Sentinel’s $140.9 Billion Cost Is Not Northrop Grumman’s Contract

Sentinel’s large figures come from different ledgers. The Office of Cost Assessment and Program Evaluation (CAPE) estimates the government’s acquisition program. Contract systems record awards and obligations. Annual budgets propose authority. Northrop reports backlog and sales. One ledger cannot substitute for another.

A Sentinel figure is incomplete unless it identifies five things: whose ledger, what scope, what legal status, what dollar basis, and what date?

The answer in three numbers

The public debate most often mixes three figures:

• $140.9 billion is CAPE’s estimate for acquiring a reasonably modified Sentinel program after the 2024 Nunn-McCurdy review. It is expressed in constant fiscal year 2020 dollars. [2] [3]
• $13.720 billion is the cumulative obligation shown on the public award profile for the prime engineering and manufacturing development contract through the latest action dated April 29, 2026. At this cutoff, USAspending also reports the same amount as the current and potential award amounts. [9]
• $7.6 billion is the increase Northrop disclosed in Sentinel backlog during the second quarter of 2026 after further program work was defined and authorized. [13]

They should not be combined or expected to add to one total. The first is program-wide and government-estimated. The second is a legal commitment recorded against one contract. The third is a contractor accounting and planning measure for expected future sales.

One program, several ledgers

Figure	Date and basis	What it is	What it is not
About $78B	2020 baseline, constant FY2020 dollars	Original acquisition-program estimate summarized by CBO	Prime award or appropriation
$13.294B	Sept. 8, 2020	Air Force engineering and manufacturing development contract award value	Cash paid on award
$85M	Sept. 8, 2020	Research and development funding obligated at award	Total contract funding
$140.9B	July 2024, constant FY2020 dollars	Estimate for the reasonably modified acquisition program	Approved current baseline or Northrop contract
About $170.6B	GAO presentation in FY2025 dollars	Same modified-program estimate presented in FY2025 dollars	A second estimate to add to $140.9B
$13.720B	Latest public action Apr. 29, 2026	Cumulative obligations and current and potential award amounts	Outlay, revenue, profit, or program-wide spending
$7.6B	Q2 2026 company disclosure	Increase in Sentinel backlog after work authorization	Payment, revenue, or a separately priced production award
$176M	First half of 2026 versus prior period	Increase in Sentinel sales	Total Sentinel sales

The $140.9 billion and approximately $170.6 billion figures are especially easy to double count. They are different-dollar-year presentations of the same modified-program estimate. They are not two costs and must not be added.

What the $140.9 billion estimate covers

After the 2024 critical Nunn-McCurdy review, the under secretary of defense for acquisition and sustainment certified that Sentinel met the statutory criteria to continue. The certification rescinded Milestone B and directed the Air Force to restructure the program. The department reported a $140.9 billion acquisition-cost estimate for a reasonably modified program, an 81 percent increase from the program’s September 2020 estimate. CBO identifies both figures as constant fiscal year 2020 dollars.

[2] [3]

“Acquisition cost” is broader than the prime contractor’s engineering work. CBO says the review identified cost increases in research, development, test, and evaluation; procurement; and military construction. The estimate covers the missile and extensive supporting infrastructure. It excludes operations and sustainment after fielding and some separate ICBM-modernization programs, including work on a new reentry vehicle and fuze. [3]

The estimate is also not a check that Congress wrote in 2024. It spreads anticipated acquisition cost across years. Annual authority still moves through budget requests, legislation, apportionment, contract actions, obligations, and outlays.

GAO’s 2025 annual assessment presented the reasonably modified-program estimate as about $170.6 billion in fiscal year 2025 dollars. Among 30 major programs included in both its 2024 and 2025 assessments, GAO reported a $49.3 billion combined estimate increase; Sentinel accounted for more than $36 billion.

[4] The different dollar basis does not create a second pot of money.

What Northrop was actually awarded

On September 8, 2020, the Air Force announced a cost-plus-incentive-fee definitive contract with an initial value of $13,293,562,839 for engineering and manufacturing development. The same notice said $85 million in research and development funds was obligated at award. [1]

The government therefore awarded a contract with an initial stated value of about $13.294 billion and initially obligated $85 million against it. Those are distinct actions.

The live USAspending description also references early production and deployment. Northrop’s June 2026 filing nevertheless describes the $13.3 billion award as the EMD contract and says the broader Production and Deployment phases remain to be priced and negotiated. Read together, the records show that the existing award references early production and deployment scope; they do not establish that the broader phases are finally priced. [9] [13]

It does not produce an accurate sentence saying Northrop received $13.294 billion in cash that day. Nor is this public award structured as an IDIQ vehicle with task-order competitions. USAspending classifies FA8219-20-C-0006 as a definitive contract without a parent indefinite-delivery vehicle. Later public actions are funding actions and modifications to that contract, not task orders against a ceiling. [9]

How $85 million became $13.720 billion

The cumulative obligation grew through later contract actions. The public award record reported $13,720,293,923.19 in cumulative obligations through its latest listed action on April 29, 2026. It also reported that amount as both the current and potential award amounts. [9] That equality is a dated feature of the record, not a general rule. Later modifications, obligations, or deobligations can change the figures.

