By B. Thomas Edman, Strategy & Technology Lead, Xtremis.AI
In 1956, Eisenhower signed the Interstate Highway Act, authorizing 41,000 miles of road justified explicitly on national defense grounds. The highways did more than move tanks. They created the conditions for suburban development, national commerce, and the modern American middle class. Infrastructure designed for mobilization became infrastructure for prosperity.
Today, the electromagnetic spectrum occupies the same strategic position. Spectrum is the set of radio frequencies through which all wireless communication travels: every 5G call, GPS fix, radar sweep, drone command, and AI inference running at the edge. Think of it as a highway system for information, with different frequency bands serving as lanes of varying speed and capacity. And like the road system before Eisenhower, it is governed by administrative arrangements built for an era that no longer exists.
Figure 1. U.S. mid-band spectrum allocation and utilization, 1-6 GHz. Federal holdings show occupancy below 5 percent across most bands studied (PCAST, 2012). Commercial licensed bands reach approximately 40 percent utilization at peak hours and fall below 14 percent off-peak. The CBRS band (3.55-3.7 GHz) demonstrates a three-tier dynamic allocation model with over 500,000 deployed base stations. Bottom panel illustrates the highway analogy: of ten allocated lanes, only three carry traffic at any given time.
The near-term opportunity is concentrated in mid-band spectrum, roughly 1 to 6 GHz, which offers the best balance of coverage and capacity and is where most demand growth, federal underutilization, and policy action are focused. The United States has managed spectrum through auctions since 1994, generating over $233 billion in federal revenue. Those figures create an illusion of a functioning market, but they are not a market. They are one-time administrative sales with no secondary pricing, no demand-driven allocation, and no mechanism for dynamic access. Imagine if highway lanes were sold once and locked in for decades, with no congestion pricing and no way to lease unused capacity. That is how spectrum works today.
Figure 2. Commodity market infrastructure comparison across nationally significant resources. Every critical resource from crude oil to timber has developed spot markets, futures contracts, options, formal exchanges, and price indices. Spectrum, despite underpinning trillions in downstream economic activity, has none. FCC auctions have generated $233 billion since 1994, but these are one-time administrative sales with no secondary trading mechanism.
The consequences are stark. Government mid-band holdings see utilization of only 20 to 35 percent. Commercial bands sit below 40 percent at peak hours and under 15 percent off-peak. This is not a demand problem. It is a structural failure. The FAA has authorized over 425,000 commercial drone pilots. Private 5G networks are projected to grow from $4.7 billion to over $24 billion by 2030. Autonomous vehicles, edge AI, and augmented reality each represent fundamentally new categories of spectrum consumer. The spectrum they need is technically available but institutionally locked. The cost falls hardest on those farthest from Washington: the rural hospital that cannot deploy remote monitoring, the farmer who cannot run autonomous equipment where allocations sit idle.
We have seen this pattern before. In the 1970s, government price controls turned a manageable energy supply into a national crisis: five-mile gas lines, odd-even rationing, shuttered factories. The root cause was not physical scarcity. It was that administered markets prevented supply from meeting demand. The breakthrough came through a coalition of regulators and industry. FERC Order 636 in 1992 mandated unbundling. NYMEX Henry Hub futures became the pricing benchmark. The result transformed a $50 billion industry into one of the world’s most liquid commodity markets. Critically, the supply-side effects were as significant as the demand-side gains. Long-dated futures gave producers the ability to hedge prices years forward, which de-risked exploration and production investment, which unlocked supply. The spectrum equivalent: if an operator can sell or lease access on a forward contract at a known price, it de-risks investment in small cells, CBRS base stations, private network buildouts, and sensing infrastructure. Today, nobody builds speculative spectrum access infrastructure because there is no way to monetize unused capacity. A market changes that calculus entirely.
Industry estimates suggest that every 100 MHz of mid-band spectrum released for commercial use contributes roughly $260 billion in GDP over 15 years. Even conservative estimates at half that figure imply tens of billions in annual output from bands that currently sit idle. The goal is not to drive every band to 100 percent utilization simultaneously. That would create interference and backhaul bottlenecks. The goal is dynamic allocation: filling idle windows, matching supply to demand in real time, and raising effective throughput without requiring new physical infrastructure. The gas pipelines did not get bigger after deregulation. They just moved more gas.
The technology to enable this already exists. The Citizens Broadband Radio Service has deployed over 500,000 base stations using automated, interference-managed dynamic allocation. Ukraine’s electronic warfare environment has accelerated real-time spectrum sensing and AI-driven signal classification. The Golden Dome initiative, with $24.4 billion appropriated and MDA’s SHIELD contract at up to $151 billion, requires the same capabilities: real-time sensing, dynamic frequency allocation, automated coordination at machine speed.
The parallel to the internet is direct. ARPANET was built on defense infrastructure, packet-switching networks designed for military resilience, that became the substrate for commercial innovation. The same pattern is unfolding in spectrum. Defense companies building electromagnetic spectrum operations capability for contested environments are creating the sensing, orchestration, and dynamic allocation systems that a commercial spectrum exchange requires. The defense industrial base is building the internet of spectrum.
Domestically, unauthorized drones threaten power grids, prisons, and stadiums. Federal C-UAS spending is projected to exceed $4 billion in 2026. Every counter-drone system depends on spectrum sensing, generating infrastructure that both requires and enables dynamic access.
Policy momentum is accelerating. The One Big Beautiful Bill Act restored FCC auction authority through 2034, mandating at least 800 MHz for auction. The Government Spectrum Valuation Act would require periodic valuation of federal holdings. Congressional appetite is bipartisan.
What remains is market architecture: platforms for liquidity, standardized contracts defining transmission rights, pricing benchmarks, and clearing infrastructure. This is not speculative R&D. It is market infrastructure investment with defense procurement pulling in the same direction. The country that builds the first functioning spectrum exchange will set the rules for every nation that follows.
This is the kind of cross-sector challenge that demands what CGCI calls “GCI investing“: market-led strategies channeling capital toward assets of strategic importance to national security and economic resilience. The vision is audacious. But so was the Interstate Highway Act, and so was energy deregulation. Both delivered their greatest benefits not to the institutions that built them but to the millions of Americans who used them. Spectrum is next. The question is not whether it becomes a traded commodity. It is who builds the market.
