The lab system is not one buyer
Companies routinely talk about "selling to the national labs" as though the seventeen DOE laboratories were a single procurement office with a common set of terms. They are not. Each lab is a separate legal buyer with its own purchasing system, its own supplier portal, and its own small business program manager. Oak Ridge and Sandia both sit inside DOE, and getting on contract at one teaches almost nothing procedural about the other. What the labs do share is a legal structure, and that structure explains nearly every surprise a new supplier runs into.
Sixteen of the seventeen labs are federally funded research and development centers operated by a contractor under a management and operating agreement. The one exception is the National Energy Technology Laboratory, which is government-owned and government-operated with federal staff. Everywhere else, the entity that signs your purchase order is a private company or a university consortium: UT-Battelle at Oak Ridge, Battelle Energy Alliance at Idaho, Triad National Security at Los Alamos, National Technology and Engineering Solutions of Sandia at Sandia, Alliance for Sustainable Energy at NREL, UChicago Argonne at Argonne, Battelle Memorial Institute at Pacific Northwest. Ames National Laboratory, ten minutes from our office in Iowa, is operated by Iowa State University.
Management and operating contracts are defined in FAR Subpart 17.6, and DOE's own supplement lives at 48 CFR Part 970. The short version of what those documents establish: DOE owns the site and sets the mission, and the M&O contractor runs the place, hires the staff, and buys the goods and services. Your counterparty is that contractor. That single fact drives the rest of this article.
Where a small technical firm's effort tends to pay off
Editorial weighting from public program documents and practitioner reading, ordered by speed to a signed agreement for a firm with no prior lab history. Illustrative, not a measured statistic.
What the M&O structure changes for a supplier
When a lab buys your work, you are not holding a federal prime contract. You hold a subcontract under the M&O contractor's prime, and that has four practical consequences most first-time suppliers discover the hard way.
Set-aside rules do not apply the way you expect. The small business set-aside machinery in FAR Part 19 governs how the government awards prime contracts. It does not reserve lab subcontracts for small firms. What operates instead is the small business subcontracting program: the M&O contractor carries DOE-negotiated goals across the socioeconomic categories and reports achievement against them. That is a preference with teeth in practice, but it is a goal, not a legal set-aside, and it means you compete on the lab's terms rather than on a reserved list.
Most opportunities never appear on SAM.gov. Lab subcontract solicitations run through each lab's own supplier portal and sourcing team. A company that monitors only SAM.gov sees a small fraction of what the labs actually buy. Registering as a supplier at each lab you care about is unglamorous administrative work and it is the entry ticket.
You have no privity with the government. Disputes run against the M&O contractor under your subcontract terms. The Contract Disputes Act path at 41 U.S.C. chapter 71 belongs to the prime. Read the flowdowns closely, because selected FAR and DEAR clauses arrive in your subcontract and the rest do not.
On-site work brings DOE's safety regime. Any work performed on a DOE site sits under the worker safety and health program at 10 CFR 851, plus site access and badging that take weeks, not days. For software and data work this is usually a scheduling issue rather than a cost issue, but it belongs in the proposal schedule from the start.
What a lab subcontract does for your past performance
This question comes up on every teaming call we have about lab work, and the honest answer has two halves. A lab subcontract will not generate a CPARS record. CPARS covers federal prime contracts; subcontractors get no entry, so the work will not appear in the database a source selection team queries.
The other half matters more. FAR 15.305(a)(2)(iii) directs evaluators to consider past performance of subcontractors that will perform major or critical parts of the requirement, and of key personnel with relevant experience. A performance letter from an M&O contractor's technical monitor, naming the scope, the period, and the outcome, is admissible evidence and it carries real weight because the evaluator knows what a national lab's technical review is like. Ask for that letter at the end of the period of performance, while the sponsor still remembers the work.
CRADAs: joint work where no federal money reaches you
The Cooperative Research and Development Agreement comes from the Federal Technology Transfer Act of 1986, codified at 15 U.S.C. § 3710a. It authorizes a federal laboratory to enter a written agreement with a company, a university, or a state or local government to conduct research that serves both parties.
The defining constraint sits at 15 U.S.C. § 3710a(d)(1): under a CRADA the laboratory may provide personnel, services, facilities, equipment, intellectual property, and other resources, but it may not provide funds to the non-federal participant. The participant may provide funds to the lab. Money flows one direction, and it is not toward you. A company that walks into a CRADA conversation expecting a check has misread the instrument.
What a CRADA buys instead is access and intellectual property position. Under § 3710a(b), the collaborating party may retain title to inventions its own people make, and the agreement can grant an option for an exclusive license, in a negotiated field of use, on inventions made by lab staff during the collaboration. The government keeps a nonexclusive, irrevocable, paid-up license to practice the invention for government purposes. Two further provisions favor firms like ours: § 3710a(c)(4) directs labs to give special consideration to small business participants and to give preference to companies that will manufacture substantially in the United States.
