On November 3, Raytheon, an RTX business, announced it has been awarded a phase-one contract from the Defense Advanced Research Projects Agency (DARPA) to develop an all-optical processing system under the Photonic Integrated Circuit Architectures for Scalable System Objectives, or PICASSO, program. Raytheon’s Advanced Technology team will design and demonstrate photonic circuit technology that processes signals with light rather than electricity. Synopsys, the electronic design automation company, will provide design and simulation tools to support the work.

That is, in substance, the whole announcement — and the restraint in it is the story. No contract value was disclosed. No architecture was described. No performance target, delivery schedule, or platform was named. What DARPA has bought is a design and demonstration effort inside an 18-month first phase of a program that exists precisely because the thing Raytheon is being asked to build does not yet work at scale. This is the start of a research program, and the interesting questions are still ahead of it.

What was actually announced

Precision first. The award is a phase-one contract under PICASSO, a DARPA program solicited in January 2026. Raytheon’s Advanced Technology team will “design and demonstrate new photonic circuit technology that uses light rather than electricity to process signals,” in the release’s words. The stated purpose is operational: operators can “detect threats sooner, respond more quickly and maintain better situational awareness in contested environments,” because processing that stays in the optical domain avoids the conversion step that today’s photonic circuits require.

The release gives the program one executive voice. “This program recognizes a fundamental shift in how processing can be achieved,” said Colin Whelan, president of Advanced Technology at Raytheon. “By moving critical processing entirely into the optical domain, we can enable significantly faster response times for a broad range of applications.” That is a claim about direction, not a result: no response-time figure or benchmark accompanies it, and none should be read into it.

The commercial partner is named but the tooling is not specified. Synopsys will provide “design and simulation tools to support development.” The release does not say which products, and this article does not assume: Synopsys’ photonic portfolio includes the OptoCompiler design platform and the OptSim circuit simulator, but whether either is involved in the Raytheon work is not on the record.

Raytheon’s stated qualifications are its own framing: “deep experience in photonics process development, radio frequency design knowledge and fielded military solutions.” The release names no prior photonic program, foundry partner, or fabrication process — the manufacturing question is entirely undisclosed, which matters for a program whose solicitation says it strongly prefers U.S.-based manufacturing.

The conversion problem PICASSO exists to solve

Today’s photonic circuits move information as light, which is fast and wideband. But when that information needs processing — filtering, combining, transforming, deciding — the circuit almost always converts the optical signal back into an electrical one, does the work in electronics, and converts the result back to light. Each optical-electrical-optical conversion costs latency and power, and each is a place where signal strength and integrity are lost.

The all-optical alternative is to keep the signal in the optical domain through the processing itself — to filter, switch, and compute with light acting on light. In principle that removes the conversion penalty entirely. In practice it runs into two problems that have kept all-optical processing a laboratory pursuit for decades, and DARPA’s solicitation names both as the program’s technical challenges.

The first is signal integrity. Unlike digital electronic circuits, which regenerate and filter noise at every stage, analog optical signals in long processing chains attenuate and accumulate noise that cannot simply be restored with optical amplification, because amplification adds its own amplified spontaneous emission noise. Every stage degrades the signal a little, and the degradation compounds: a circuit that works at four stages may be unusable at forty — and the systems PICASSO imagines need far more than forty.

The second is predictability. Optical systems suffer spurious wave interactions — scattering, back-reflections, leakage, crosstalk between modes — that have no direct analogue in electronics and grow unmanageable as circuits get larger. A designer cannot simply scale up a working small circuit and expect the large one to behave the same way. The solicitation calls this the mode-control problem, and it is why PICASSO is framed as an architectures program rather than a devices program: the challenge is not a better modulator or waveguide, but circuit-level strategies that keep a large photonic system predictable.

PICASSO is therefore not asking performers to invent a new component. It is asking them to invent ways of composing components into systems that stay clean and predictable at scale — a design problem as much as a physics problem, which is also why a design-tools partner like Synopsys is part of the announcement rather than an afterthought.

What PICASSO is, as a program

PICASSO was solicited on January 14, 2026, as DARPA-PS-26-13, out of the agency’s Microsystems Technology Office, with proposals due March 6, 2026 and an estimated period of performance starting July 1, 2026. The program is managed by Anna Tauke-Pedretti of DARPA’s Multi X Office, according to trade coverage. The solicitation put anticipated funding at $35 million; Washington Technology reported DARPA planned several awards worth up to $35 million each. The two figures are not obviously the same number, and neither Raytheon nor DARPA has said what Raytheon’s contract is worth — so this article asserts no dollar figure for the award.

