Originally published Making edge AI work for government on by https://federalnewsnetwork.com/commentary/2025/08/agentic-ai-is-redefining-software-development-in-government-2/ at Federal News Network
Federal IT leaders have seen game-changing technology before. Mobile devices. Cloud computing. Each arrived with enormous potential and a long list of questions. Who owns it? Who secures it? What happens to the data?
Now AI PCs are shaking things up. This new class of endpoint devices can run powerful artificial intelligence models locally, eliminating the need for network access. They could be game-changers for tactical and edge environments, or they could be another innovation delayed by uncertainty. What happens next will depend entirely on whether agencies move with purpose or wait for policy to catch up to them.
AI PCs integrate traditional central processing units (CPUs) and graphics processing units (GPUs) with a third engine: neural processing units (NPUs) designed specifically for AI inference. The hardware matters, but the real shift is the ability to run real-time AI workloads locally without relying on the cloud, draining battery or slowing down performance. It’s a major advantage anywhere speed, autonomy and privacy matter.
A shift from refresh to reinvention
AI PCs process workloads in a smarter way. Thanks to orchestration software, each task is routed to the most efficient engine — CPU, GPU or NPU — to maximize performance and efficiency.
Some of these changes show up in small, everyday ways: Background noise fades on a video call, captions appear automatically and meeting notes draft themselves. But the same intelligence can enable more powerful use cases for the people doing critical work in the field.
With the right models onboard, AI PCs can help frontline personnel translate languages, analyze sensor data, detect threats or anticipate maintenance needs, especially in a denied, disconnected, intermittent or limited (DDIL) environment. And because this happens on-device, they don’t need a signal to get results. Soldiers or field technicians can find the answer they’re looking for when it’s needed.
Edge-ready is just the starting point
Disconnected operations demand more autonomy, not less.
Picture a young technician at a forward operating base holding a rugged laptop up to a damaged aircraft part. Instead of flipping through manuals or waiting for a signal, they point a camera and ask, “What is this?” A lightweight AI model on the device instantly identifies the component and pulls up the repair procedure. The job gets done.
Or imagine an aircraft or drone offloading terabytes of sensor data at a forward location. Rather than shipping that data back to a data center, an AI PC in the field can filter and analyze it on the spot. In addition to flagging anomalies, users can use natural language queries like “what changed?” to get faster, clearer insights.
These aren’t speculative scenarios. These are use cases that are already being explored, piloted, and in some cases deployed. And they point to a bigger opportunity: What else could agencies do with computing power that’s both smarter and closer to the mission?
If this sounds familiar, it should. The government’s adoption of mobile devices and cloud computing followed a similar arc. Early conversations rightly centered on security, ownership and control. But once those tools were integrated, they quickly became mission-essential. And then creativity took over, with staff discovering new uses once the tech was in place. The same will likely hold true for AI PCs.
Making AI work at the edge
Agencies must get two things right for AI PCs to deliver in disconnected, high-pressure or high-security environments: the models and the security.
Cloud-scale AI may dominate headlines, but the edge calls for something lighter and more nimble. Pre-trained models in the 3 billion to 10 billion parameter range can be optimized ahead of time to run offline at the source.
At the same time, AI PCs bring a new kind of trust model. By shifting processing and data containment to the device, they reduce the exposure to and reliance on external networks. Features like secure boot, encrypted memory and workload isolation work alongside NPUs that offload tasks like malware scanning or anomaly detection.
On-device inference can also lower operational costs by reducing cloud data calls and minimizing the need to transport or retransmit data. And because sensitive information stays local, agencies gain stronger control over data sovereignty and compliance.
The private sector isn’t waiting
The private sector is already moving fast. Companies in content creation, collaboration, cybersecurity and IT operations are actively building software that takes advantage of AI PCs. Used for more than flashy demos, they’re driving real improvements like deep fake detection, local language processing and predictive endpoint monitoring.
In fact, the number of commercial applications optimized for AI PCs has grown from just a few dozen to over 400 in under two years. The ecosystem is maturing fast. Government leaders have a chance to shape how those same capabilities are applied to the mission edge.
Translation at the edge. Local generative tools for ops planning. Adaptive assistance for frontline personnel. Many of these capabilities already exist. Others are just one update away.
What makes AI PCs exciting isn’t just what we know they can do; it’s what we haven’t imagined yet. Like mobile and cloud before them, their value will emerge through the new workflows, insights and efficiencies agencies uncover. The trajectory is clear. The only question is: Are you ready to use them to do more than you could yesterday?
Andrew Awalt is industry technical specialist for Defense and national security at Intel Corporation.
Originally published Making edge AI work for government on by https://federalnewsnetwork.com/commentary/2025/08/agentic-ai-is-redefining-software-development-in-government-2/ at Federal News Network
A 2021 law provided $551 billion to the Department of Transportation for about 100 grant programs. Most of the funds will go to state and local governments and others to build things like roads and bridges.
About Us
To assist commercially facing small and startup technology companies, and help determine if there is value in engaging with defense, intelligence community.
Making edge AI work for government
Originally published Making edge AI work for government on by https://federalnewsnetwork.com/commentary/2025/08/agentic-ai-is-redefining-software-development-in-government-2/ at Federal News Network
https://federalnewsnetwork.com/wp-content/uploads/2024/10/GettyImages-1628553826-1024x683.jpgFederal IT leaders have seen game-changing technology before. Mobile devices. Cloud computing. Each arrived with enormous potential and a long list of questions. Who owns it? Who secures it? What happens to the data?
