Bengaluru, 20th November, 2025: NimbleEdge, the on-device AI infrastructure company
reshaping how intelligence runs on mobile, today announced its collaboration with Microsoft
in bringing Foundry Local to Android, revealed on stage at Microsoft Ignite 2025 in San
Francisco. The launch marks a major leap for the global developer community, enabling truly
powerful, low-latency, privacy-preserved AI to run directly on smartphones at scale.
One of India’s largest digital payments platforms has become the first enterprise to adopt
Foundry Local for Android, enabling it to bring agentic experiences into everyday consumer
applications while preserving user privacy.
Foundry Local introduces optimized small language models to Android, allowing developers
to build agentic apps that reason, respond, and collaborate with users in real time. By
executing models locally through a unified, optimized runtime, Foundry Local ensures that
intelligence stays on the device, resulting in low-latency performance, offline reliability, and
privacy-aware architecture.
Rajat Monga, CVP, AI Frameworks, Microsoft, stated, “At Microsoft, we believe the next
breakthrough in AI lies not only in building smarter models, but in making AI scalable,
efficient, and accessible. With Foundry Local for Android, built with contributions from
NimbleEdge, we’re unlocking a future where developers can build real-time AI native app
experiences with powerful, low-latency intelligence on users’ phones.”
NimbleEdge is contributing in shaping Microsoft’s Foundry Local into a secure, on-device AI
layer – described as a “mini AI server inside your phone.” The two companies have a long-
standing collaboration, starting from the ONNX runtime to scaling deployments to 30M+
devices. NimbleEdge architected the background service that handles robust long-running
SLM downloads, shared resources, and inference over a secure-bound AIDL service with
mutual certificate verification, across diverse Android devices. NimbleEdge’s DeliteAI works
with Foundry Local to simplify deployment and scaling challenges, while orchestrating real-
time agentic workflows on user interactions.
Neeraj Poddar, Co-founder & CTO, NimbleEdge, shares that “This collaboration brings
our core mission to life: enabling real-time, on-device AI for billions of users. With Foundry
Local on Android, developers can offer secure and personalized experiences without the
cost and latency of cloud-based inference. We’re excited to see global developers across
the world build the next generation of mobile AI applications.”
Baked from witnessing high-frequency event streams on edge devices, DeliteAI, by working
with Foundry Local provides developers with key advantages in:
1. In-session Context Engineering: Capture and process event streams directly from
the user interface for real-time context.
2. Advanced Agent Orchestration: Facilitate agent orchestration, prompt templating,
tool calling, persistent memory, and voice interactions.
3. Unified Performance: Ensure consistent performance across various hardware,
along with cross-compatible cloud-edge inference APIs.
“The next wave of AI will be driven by agents that collaborate in real time with humans and
other AI agents alike – and that requires an on-device architecture,” said Varun Khare, Co-
founder & CEO, NimbleEdge. “At NimbleEdge, we’ve been building the infrastructure
catering to the needs of mobile and edge devices. By unifying fragmentation across devices,
chipsets, and operating systems, Foundry Local and DeliteAI support developers in
deploying AI agents locally while upholding user privacy.”
On-device intelligence is, by design, more supportive of a dynamic, evolving environment
found in modern mobile applications. The collaboration envisions a global shift toward AI that
is faster, more efficient, and fundamentally user-centric.