The program has two technical areas. Technical Area 1, the one Raytheon is working under, runs in two 18-month phases with a DARPA down-select between them. Phase 1 is the circuit-integrity phase: develop circuit-level strategies to preserve optical signal integrity and suppress and decouple parasitic interactions. Phase 2 generalizes those approaches to distinct circuit functionality — building working circuits from the integrity techniques proven in phase one. Technical Area 2, a single phase applying TA1 designs to application-specific improvements, was not solicited under this program solicitation; DARPA said it would hold an industry day before a separate TA2 solicitation.

The target applications DARPA lists are computing, analog signal processing, and sensing. The breadth is deliberate — a circuit architecture that preserves optical signal integrity at scale would serve all three — and a hedge: at phase one, nobody knows which application benefits first, and the down-select exists so DARPA can concentrate funding on the approaches that prove out.

The timeline matters for expectations. With an estimated start of July 1, 2026 and an 18-month first phase, Raytheon’s phase-one work would run into early 2028 before the down-select. Nothing Raytheon announces before then — no demonstration, no paper, no prototype — is the program’s outcome. It is a milestone inside a phase whose purpose is to establish whether the approach works at all. As of publication, no other phase-one performer has publicly announced an award, and DARPA has not published an awardee list — so Raytheon is the public’s first look at the cohort, not evidence that it is the only performer, or the lead one.

Why this is a photonics story before it is a defense story

It is tempting to read a DARPA award as a defense story and stop there. The operational framing in the release invites it: faster threat detection, contested environments, situational awareness. But the technology question PICASSO poses is a photonics question, and it is the same question the commercial photonics industry is asking for different reasons.

In data centers, the optical interconnect industry is pushing toward co-packaged optics and, eventually, optical compute, for the same underlying reason: the conversion between optical and electrical domains is where latency and power are lost. The devices differ — a data-center switch fabric and a radar signal processor are not the same system — but the physics of the conversion penalty is identical, and so is the design problem of keeping a large photonic circuit predictable. A circuit architecture that solves signal integrity for a defense sensor would be directly relevant to a commercial optical processor, and vice versa. That is the dual-use logic DARPA programs are built on, and the reason this award belongs on a photonics desk rather than only on a defense desk.

Raytheon’s qualifications, as the company presents them, map onto that logic. Photonics process development is the foundry-facing skill: getting a design fabricated with acceptable yield. Radio frequency design knowledge is the systems skill: defense signals are often RF signals carried on optical carriers, and the interface between the two domains is its own discipline. Fielded military solutions are the credibility claim: Raytheon has shipped systems that work outside a laboratory. None of that guarantees the PICASSO work will succeed, but it explains why DARPA would put a defense prime on a problem that is, at its core, about photonic circuit architecture. The Synopsys partnership points the same way: circuit design at the scale PICASSO imagines requires simulation that models optical and electrical behavior together, and the commercial EDA industry has spent the last several years building exactly that capability for the data-center photonics market.

The defense framing, kept factual

The release frames the work in terms of operators detecting and responding to threats faster. That framing is accurate as far as it goes, and it should not be inflated. No specific weapon, sensor, or platform is named as a user of the technology. No service branch is identified as a customer. No deployment timeline is given, and the release’s own logic — a phase-one design and demonstration contract — places any fielded system years away at the earliest, behind a down-select and a second phase that has not yet been funded for Raytheon.

There is a debate worth naming rather than settling: defense funding may be the only vehicle for circuit-architecture research whose payoff is years out, or defense requirements — extreme environments, low volumes — may pull designs away from the commercial mainstream, where photonics scales. PICASSO’s structure — general circuit integrity first, application-specific work deferred to TA2 — reads as DARPA’s attempt to manage that tension. A narrower caution: “significantly faster response times” is a comparative claim with no comparator. Faster than what? The release does not say, and the number does not exist yet, because the circuit has not been built. Speed is the program’s goal, not its result.

What is undisclosed, and why it matters

An honest accounting of what this announcement does not contain is part of the story. No contract value. No delivery schedule. No performance target — no bandwidth, latency, energy-per-operation, or scale figure of any kind. No architecture: no material platform, no circuit topology, no statement of whether the work builds on silicon photonics, indium phosphide, silicon nitride, or something else. No foundry partner. No named customer inside the Department of Defense. No statement about whether the work will produce hardware, simulation results, or both by the end of phase one — the release says “design and demonstrate,” which could mean either, and the distinction matters for judging progress.

