Two firms starting the same programme on the same day - SABER, securing the robustness of battlefield AI (BAE Systems, about $1.65 million)
A DARPA contract to BAE Systems obligating about $1,647,757, under a programme securing the robustness of AI used on the battlefield. Another firm is running the same programme from the same start date.
Contract key facts
- RecipientBAE SYSTEMS INFORMATION & ELECTRONIC SYSTEMS INTEGRATION INC
- Contract value$1,647,757 (≈$1.6M)
- Awarding agencyDepartment of Defense
- Awarding sub-agencyDefense Advanced Research Projects Agency
- Award typeDEFINITIVE CONTRACT
- Period of performance2026-02-01 〜 2028-01-31
- Contract ID (PIID)HR001126CE030
Contract scope (original)
SECURING ARTIFICIAL INTELLIGENCE FOR BATTLEFIELD EFFECTIVE ROBUSTNESS
Key points
- Robustness covers both withstanding accidental input and noise and withstanding interference an adversary sets deliberately.
- Of the 222 records this site holds as of 2026-09-02, two touch this programme, both starting 1 February 2026.
- The other is about $2,313,263 to Lockheed Martin, a factor of 1.4 above this contract at about $1.65 million.
- Whether the difference follows from scope or from the scale of the proposal is not stated in the record.
- DARPA running several teams on one question also appears in SAIL-ON (seven), AIR (four) and PTG (three).
- Resistance to adversarial attack is shared with civilian AI in autonomous driving and critical infrastructure control.
1What robustness means here
Robustness here covers both holding performance under unexpected input and noise, and withstanding interference an adversary sets deliberately. The first arises by chance; the second is worked at by someone probing for the weakness. In a battlefield environment, the second is the premise.
2The same programme run by two firms from the same day
| Recipient | Amount and period |
|---|---|
| BAE Systems (this contract) | about $1,647,757, 1 Feb 2026 - 31 Jan 2028 |
| Lockheed Martin | about $2,313,263, starting the same 1 February 2026 |
Of the records this site holds as of 2026-09-02, two touch this programme and both begin on the same date. DARPA setting several teams on one research question at once appears also in SAIL-ON (seven firms), AIR (four) and PTG (three), covered elsewhere on this site. At the research stage no one knows in advance which line will succeed, so running them in parallel is itself the method.
3The difference in amount need not mean anything
The two amounts, about $1.65 million and about $2.31 million, differ by a factor of 1.4. The record does not say whether that follows from a difference in scope or in the scale of what was proposed. What can be read is that two firms took up the same programme from the same day, each with a two-year period.
- 1Run several teams at onceContracts of a few million dollars are the entrance
- 2Judge the resultsWhich line succeeds is sieved here
- 3Grow, or endSuccess grows into a large programme; failure ends quietly
Placing AI in the harsh, unpredictable environment of a battlefield surfaces weaknesses peacetime testing did not show. If an AI carrying a judgement malfunctions under attack or disruption, the consequence is serious. Resistance to adversarial attack is a problem shared with civilian AI in autonomous driving and critical infrastructure control.
Why it matters
Resistance to adversarial attack becomes a condition of acceptance as AI enters operation. Questions raised first in defense research descend in the same shape to autonomous driving, finance and critical infrastructure control. Procurement that runs several teams at once keeps a plan alive even where part of it misses, which from the supply side means a small entry contract can still lead somewhere large.
FAQ
What is SABER?
Are there other contracts under the same programme?
Why award two firms at once?
Sources (primary)
This article is an independent organization based on the U.S. official spending data below. Verify the exact, latest details with the official source.
- USAspending (award details)
- Contract ID (PIID)：HR001126CE030