Through the Lens
Navigating the Genesis Mission
Star Trek lovers - you will remember the movies, Star Trek II: The Wrath of Khan, Star Trek III: The Search for Spock, and Star Trek IV: The Voyage Home. And that the central theme for these three movies embodied the “Genesis Device”; a revolutionary but dangerous technology that creates a living planet from lifeless matter. This piece is not about that! Although I guarantee you policy peops in the Federal Government gave it some thought when considering what our government should be doing with Artificial Intelligence.
The “Genesis Mission” – launched via an Executive Order by President Trump – is not an academic or research science program. Nor is it entertainment. It’s a funding vehicle; established to allow Congress to create spending legislation and fund strategic public and private sector investment that will make America the predominant AI-capable sovereign nation. And to be economically and politically ahead of the rest of the world, most importantly China.
The importance of this policy move doesn’t rest on immediate federal appropriation, but on the ability to make future appropriations. And to anchor U.S. scientific discovery and national-security workloads inside a hybrid public–private compute ecosystem. Already, we can see that the agenda is supported by private sector capital in the order of hundreds of billions of dollars. This is real money. The developments are in the daily news cycle. We can’t avoid hearing about political and economic implications with AI.
The central analytical question is not, “How large is Genesis?” but rather, “How much incremental spending”, and infrastructure pull-through does Genesis create across the AI stack? This beyond the money that hyperscalers already are planning to spend – a level of investment forecasted to exceed $500 billion annually by 2026.
A Platform Mandate
The executive order establishing Genesis frames the initiative as a coordinated national effort to build an AI platform capable of training scientific models, deploying autonomous research agents, and automating discovery workflows; like evaluating DNA structure and modeling catastrophic weather systems. It also applies to the performance of defense assets in a myriad of offensive and defensive military scenarios.
Crucially, the executive order does not create a channel for new appropriations. Rather, it functions as a coordination instrument that re-prioritizes existing federal spending plans and signals long-term policy intentions. A good analogy would be, for example, moving water (money) from one (spending) bucket to another.
From an investment and infrastructure perspective, this distinction matters. Genesis is not a one-time spending event. It is a policy overlay that shapes capital allocation across the AI ecosystem. We need to become acquainted with discussing ‘ecosystem’ in the context of AI spending.
DOE as the Logical Anchor
Genesis will be implemented through the Department of Energy, fusing High-Performance Computing resources at DOE national laboratories with the private sector. Although competition with China is an unsaid strategic agenda, it is a logical look at the rest of the world already in a race to be the dominant benefactor of AI technologies. China would probably be considered the largest economic and political threat to the US with the use of AI. Russia, Iran, and fragments of the Middle East after that.
The Department of Energy is uniquely positioned to execute Genesis because of its stewardship of the national laboratory system, and its decades-long experience managing capital-intensive scientific infrastructure. DOE already operates the world’s most advanced, government-owned supercomputers and maintains institutional expertise with public and classified use of these critical resources.
Genesis builds upon existing federal R&D and security compute footprint rather than beginning a new spending conduit from point zero. The DOE FY2026 budget approved places total spending at approximately $46 billion, with the Office of Science accounting at roughly $7 billion of that. Within these buckets, Advanced Scientific Computing Research (ASCR) programs remain at approximately $1 billion in approved annual spending. The largest budget component, for the National Nuclear Security Administration (NNSA), exceeds $30 billion. Roughly two-thirds of FY2026 budget.
Scenario-Based Spending
Because Genesis lacks dedicated appropriations, its fiscal impact is best analyzed through scenarios. In a Re-Vectoring Scenario, Genesis primarily reorganizes existing federal spending, with incremental costs limited to software integration, staffing, and targeted procurements. Incremental federal outlays in this case would likely be in the low single-digit billions of dollars annually.
In a Platform-Build Scenario, Congress provides incremental funding to support Genesis as a national AI infrastructure priority. Annual incremental spending could rise to the mid-single-digit or low double-digit billions of dollars in scale, supporting new high performance computer systems. Networking capacity, storage, cybersecurity and natural resources like energy from the electrical grid, and water for cooling become a level of investment that can’t be ignored and will require some government backing.
In a National Mobilization Scenario, Genesis becomes a sustained, program over several decades with dedicated appropriations exceeding $10 billion per year. Even in this scenario, federal spending remains modest relative to private capital expenditures by hyperscalers (Google, Microsoft and AWS for example). But then, the private sector companies are entitled to make a profit from their CapEx. Federal spending is a necessary financial injection (or backstop if you will) so this buildout can happen rapidly; simultaneously promoting confidence that the projects will be completed. Failure to complete would be a huge black eye for everyone involved with the AI buildout not to mention the loss of investment dollars, the impact on corporate profits and the health of the economy in general. At this scale, once started, it will be difficult and painful to quit or turn back.
The Impact of Private Sector Capital
Lisa Su, CEO of Advanced Micro Devices, gave the keynote address at the 2026 Consumer Electronics Show (CES) in January. Her two-hour presentation included visiting corporate and government executives illustrating their plans for the use of AI compute once corporate spending is well underway and by the time the first stage of AI datacenter infrastructure is largely built.
