The FAA announced on June 22, 2026 that it had selected Air Space Intelligence in an award covering more than one system, with the centerpiece being SMART — Strategic Management of Airspace, Routes, and Trajectories. The agency describes SMART as “a cloud-based platform system that enhances existing FAA air traffic management systems.” The contract is valued at approximately $875 million over twelve years. September coverage revisits the programme ahead of what has been reported — though not confirmed with a specific date — as an imminent Washington, D.C.-area launch, with expansion to other regions to follow.
SMART uses AI to assess “airline schedules, weather, airport capacity, airspace conditions, and operational constraints to predict traffic flows and identify potential conflicts before they occur.” That output goes to a human air traffic controller who makes the decision. SMART assists controllers. It does not replace them. This distinction is not incidental — it is the entire design premise.
The programme launches first in the Washington, D.C. metropolitan area before expanding nationally. Reporting indicates airlines pushed for a limited initial scope; no airline is named in coverage and no motive is characterised here. The FAA is also managing a longstanding controller shortage, and separately announced a 2026 hiring plan intended to “erase the longstanding staffing shortage.” Those two facts share a context. They do not share a causal line — the system as described is decision support for controllers, not a staffing instrument.
Programme Timeline
June 22, 2026
FAA selects Air Space Intelligence; award covers more than one system, including SMART. Contract value: ~$875M over 12 years.
September 2026
Coverage revisits the programme ahead of the D.C.-area deployment. A near-term start has been reported; no specific date is confirmed.
Expansion Phase — timing unspecified
Following D.C.-area rollout, SMART is intended to expand to other regions. No schedule, milestone or sequencing is confirmed in available coverage.
The key insight: The twelve-year term is the single most informative fact in this announcement — not the dollar figure. It places this contract in a categorically different product class from standard frontier AI software, and it imposes structural obligations on both sides that a short-cycle deal never would.
Software that must behave the same way in year eleven as in year one is a different engineering problem from software that is expected to be superseded.
The Structural Read
Frontier AI software is built, priced, and sold on an implicit assumption: the underlying models will be substantially rewritten or replaced within a year or two. That assumption sits underneath how AI companies hire, how they raise capital, and what they promise at the contract stage. A twelve-year commitment inverts it entirely.
The vendor selected here must support a deployed system long after the models underneath it have been superseded by newer generations. The buyer — the FAA — must be able to certify behaviour that does not drift. In a safety-critical context, a capability improvement cannot be treated as automatically desirable: changed behaviour is behaviour that must be re-validated before it operates in live airspace. These are structural properties of long-dated safety-critical procurement. They are not claims about this vendor’s specific terms, obligations, milestones, or payment schedule — none of which are published.
Permission Layer
When the Buyer Controls the Clock
The Permission Layer framework describes how government procurement doesn’t just regulate which AI ships — it defines the temporal rules under which it must operate. A twelve-year horizon means the buyer, not the market, sets the upgrade cadence. The vendor’s product roadmap becomes subordinate to the certification cycle. That is a fundamentally different business than shipping a SaaS tool that auto-updates every sprint.
The second structural observation follows from the FAA’s own product description. The phrase “before they occur” places SMART in the prediction category, not the control category. The system assesses schedules, weather, capacity, conditions, and constraints — and produces a prediction about traffic flows and potential conflicts that it hands to a controller, who decides. SMART assists air traffic controllers. It does not replace them, and it does not make control decisions.
Held structurally, a predictive layer changes when a decision is made rather than who makes it. In congested systems, moving a decision upstream — before aircraft have departed — is generally where optionality is highest, because the available responses are cheaper and more numerous before anything has left the ground. That observation carries no performance claim and predicts no operational outcome. It is a description of where in a decision sequence a tool intervenes.
Finally, the staged rollout — D.C. first, then other regions — is the ordinary shape of safety-critical deployment, not a signal of doubt or confidence in either direction. Reporting that airlines pushed for a limited initial scope is noted as reported; no airline is named, no motive is characterised, and it is not read here as evidence of anything beyond standard practice.
