America’s AI Action Plan is a blueprint for jump-starting federal AI investment in R&D and collaboration with the private sector. It focuses on accelerating American leadership across research, development, adoption, and use of cutting-edge AI. This is important and timely, given the emergence and acceleration of Frontier AI models. The plan offers 90-plus recommendations for implementation and commits the federal government to backing the efforts of government agencies as they work to build out their AI architectures for improving their agencies and the lives of the citizens they serve.
Three Pillars
The AI Action Plan is built on three pillars:
- Accelerating AI Innovation
- Building American AI Infrastructure
- Leading in International AI Diplomacy and Security
The Accelerating AI Innovation pillar directs federal agencies to create environments to speed development by removing bureaucratic “red tape” and regulatory barriers, among other recommended actions. It specifically calls upon agencies like the Office of Management and Budget (OMB) and the Federal Communications Commission (FCC) to identify and repeal rules that may be hindering AI deployment across the federal government, and it directs agencies to prioritize the prevention of vendor lock-in and to promote transparency in contracting and development.
The Building American AI Infrastructure pillar seeks to accelerate the private sector’s development of physical and technical infrastructure needed to power the world’s leading AI systems. It will drive this revolution through initiatives in data center and semiconductor manufacturing permit reform, energy generation and electric grid modernization, and skilled workforce training for the “trades of AI,” from electricians to HVAC technicians.
The Leading in International AI Diplomacy and Security pillar seeks to establish the U.S. AI technology stack as the globally dominant AI innovation standard, through an export strategy to establish secure US software application, model, and hardware packages among allies around the world.
“The USAi Toolkit”
To help federal agencies meet the goals of the three strategic pillars, the U.S. government established the USAi.gov platform. This website provides agencies with a technical toolkit composed of a centralized suite of resources, platforms, and policy frameworks developed to speed AI adoption. The suite includes vetted large language models (LLMs), tools to modernize government software, and policy-as-code guidelines to ensure agencies use AI in conjunction with all applicable federal laws.
AI and Private Cloud
Given the unique security and governance challenges faced by federal agencies, the deployment of AI using a private cloud platform offers tremendous support and reinforcement for the goals of the AI Action Plan. An AI private cloud strategy brings together the game-changing capabilities of world class AI with the core strengths of private cloud. It uniquely offers federal agencies the ease, scalability, interoperability, and security to quickly implement the AI action plan at the speed and velocity the plan demands.
Advantages of Private Cloud
Private cloud confronts the “three C’s” of public cloud — cost, complexity, and compliance — adding significant operational flexibility while reducing complexity, lowering costs, and simplifying compliance issues.
Reduces complexity. Private cloud is built on a software-defined, virtualized infrastructure that gives it the flexibility to extend hybrid cloud and multi-cloud services across the enterprise. In the case of Broadcom’s private AI platform, agencies can deploy and manage AI with the same tools and resources with which they are already familiar, leveraging the same skilled teams and their years of experience with Broadcom’s VMware tools. With private cloud, agencies don’t need to develop bespoke technology — they have what they already need in their private cloud subscriptions.
Lowers costs. Private cloud helps lower federal agencies’ costs, including eliminating the need to move their data across multiple clouds and pay unanticipated costs. Broadcom conducted a Total Cost of Ownership Analysis, which demonstrates that moving to software-defined infrastructure with VCF can dramatically reduce the total cost of ownership. VCF delivers TCO savings of 51% and 55% infrastructure cost savings compared to traditional infrastructure, and 40% TCO savings compared to native public cloud.
Security and compliance. A public cloud solution migrates sensitive data beyond a government agency’s control to computing resources and applications shared with and accessed by other entities. A private cloud platform allows agencies to bring LLMs to their own environments, keeping the agency’s data, computing resources, and applications within its own controlled environment, where no other entity has access. The security and compliance advantages of building sensitive and capableAI solutions within a private cloud platform are superior at their inception compared to even the most ideal public cloud circumstances.
AI Private Cloud Use Cases
A modern, agnostic private cloud AI solution like VMware Cloud Foundation (VCF) makes AI a core feature of private cloud. Under the agency’s own control and ability to manage securely, AI in private cloud contributes speed-to-value as well as observability and enhanced security of the agency’s data.
Deploying an agnostic AI private cloud incorporates and provides access to a full array of best-of-breed LLMs. An AI Agent Builder built into a private cloud platform can take full advantage of model galleries of LLMs and other resources. It also allows agencies to automate multiple steps of complex workflows. The use of model galleries, world class LLMs included, also allows agencies to pick-and-choose the ones that best suit the specific workflow challenge. Agencies enjoy true freedom of choice and flexibility with these models.
By bringing the LLM to the data in a private cloud, security is reinforced by the observability into that data afforded by a well-protected single plane-of-glass dashboard.
It’s also worth noting that several of the AI Action Plan’s requirements involve standing up large-scale, virtualized lab environments for scaling and testing novel AI capabilities.
The U.S. Cyber Command’s premier training system, the Persistent Cyber Training Environment (PCTE), already offers a great example of massively-scaled virtualization. Virtual training ranges within PCTE simulate real-world IT environments where cyber warriors practice how to detect, intercept, or mitigate attacks on IT infrastructure. And in fact, a private cloud platform provides the backbone of PCTE.
America’s AI Action Plan calls for government agencies to quickly ramp up their plans to deploy AI. AI in a private cloud platform offers agencies the most efficient and secure way to meet these goals.
Learn more about VMware Cloud Foundation 9.1