑world benchmarks will be essential.
- Governance models: The platform’s ability to enforce fine‑grained access controls aligns with emerging “data‑trust” frameworks, but policies must be defined up‑front to avoid bottlenecks.
- Skill requirements: Teams will need expertise in cryptographic techniques and secure AI pipelines, a niche skill set that may affect rollout timelines.
Looking ahead
If the partnership delivers on its promise, it could accelerate the shift from AI pilots to production‑grade deployments in regulated environments. The ability to tap into previously inaccessible data reservoirs may unlock new use cases—such as federated clinical research, cross‑institutional fraud detection, and joint defense analytics—while keeping sensitive information under the custodianship of its owners.
The collaboration also signals a broader industry trend: privacy‑preserving AI is moving from academic research into commercial offerings. As more enterprises confront the reality of data‑centric regulation, solutions that reconcile security with scalability are likely to gain traction.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