It is also not a program-wide spending total. The acquisition estimate includes government program work, infrastructure, and activity outside this one prime contract. And an obligation is not necessarily an outlay, cumulative Northrop revenue, or profit. Those measures respond to different events and can be reported on different schedules.

This is why describing $13.720 billion as “spent on Northrop” can be misleading. The public record establishes committed budget authority on the contract. A complete cash and income statement would require outlay, billing, revenue-recognition, and cost information that the public program record does not provide at the same granularity.

Annual budget requests are another layer

The fiscal year 2027 President’s Budget requested $4,491.775 million for Sentinel research and development, $107.602 million in net missile procurement after accounting for $3.220 million of prior-year advance procurement, and about $1.3 billion for Sentinel military construction. [11] [12] [10] These amounts sit in separate appropriation titles. They are not an official program total, enacted funding, or new contract awards.

The sequence matters:

1. The executive branch requests budget authority.
2. Congress authorizes policy and appropriates budget authority.
3. The government apportions and allots available funds.
4. An authorized official obligates funds through a contract or other action.
5. Treasury records an outlay when payment occurs.

Even after enactment, not every Sentinel appropriation dollar necessarily flows through the Northrop prime contract. Military construction, government activities, other contractors, and separate program elements can sit elsewhere in the acquisition estimate.

Why a $7.6 billion backlog increase is not a check

Northrop’s second-quarter 2026 filing said further definition and authorization of program-plan elements increased Sentinel backlog by $7.6 billion. The same filing continued to say that production and deployment scope remained to be priced and negotiated.

[13]

Northrop defines backlog as future sales expected from firm orders, including funded and unfunded work. It converts into sales as the company incurs costs or makes deliveries under the applicable accounting method. The 2025 annual report likewise described Sentinel as remaining in the engineering and manufacturing development phase and said production and deployment were not yet finally priced.

[6]

Northrop included the backlog increase among its “significant second-quarter new awards,” but that company presentation does not establish a separately priced production contract in the federal award record. It is not revenue. The same quarterly filing reported that first-half Sentinel sales increased by $176 million from the prior-year period. That is a period-over-period change in sales, not cumulative program revenue and not the $7.6 billion backlog amount.

Revenue can precede delivery

Northrop says it generally recognizes revenue from long-term contracts over time using a cost-to-cost measure of progress. Revenue can therefore appear before a complete system is delivered or fielded, and sales alone do not establish physical completion.

[6]

Backlog is not revenue. It represents expected future sales under the company’s definition. Billings and cash collections do not necessarily equal revenue. Margin and profit require cost information. No cited public filing provides a defensible cumulative figure for Sentinel revenue or profit.

Test, delivery, and fielding are separate claims

The Air Force said in February 2026 that restructuring was expected to finish during 2026 and that initial capability was targeted for the early 2030s. [7] GAO reported one day later that the previous Milestone B approval and related baselines had been rescinded, with a new baseline still to be established. It also described the first flight as planned for March 2028. [8]

A component test or flight is not government acceptance of a completed system. Acceptance is not operational fielding. Initial capability is not full deployment. Each claim needs its own dated record.

GAO’s September 2025 public transition report is deliberately limited because it omits information the government identified as sensitive or classified. It still establishes high-level planning and risk-management gaps without supplying operational detail. [5] That is the appropriate boundary for business analysis: cost, schedule, acquisition structure, and public accountability, not sensitive capabilities or operating plans.

A checklist for the next Sentinel number

Before repeating a new figure, ask:

1. Whose ledger is it? CAPE, Air Force, Congress, USAspending, or Northrop?
2. What is the scope? Whole acquisition program, one appropriation, one contract, or one company?
3. What dollar year applies? Constant FY2020, FY2025, or nominal dollars?
4. What is the legal status? Request, appropriation, award, obligation, or outlay?
5. What is the accounting status? Backlog, performance obligation, sales, cash, or profit?
6. What event occurred? Authorization, test, delivery, acceptance, initial capability, or fielding?
7. What is the cutoff date? Every obligation total and program baseline can change.

The framework does not make the cost growth smaller. It makes the accountability sharper. The $140.9 billion estimate applies to CAPE’s reasonably modified government acquisition program. The $13.720 billion cumulative-obligation total belongs to a dated contract record. The $7.6 billion belongs to Northrop’s future-sales ledger. Accurate reporting starts by keeping them apart.