About $5.98M for the Army NAUTILUS broad agency announcement — a form of procurement in which the government does not specify what to build (ECS Federal)

1Procurement without a specification

Ordinary government procurement decides what is to be delivered, how many and to what performance, and then selects whoever can meet it. A broad agency announcement runs close to the reverse. The government indicates the research areas it is interested in, and what to build and how is brought by whoever proposes. At the research stage the government does not know the answer either.

What is not known cannot be written into a specification, so a form exists for collecting proposals and choosing among them.

2Competition by proposal rather than price

Under this form, competition does not turn on lowest price. Selection depends on what is proposed and whether it looks technically promising. Put the other way, because the criteria are not quantitative, the evaluation process carries weight.

Among the records this site holds, five touch a broad agency announcement, 21 mention SBIR and 25 mention Phase II, so research-stage funding is distributed through several distinct mechanisms, each shaping who brings what.

3The limits of what a description reveals

The description on this record is a single line. Beyond the name of the announcement and that it concerns artificial intelligence research and development, nothing is written. What NAUTILUS studies, and what outcome is sought, cannot be determined from this record. Short descriptions of this kind are not uncommon among the federal spending records this site holds.

Research contracts, unlike procurement of equipment, do not lend themselves to a one-line statement of what was bought, and the granularity of records does not line up. Amount, period and recipient are visible; the substance is not.

4Which programmes carry research-stage funding

At the research stage the government does not hold the answer either. What is not known cannot be written into a specification, so the system provides several vessels instead.

Dividing the 222 federal spending records this site holds as of 2026-08-31 by the wording of their summaries produces this distribution. Each programme differs in who brings what. Under a BAA, competition is not on lowest price but on what is proposed and whether it is technically promising — which, conversely, makes the evaluation process important, since the criteria are not quantitative.

The summary of this record is a single line, and what NAUTILUS researches cannot be determined.