Michael Kratsios, Director of the White House Office of Science and Technology Policy in the Trump Administration, joined the discussion and pointed to the already funded development of two new supercomputers at the Oak Ridge Laboratory in California; one code named ‘Lux’ (planned to be online in H1 this year), the other named ‘Discovery’. DOE already manages the world’s fastest supercomputer, El Capitan, at the Lawrence Livermore National Laboratory near San Francisco.
Given the proximity of these three ultra-high-powered machines to Silicon Valley, expect a meaningful and robust public-private partnership with the labs to build the ecosystem. Some of this work will be classified, other work will likely be ready for immediate industrial and academic use. Beyond the energy, defense and intelligence communities, the two largest private industries slated for research will include healthcare (biotechnology) and finance. Followed or coincident with fringe network devices, manufacturing systems and robotics.
Industry analyses project total capital expenditures by major cloud service providers to exceed $520 billion in 2026, with some estimates higher than $600 billion. Of this total, roughly 70 to 75 percent is expected to be directly related to AI infrastructure, including servers, HPC accelerators, the networking matrix, data storage, and data center construction.
The market forecast for servers reinforces this view. The global server market in terms of money spent for rack-built processors and networking gear is projected to rise from approximately $253 billion in 2024 to over $565 billion this year.
Building the Relevant Infrastructure
Genesis reinforces several durable investment themes. First, accelerated computing remains central. Scientific AI workloads are among the most compute-intensive applications with continuous demand for high-bandwidth CPU/GPU/TPU chips and memory systems. Lisa Su at AMD and Jason Huang at Nvidia regularly point this out in presentations. Dr. Su forecasts the worldwide compute demand in the next five years to exceed 10 Yottaflops1 (10 followed by 24 zeroes) based on what we know now – and the estimate is subject to change.
Second, Large Language Models rely on the movement and retention of bits and bytes, ones and zeros, for workflow orchestration, and instrument integration rather than raw compute alone. Imagine the amount of networking and storage needed as billions of individuals, government and corporate users grow their own storage space in the cloud using LLM.
Third, secure hybrid platforms become essential. Genesis incorporates the use of national security methods and sensitive datasets: private US government records regarding defense programs, government agency financing, and spending patterns for example. Cybersecurity solutions for AI will have a big role here.
Finally, power and facility infrastructure emerge as a second order but unavoidable constraint. The scale of AI infrastructure investment implied by current forecasts investment in power generation: the ‘grid’, meaning power plants, transmission facilities, and cooling systems, is capital intensive and in need of expansion. Already, the most advanced two-nanometer and three-nanometer process manufactured chips – which are actually trays of microprocessors and circuit boards built into rack systems in data centers – cannot operate without substantial closed loop water-cooling systems.
Through the Lens of New Edge Analytics
John Chambers, former CEO of Cisco, once famously said that Cisco could grow at a rate of 30 percent annually, indefinitely, and indeed in perpetuity, forever. Of course, he was terribly wrong, and I guess his economics staff on Cisco’s payroll (with stock options) wouldn’t tell him. Compounding at 30 percent, even with Cisco’s valuation at the time, would mean over the course of a decade or so, Cisco as a revenue-producing company would overtake US Nominal GDP. The Tech Wreck followed and nixed that idea without involving the economists.
People say sometimes about fabulous growth in the IT sector, that ‘this time it’s different’, the four most dangerous words when investing. But it might be a little different this time. Maybe we learned something from the Tech Wreck in the early 2000’s and the Financial Crisis of 2009 that ushered in the Great Recession. With the latter, Congress mandated all the Wall Street banks to be placed under the jurisdiction of the Federal Reserve, like traditional banks. And their ability to create junk bond financing and pools of subprime mortgages would be severely curtailed if not outright eliminated. Regulated in other words.
Genesis is a catalyst creating an ecosystem to motion public and private investment to converge around AI infrastructure as an industry. Genesis is American Capitalism. Genesis serves as a strategic guidepost for the broader AI economy. Personally, I would vote for the National Mobilization Scenario as the most likely format for public policy going forward, signaling a long-term commitment by the government, integration of data centers with the energy grid, and mining for fuels. And remember these are ‘reappropriations’, meaning it is allocation of existing capital created through existing federal legislation that can be directed elsewhere without new and likely partisan spending bills.
The Genesis Mission, implemented through the Department of Energy, amplifies several key infrastructure themes – enormous requirements for compute, provisions for data networking and storage, secure hybrid platforms and all of this while managing growing constraints due to the need for power and cooling systems. It’s well known that the electrical grid in America is aging rapidly and needs a facelift anyway. Taken together, Genesis emerges as a coordinated effort allowing profit in the private sector while backing the effort with federal spending for multiple beneficiaries. Recognize that AI is here to stay and Genesis is as good an idea as we have now to build it.
1FLOPS are ‘floating point operations per second’, the common measure of RISC processor computing performance in terms of speed. By the way of one acronym referencing another, RISC means ‘reduced instruction set computer’, a chip architecture made popular in the scientific, open-source community, compared with Intel x86 processors, an architecture so dominant with office automation and industrial use.
Copyright © 2026 New Edge Analytics. All rights reserved.