Each omission has a plausible explanation — contract values are often undisclosed at the performer’s discretion; architectures may be proprietary; phase-one performance targets are often proposer-defined metrics, as DARPA’s own FAQ confirms. But the pattern of omissions sets the limit of what can responsibly be claimed here. This article asserts no dollar figure, no timeline, no performance number, and no architecture for Raytheon’s PICASSO work, because none is on the record.

One more gap will determine whether this story matters in two years: the program’s success criterion is not a demonstration but survival through the down-select, where Raytheon’s work will be judged against other performers whose work is not public. Until then, the award is evidence of DARPA’s confidence in Raytheon’s proposal, not of the proposal’s correctness.

What to watch

The signals that will move this story from announcement to evidence are specific. First, the cohort: whether other phase-one performers announce their awards, and who they are. Second, a technical disclosure: any publication, presentation, or patent filing describing the circuit approach, material platform, or integrity strategy. Third, a demonstration marker: what “demonstrate” means by the end of phase one — a simulated circuit, a fabricated chip, or a system-level test. Fourth, the down-select, expected around early 2028: whether Raytheon advances to phase two, and on what evidence. Fifth, the TA2 solicitation: which applications DARPA prioritizes will reveal where the agency thinks the technology is headed. Sixth, manufacturing: whether a U.S. foundry partner emerges, given the solicitation’s stated preference for U.S.-based manufacturing. And a commercial signal: if PICASSO’s design methodologies show up in commercial photonic design flows, the dual-use logic is working as DARPA intends; if they do not, the program will have to justify itself on defense results alone.

The fair summary is a narrow one. On November 3, Raytheon announced a phase-one DARPA contract to design and demonstrate all-optical processing circuits under the PICASSO program, with Synopsys providing design and simulation tools. The program’s goal — photonic circuits that preserve signal integrity at scale without converting to electronics — is one of the field’s central unsolved problems, and DARPA has structured the effort, in two 18-month phases with a down-select, to find out whether it can be solved. No value, architecture, performance target, or timeline has been disclosed. That is a modest claim — and precisely because it is modest, it is believable.

Sources and further reading

RTX, “RTX’s Raytheon to develop new optical signal processing capability,” November 3, 2026 — phase-one DARPA contract for the PICASSO program; Raytheon’s Advanced Technology team to design and demonstrate photonic circuit technology using light rather than electricity to process signals; quote from Colin Whelan, president of Advanced Technology at Raytheon; collaboration with Synopsys for design and simulation tools; no contract value, delivery schedule, architecture, or performance target disclosed. RTX background figures from the same release: more than 180,000 global employees; 2025 sales of more than $88 billion; headquartered in Arlington, Virginia.

DARPA, program solicitation DARPA-PS-26-13, “Photonic Integrated Circuit Architectures for Scalable System Objectives (PICASSO),” January 14, 2026, Microsystems Technology Office — TA1 in two 18-month phases with a down-select; phase one on circuit-level signal-integrity strategies, phase two generalizing them to distinct functionality; TA2 not solicited under this solicitation; anticipated funding $35 million; proposals due March 6, 2026; estimated period of performance start July 1, 2026. DARPA PICASSO FAQ, version 3, March 6, 2026 — circuits must include both processing and signal regeneration; retiming left to proposer-defined metrics.

Washington Technology, “DARPA seeks to scale photonic computing with new program,” February 5, 2026 — reported DARPA planned several Other Transaction Authority awards worth up to $35 million each; note that this figure and the solicitation’s $35 million anticipated funding are reported differently, and neither is the value of Raytheon’s contract, which was not disclosed. Military Aerospace & Electronics, January 16, 2026 — solicitation coverage and deadlines. Defence Blog, November 3, 2026 — program managed by Anna Tauke-Pedretti of DARPA’s Multi X Office; target areas computing, analog signal processing, and sensing; no specific weapon, sensor, or platform named. ExecutiveBiz, November 4, 2026 — phase structure and down-select; Synopsys design and simulation tools.

Synopsys, product documentation for OptoCompiler and OptSim — background on Synopsys’ photonic tooling only; the RTX release does not name the tools Synopsys will provide, and no specific tool is asserted here as part of the Raytheon work. No other PICASSO phase-one performer had publicly announced an award as of publication, based on searches of trade coverage; DARPA has not published an awardee list. This article is analysis, not investment advice.