Why Former Energy and NSA Chiefs Are Joining the AI Infrastructure Race

BOSTON, MA – November 12, 2026 – The artificial intelligence arms race has quietly entered a new, highly pragmatic phase. For the past few years, the dominant narrative was a frantic scramble for silicon—companies hoarding highly coveted graphics processing units (GPUs) to train increasingly massive models. Today, the bottleneck has shifted from the supply chain to the power grid and the security perimeter.

This pivot is perfectly encapsulated by the recent strategic moves of Boston-based CambridgeNexus Inc., an AI Factory-as-a-Service (AIFaaS) operator. The company recently announced the appointment of retired U.S. Navy Admiral Michael Rogers, technology entrepreneur Brad Hunstable, and former U.S. Energy Secretary Ernest Moniz to its Advisory Committee.

The enlistment of prominent former federal national security and energy officials highlights a stark reality: AI infrastructure scaling is no longer just a compute race. It is now a critical infrastructure challenge involving grid capacity, federal energy policy, cybersecurity, and sovereign-scale capital deployment.

“Building AI infrastructure at scale requires more than technology. It demands expertise in energy, capital investment and the operational challenges of deploying complex systems,” said Ken Huang, founder and CEO of CambridgeNexus. “These distinguished leaders bring extraordinary experience across several critical disciplines, and their perspectives will be invaluable as we expand our capabilities and help customers bring increasingly sophisticated AI workloads into production.”

The Power Bottleneck: Diplomacy in the Grid

As data center electricity demands surge, AI compute providers are running into hard limits on power availability and grid interconnects. A single large-scale AI data center can now require between 100 megawatts and a full gigawatt of power. Consequently, global data center electricity consumption is projected to more than double by 2030, creating a "time to power" crisis where interconnection wait times stretch beyond five years in many regions.

The appointment of Ernest Moniz signals that future AI infrastructure winners will likely be decided by energy policy and grid access rather than chip supply alone. Moniz, who served as the 13th U.S. Secretary of Energy and is a noted physicist at the Massachusetts Institute of Technology, brings deep expertise in electricity infrastructure and public-private collaboration.

Industry analysts note that the inability of public grids to deliver sufficient power is forcing a pivot toward direct investment in dedicated, on-site power generation. This includes exploring nuclear energy, particularly Small Modular Reactors (SMRs), which offer the constant, clean, and reliable baseload electricity that AI factories demand.

“AI is creating new opportunities for innovation across the energy and technology sectors,” Moniz said. “Supporting its growth will require investment, collaboration and practical approaches to infrastructure development. CambridgeNexus has an opportunity to contribute to that progress as demand for advanced computing continues to expand.”

For an infrastructure operator like CambridgeNexus, navigating these regulatory hurdles and securing reliable grid access is paramount. The company's strategy relies on deploying high-density computing clusters in Tier III data center environments, which require meticulous responsible energy planning to remain sustainable and operational.

Securing the AI Factory for Sovereign Workloads

With AI infrastructure increasingly deemed essential to national competitiveness, providers must also guarantee military-grade security against espionage and cyber disruption. Bringing former National Security Agency Director and U.S. Cyber Command chief Michael Rogers onto the advisory board positions the Boston startup to target sovereign and defense-tier workloads.

During his 37-year Navy career, Rogers managed large, technically sophisticated organizations focused on intelligence and complex information systems. His presence on the advisory committee strongly suggests an intent to pursue high-security public sector contracts. Governments worldwide are accelerating investments in AI for intelligence analysis, cybersecurity operations, and critical infrastructure protection, requiring robust environments that can withstand sophisticated state-sponsored attacks.

“As AI becomes more central to business operations, reliability and security will be just as important as computing power,” Rogers said. “Organizations need confidence that the systems they depend on can perform consistently and withstand disruption. CambridgeNexus is focused on helping meet that need.”

The intersection of high-performance computing and intelligence-grade cyber defense is becoming a lucrative niche. While hyperscalers like Amazon Web Services and Microsoft Azure have established government cloud regions, specialized providers that can offer bare-metal, single-tenant deployments with extreme security guarantees are carving out their own space in federal procurement.

Beyond the GPU Crunch: The Rise of End-to-End Operators

While the initial wave of the AI boom focused on securing hardware, enterprises now struggle with the operational complexity of cooling, networking, and sustaining massive clusters. Deploying next-generation hardware, such as NVIDIA's GB300 NVL72 rack-scale systems, requires highly specialized environments capable of supporting up to 140 kilowatts per rack with advanced liquid cooling.

CambridgeNexus is betting on a turnkey AIFaaS model to solve this deployment friction for companies unwilling to build their own facilities. By offering bare-metal dedicated full-rack GPU infrastructure leasing, the company aims to bridge the gap between massive hyperscalers and specialized GPU cloud providers.

Hyperscalers offer vast ecosystems but are generally more expensive per GPU-hour and often charge significant egress fees. Conversely, specialized GPU cloud providers—such as CoreWeave, Lambda Labs, and Nebius—offer raw compute at significantly lower prices but often require customers to handle more of the software stack and infrastructure management.

Brad Hunstable, a technology entrepreneur with extensive experience in commercializing emerging technologies, was brought on to help navigate this competitive landscape. As the co-founder of Linear Labs and former CEO of Ustream, Hunstable understands the friction of scaling complex technological systems.

“For many businesses, the biggest obstacle to adopting AI isn't the technology itself. It's the cost and complexity of putting it to work,” Hunstable said. “CambridgeNexus offers a model that could make advanced computing more practical and accessible for companies that don't want to build the infrastructure themselves.”

The AI cloud market is fracturing. Specialized providers are undercutting legacy tech giants on raw compute, while the legacy giants maintain an advantage in integrated ecosystems. Operators that can offer the raw power of bare-metal deployments combined with the operational resilience of an integrated factory platform are aiming to capture the enterprise market.

“The next generation of AI infrastructure will depend on our ability to solve challenges that extend well beyond computing hardware,” Huang said. “Power availability, infrastructure resilience and the ability to scale efficiently will increasingly determine how quickly organizations can put AI to work.”