All essays
§031 · Positioning

Federal facilities contracting is pulling away from software

This week's obligation data shows eight major contract actions — seven of them anchor physical infrastructure, legacy industrial systems, or national-lab M&O agreements. Only one touches modern IT services, and it's legacy aviation infrastructure modernisation, not software.

3 min · Published 2026-11-09 · By Bridger

The federal government obligated funds this week for rocket engine production restarts, strategic petroleum reserve management, border barrier construction, and the continued operation of Oak Ridge National Laboratory. It did not obligate funds for cloud architecture, software modernisation, or digital services — at least not at a scale that registers in the public appropriations record. The gap is structural, not anecdotal.

Of the eight enacted obligations tracked this week, six belong to the machinery of physical federal presence: energy stockpiles, national laboratories, border infrastructure, launch systems. The seventh is a State Department incremental funding action to an incumbent systems integrator. The eighth — FAA's T5 contract to support Aviation Capital Improvement Programme implementation — is the only signal that touches anything resembling modern IT, and even there the scope description foregrounds "implementation" of capital infrastructure, not software development or platform engineering.

The industrial base still owns the appropriations calendar

This is not a temporary distortion. It reflects the durable reality that federal spending authority remains concentrated in programmes where incumbency, facility security clearances, and multi-decade contract vehicles create insurmountable entry costs for firms outside the traditional defence and energy industrial base. UT-Battelle has operated Oak Ridge since 2000. Sandia has been managed under continuous contract since 1949. The Strategic Petroleum Reserve operator was selected under a long-cycle procurement with exhaustive operational-security requirements. These are not markets; they are franchises.

Federal software procurement happens, but it happens under the noise floor of the appropriations process — in task orders against existing IDIQs, in micro-purchases, in agency working capital funds that never surface as line items.

Federal software procurement happens, but it happens under the noise floor of the appropriations process — in task orders against existing IDIQs, in micro-purchases, in agency working capital funds that never surface as line items. When allied tech firms ask where the "digital transformation" budget is, the answer is: it exists, but it is atomised across hundreds of small actions that do not generate the kind of obligation signal that appears in public reporting. The large, visible contract actions remain anchored to steel, concrete, uranium, and jet fuel.

What this means for positioning

Allied tech companies entering the US federal market tend to track RFP calendars and assume that software modernisation programmes will eventually mature into large, visible contract vehicles. The evidence suggests otherwise. The programmes that generate eight-figure public obligations are infrastructure programmes, and the firms that win them are infrastructure firms. Software work is packaged differently: smaller, faster, more modular, and nearly invisible in the appropriations record until it reaches operational scale — by which point the incumbent is entrenched.

The implication is not that software opportunities do not exist. It is that they do not look like the opportunities that dominate the public federal-contracting record. Positioning for them requires tracking agency IT roadmaps, identifying uncommitted modernisation funds before they are packaged into task orders, and building relationships with programme offices before the procurement is scoped. The window for entry is earlier, smaller, and less visible than the industrial-base model — and it closes faster.

This week's obligation data is a useful reminder: the federal market is not one market. It is at least two, operating on different calendars, with different incumbents, and different rules of entry. The industrial base still owns the appropriations calendar. Software firms need to stop waiting for it to show up there.

Built on this thesis

See how it shows up in the product.

Request access