Now AI PCs are shaking things up. This new class of endpoint devices can run powerful artificial intelligence models locally, eliminating the need for network access. They could be game-changers for tactical and edge environments, or they could be another innovation delayed by uncertainty. What happens next will depend entirely on whether agencies move with purpose or wait for policy to catch up to them.
AI PCs integrate traditional central processing units (CPUs) and graphics processing units (GPUs) with a third engine: neural processing units (NPUs) designed specifically for AI inference. The hardware matters, but the real shift is the ability to run real-time AI workloads locally without relying on the cloud, draining battery or slowing down performance. It’s a major advantage anywhere speed, autonomy and privacy matter.
A shift from refresh to reinvention
AI PCs process workloads in a smarter way. Thanks to orchestration software, each task is routed to the most efficient engine — CPU, GPU or NPU — to maximize performance and efficiency.
Some of these changes show up in small, everyday ways: Background noise fades on a video call, captions appear automatically and meeting notes draft themselves. But the same intelligence can enable more powerful use cases for the people doing critical work in the field.
With the right models onboard, AI PCs can help frontline personnel translate languages, analyze sensor data, detect threats or anticipate maintenance needs, especially in a denied, disconnected, intermittent or limited (DDIL) environment. And because this happens on-device, they don’t need a signal to get results. Soldiers or field technicians can find the answer they’re looking for when it’s needed.
Edge-ready is just the starting point
Disconnected operations demand more autonomy, not less.
Picture a young technician at a forward operating base holding a rugged laptop up to a damaged aircraft part. Instead of flipping through manuals or waiting for a signal, they point a camera and ask, “What is this?” A lightweight AI model on the device instantly identifies the component and pulls up the repair procedure. The job gets done.
Or imagine an aircraft or drone offloading terabytes of sensor data at a forward location. Rather than shipping that data back to a data center, an AI PC in the field can filter and analyze it on the spot. In addition to flagging anomalies, users can use natural language queries like “what changed?” to get faster, clearer insights.
These aren’t speculative scenarios. These are use cases that are already being explored, piloted, and in some cases deployed. And they point to a bigger opportunity: What else could agencies do with computing power that’s both smarter and closer to the mission?
If this sounds familiar, it should. The government’s adoption of mobile devices and cloud computing followed a similar arc. Early conversations rightly centered on security, ownership and control. But once those tools were integrated, they quickly became mission-essential. And then creativity took over, with staff discovering new uses once the tech was in place. The same will likely hold true for AI PCs.
Making AI work at the edge
Agencies must get two things right for AI PCs to deliver in disconnected, high-pressure or high-security environments: the models and the security.
Cloud-scale AI may dominate headlines, but the edge calls for something lighter and more nimble. Pre-trained models in the 3 billion to 10 billion parameter range can be optimized ahead of time to run offline at the source.
At the same time, AI PCs bring a new kind of trust model. By shifting processing and data containment to the device, they reduce the exposure to and reliance on external networks. Features like secure boot, encrypted memory and workload isolation work alongside NPUs that offload tasks like malware scanning or anomaly detection.
On-device inference can also lower operational costs by reducing cloud data calls and minimizing the need to transport or retransmit data. And because sensitive information stays local, agencies gain stronger control over data sovereignty and compliance.
The private sector isn’t waiting
The private sector is already moving fast. Companies in content creation, collaboration, cybersecurity and IT operations are actively building software that takes advantage of AI PCs. Used for more than flashy demos, they’re driving real improvements like deep fake detection, local language processing and predictive endpoint monitoring.
In fact, the number of commercial applications optimized for AI PCs has grown from just a few dozen to over 400 in under two years. The ecosystem is maturing fast. Government leaders have a chance to shape how those same capabilities are applied to the mission edge.
Translation at the edge. Local generative tools for ops planning. Adaptive assistance for frontline personnel. Many of these capabilities already exist. Others are just one update away.
What makes AI PCs exciting isn’t just what we know they can do; it’s what we haven’t imagined yet. Like mobile and cloud before them, their value will emerge through the new workflows, insights and efficiencies agencies uncover. The trajectory is clear. The only question is: Are you ready to use them to do more than you could yesterday?
Andrew Awalt is industry technical specialist for Defense and national security at Intel Corporation.
Copyright
© 2025 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.
Originally published Making edge AI work for government on by https://federalnewsnetwork.com/commentary/2025/08/agentic-ai-is-redefining-software-development-in-government-2/ at Federal News Network
Originally published Federal News Network
Related Posts
Making edge AI work for government
Originally published Making edge AI work for government on August 6, 2025 03:43
House committee’s top Democrat presses for details on HHS layoffs
Originally published House committee’s top Democrat presses for details on HHS layoffs on
GAO: Report Finds DOT Needs to Better Track and Communicate Infrastructure Act Funding
A 2021 law provided $551 billion to the Department of Transportation for about 100 grant programs. Most of the funds will go to state and local governments and others to build things like roads and bridges.
About Us
To assist commercially facing small and startup technology companies, and help determine if there is value in engaging with defense, intelligence community.
Let’s Socialize
Making edge AI work for government
House committee’s top Democrat presses for details on HHS layoffs