Three Implications
IMPLICATION 1 — THE TWELVE-YEAR HORIZON IS A DIFFERENT PRODUCT CATEGORY
AI software built for annual replacement cycles competes on capability velocity. AI software built for twelve-year safety-critical deployments competes on behavioural stability, auditability, and the ability to certify a fixed envelope of behaviour over time. Those two markets require different engineering cultures, different sales motions, and different definitions of what “better” means. A capability jump that delights a consumer product manager may require a full re-validation cycle before it can be deployed in a safety-critical context — and re-validation takes time and resources that are not free.
IMPLICATION 2 — PREDICTION BEFORE DEPARTURE IS WHERE THE LEVERAGE SITS
The structural value of moving a decision earlier in a constrained system is that options remain open. A conflict flagged before departure can be resolved with schedule adjustments, routing changes, or ground delays that cost far less than resolving the same conflict once aircraft are airborne and airspace is committed. SMART is described as doing exactly that: surfacing potential conflicts before they occur, and giving that information to a controller who decides. The human decision role is explicit and central. No claim is made here about any improvement in delay, capacity, fuel, or safety — none is established in available coverage.
IMPLICATION 3 — GOVERNMENT PROCUREMENT SETS THE TERMS THAT THE MARKET CANNOT
The Permission Layer is visible here in an unusually direct form. The FAA controls not just which AI enters the national airspace system but at what pace, in what geographic scope, and under what certification constraints. Staged deployment — beginning in one metropolitan area — is standard safety-critical practice; it is not a commentary on the technology. But it does mean the buyer’s institutional process, not the vendor’s capability roadmap, determines the deployment timeline. For AI vendors, winning a twelve-year government safety contract is a different kind of win than closing enterpriseSaaS: the revenue is long-dated, but so is the obligation to operate within a certified behavioural envelope that the buyer, not the vendor, defines.
Business Engineer Framework
The Permission Layer
The SMART contract is a textbook Permission Layer case: government procurement doesn’t just decide which AI enters a regulated domain — it sets the temporal, behavioural, and geographic constraints under which that AI must operate for a decade or more. Understanding where the Permission Layer sits in the AI stack explains why long-dated safety-critical contracts represent a structurally distinct market, not a larger version of the same market. The Map of AI maps all nine layers, including where permission and certification sit relative to the model, the application, and the distribution layer.
The FAA’s selection of Air Space Intelligence and SMART is not primarily a story about an $875 million contract — it is a story about what the AI industry looks like when the Permission Layer is load-bearing, when the buyer’s certification cycle outranks the vendor’s capability roadmap, and when “before they occur” is the entire product thesis: a predictive layer that changes when a decision is made, hands the result to a controller who makes the call, and must hold that behavioural envelope steady for twelve years while the rest of the AI industry ships new models every quarter. Those two clocks do not run at the same speed, and the gap between them is the market this contract defines.
Disclosure: This article is businessanalysis only. It is not aviation or safety advice, and it is not investment advice. No view is expressed on any security and no recommendation is made.
SMART as described assists air traffic controllers with workflow and routing decisions. It does not replace them, and nothing above should be read as suggesting the automation of control decisions or the removal of any human role. No accuracy, performance, safety, delay, fuel or capacity metric is published in the coverage used here, and none appears above. No benefit is quantified. The controller shortage and the 2026 hiring plan are context. Nothing here claims the system is intended to substitute for controllers, to reduce headcount or to address the shortage, and nothing here speculates about employment effects in either direction. Any reported start date is treated as reported rather than confirmed. No claim is made about the contract’s terms, obligations, milestones, termination rights or payment schedule, none of which is published here, and no model or vendor technology detail, count of controllers, airports or facilities, competing bidder, or certification or approval status is stated. The report that airlines pushed for a limited initial scope is attributed as reported; no airline is named and no motive is characterised. A staged geographic rollout is ordinary practice for safety-critical deployment and is evidence of neither doubt nor confidence. Nothing here comments on aviation safety generally, references any incident or accident, or compares this to any other country’s air traffic system. No government or official’s motives are characterised and no position is taken on public procurement or policy. Nothing is predicted. No claim is made about whether Air Space Intelligence is publicly or privately held, or about any valuation, share price or market capitalisation. This is businessanalysis. It is not aviation advice, not safety advice and not investment advice; no view is expressed on any security and no recommendation is made.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
