DARPA Firefox Seeks Scalable Technologies for High-Resolution, Noninvasive Brain-Computer Interface

The Defense Advanced Research Projects Agency’s (DARPA) Biological Technologies Office (BTO) has released the Firefox Program Solicitation, DARPA-PS-26-137, to advance a scalable, noninvasive brain-computer interface (BCI). The program seeks to build on DARPA’s existing coherent optical sensing system by addressing two scale up barriers: steering coherent light across multiple neural sensing regions and processing the resulting optical speckle data before digitization. Firefox’s ultimate objective is a 100-channel, noninvasive BCI capable of 50-micron spatial resolution and 100 microsecond temporal resolution.

Firefox Program/Solicitation Overview

Firefox is a 36-month program divided into three 12-month phases. TA1 and TA2 teams will work independently to develop their respective component technologies. DARPA’s designated Independent Verification and Validation (IV&V) partner is expected to test and integrate performer outputs into the government-retained coherent optical system, with the resulting system benchmarked against fMRI in human subjects.

The selection process begins with a required abstract submission. DARPA will review conforming abstracts and may invite selected teams to submit an Oral Presentation Proposal (OPP) package and participate in an oral presentation. Favorably reviewed OPPs may lead to negotiation and award of a Phase 1 OT for Prototype agreement, with Phases 2 and 3 structured as option periods.

Technical Objectives/Capability Areas

Firefox is structured around two distinct Technical Areas (TAs). Offerors may submit to either TA; teams pursuing both must submit separate proposals, each addressing only one TA.

Technical Area 1: Dynamic Optical Rastering and Steering

Technical Area 1 seeks prototype components or systems of components that dynamically steer or raster coherent light across an operationally relevant sensing volume. The goal is to expand DARPA’s current single-point sensing capability to a rastered field of view supporting multiple independent neural sensing targets.

TA1 proposers must develop approaches compatible with the government-retained coherent optical system, including:

• Maintaining coherence of incoming light centered at 1310 nm.
• Supporting a 3-10 degree bi-static angle between illuminator and receiver fields of view.
• Accommodating the specified optical system parameters, including a single mode transmit fiber and multimode receive fiber.
• Enabling optical transmit power of at least 10 mW average power in Phase 1 and 100 mW average power in Phase 2, while complying with ANSI skin/tissue laser-safety limits.
• Supporting miniaturization toward a fiber tethered, head-mountable configuration.
• Demonstrating performance through benchtop testing and phase gated validation activities.

DARPA identifies spatial light modulators, MEMS, Risley-prism scanning assemblies, and optical phased arrays as example approaches, while remaining open to other technically responsive architectures.

TA1 Performance Targets

• Phase 1: Minimum 0.5 steradian field of view, fewer than 100 ms for a full volume scan, and 20 independent coordinate points.
• Phase 2: More than 0.5 steradians, fewer than 10 ms for a full volume scan, 100 mW average optical transmit power, and a tethered optical head form factor.
• Phase 3: More than 1.0 steradian field of view, fewer than 100 microseconds for a full volume scan, and 100 independent coordinate points.

Technical Area 2: Analog Optical-Speckle Processing

Technical Area 2 seeks a scalable, high bandwidth, low power prototype architecture for optical-speckle processing. DARPA’s current single point coherent optical system generates approximately 24 Mb/s of raw data per channel. Firefox’s TA2 objective is to process, combine, and compress the resulting optical streams in the analog or optical domain before digital conversion, reducing the downstream data processing burden for a multi-channel interface.

TA2 performers must execute data compression, phase alignment, and speckle processing entirely in the analog or optical domain before digitization. DARPA states that proposals primarily relying on digital signal processing or digitization for the required processing will be considered nonresponsive. TA2 teams must also address miniaturization, optical collection efficiency, and processing response times faster than the stated biological-tissue speckle decorrelation time of less than one millisecond.

TA2 Performance Targets

• Phase 1: Process 16 individual speckles per sample per channel, achieve 24-bit dynamic range, align 16 speckles within 10 ms, and exceed 10% optical collection efficiency.
• Phase 2: Process 32 speckles per sample per channel, align 32 speckles within 100 microseconds, and exceed 20% optical collection efficiency.
• Phase 3: Process 100 individual speckles, align 100 speckles within 1 microsecond, and exceed 50% optical collection efficiency.

Key Dates & Submission Timeline

• Proposers Day: September 21, 2026.
• Questions due: October 13, 2026.
• Abstract due: October 19, 2026, by 3:00 PM ET.
• Abstract submission method: DARPA Broad Agency Announcement Tool (BAAT).

Funding Overview

DARPA anticipates making multiple Firefox awards, with total funding estimated between $4 million and $7 million across Phases 1-3. Awards may be made as OT for Prototype agreements, and the Government may make no awards.

The solicitation encourages participation from large and small businesses, nontraditional defense contractors, universities, and research institutions. Cost sharing is not expected unless a traditional defense contractor is not working to a significant extent with a nontraditional defense contractor.

The solicitation also contains significant intellectual property and data rights considerations. DARPA expects to retain Government Purpose Rights in program subject data and asks proposers to address a licensing structure through which the Government could obtain Unlimited Rights for an entire TA technology, including background and foreground IP, when the performer does not intend to commercialize the technology. Teams should review the model OT agreement and evaluate the requested rights structure before committing to a proposal strategy.

Positioning for Firefox

Firefox is a technically demanding opportunity for teams with credible capabilities in coherent optics, photonics, beam steering, optical hardware, high-speed analog processing, neuromorphic photonics, in-sensor computing, and prototype system integration. The strongest submissions will connect a specific architecture to DARPA’s phase-by-phase metrics, show how the solution will interface with the required 1310 nm coherent optical system, define meaningful technical risks and mitigations, and establish a practical demonstration and IV&V handoff path.

For TA2, the clearest gate is straightforward: a solution that sends raw or high-bandwidth optical signals into the digital domain for primary processing does not meet Firefox’s stated objective. Teams should ensure that their technical narrative, architecture diagram, test plan, and budget all substantiate genuinely pre-digitization analog or optical phase-processing and data-compression capability.

Firefox presents an opportunity for government contractors and research performers to contribute enabling components for DARPA’s pursuit of a scalable noninvasive BCI. Organizations considering a response should begin immediately with technical-area selection, teaming alignment, IP and data-rights review, CMMC and registration readiness, and development of a concise, metrics-traceable seven-page abstract.