The other underrated CRADA feature is data protection. Information produced under a CRADA that would be a trade secret or confidential commercial information if it had been obtained from a non-federal party may be shielded from public disclosure for up to five years from its development, under § 3710a(c)(7)(B). For a firm building a product on top of a joint research result, that window is the difference between a defensible position and a published head start for everyone else.
Strategic Partnership Projects: when you fund the lab
Strategic Partnership Projects is the mechanism DOE uses when a non-DOE sponsor pays a lab to perform work. It was called Work For Others until DOE renamed it in 2015, and plenty of lab staff still say WFO in conversation. The governing document is DOE Order 481.1.
SPP is the instrument behind a large share of what the labs do for other federal agencies, and it is available to companies as well. Three screening criteria apply before a lab can accept the work: the project must be consistent with the laboratory's mission, it must not adversely affect DOE programs, and it must not place the laboratory in competition with the domestic private sector. That last criterion is the one that kills company-sponsored proposals. If a commercial firm could do the work, the lab is not supposed to take it. Frame an SPP request around the unique instrument, the unique facility, or the unique expertise, or expect a decline.
Non-federal sponsors pay full cost recovery, generally in advance, and DOE adds an administrative charge on top for non-federal work. Budget accordingly. SPP is the right tool when a company needs a specific measurement, a materials characterization, an accelerated-life test, or a validation run on equipment that exists in four places on earth. It is the wrong tool when a company wants development labor.
The five instruments, side by side
| Instrument | Who pays whom | Best used for |
|---|---|---|
| Lab subcontract | Lab pays you | Delivering software, data, engineering, or analysis the lab has scoped and budgeted. Revenue today, relationship for later. |
| CRADA | No federal funds to you; you may fund the lab | Joint research on a shared question, with an option on an exclusive license to lab inventions and up to five years of data protection. |
| Strategic Partnership Project | You pay the lab, in advance, at full cost | Access to a one-of-a-kind facility, instrument, or specialty. Not for work the private sector can perform. |
| ACT agreement | You pay the M&O contractor | Negotiable commercial terms on IP, payment, and liability when standard SPP terms will not close a deal. |
| License | You pay royalties and fees | Taking an existing lab-developed technology to a market the lab cannot reach. |
| SBIR / STTR subaward | The awarding agency funds both parties | Federal money on both sides of the table, with the lab as your research partner and no cost share. |
ACT agreements: the flexible-terms path
Agreements for Commercializing Technology grew out of a DOE pilot and are now offered at most labs through the Office of Technology Transitions. The distinguishing feature is legal: the M&O contractor signs in its own name rather than as an agent of the government, which lets the parties negotiate terms that standard SPP agreements fix in place. Intellectual property allocation, payment schedules, indemnification, and liability caps all become negotiable.
The trade is priced in. Because the contractor takes on commercial risk it cannot pass to DOE, ACT work generally costs more than the same work under SPP. For a company whose investors or customers need commercial-style IP terms and a real indemnity, that premium is usually cheaper than the deal not happening.
Licensing what the labs already built
Every lab maintains a technology transfer office holding a portfolio of patents, copyrighted software, and know-how. Title to those inventions usually rests with the M&O contractor, either through Bayh-Dole election under 35 U.S.C. §§ 200 through 212 or through a patent waiver granted under DOE's authority at 42 U.S.C. § 5908. The government retains its practice license, and march-in rights under 35 U.S.C. § 203 sit in the background. In either case, the office that negotiates with you is the lab's, not DOE headquarters'.
Licenses come in the familiar shapes. Non-exclusive is fast, cheap, and available to competitors. Exclusive in a defined field of use is slower, costs more, and requires a commercialization plan the lab believes, with milestones that have teeth. Option agreements hold a position for six to twelve months during evaluation. Lab software is often released under open-source terms already, which is worth checking first.
Two federal resources make the search tractable. DOE's Lab Partnering Service publishes available technologies and identifies lab experts by subject. The Federal Laboratory Consortium, established at 15 U.S.C. § 3710(e), indexes technologies across all federal labs, not only DOE's. There is also money attached: the Technology Commercialization Fund, created by section 1001 of the Energy Policy Act of 2005 and funded from 0.9 percent of DOE's applied energy research appropriations, pairs lab projects with private partners on a 50 percent non-federal cost share. A company that shows up with a real market and a matching contribution is the profile that program was built for.
The SBIR and STTR route, which is the cleanest one
For a small technical firm, the mechanism where federal money funds both sides is SBIR and STTR. DOE runs its own program through the Office of Science, releases topics annually, requires a letter of intent before the Phase I application, and has funded Phase I awards in the $200,000 range with Phase II above a million.
The structural point that gets missed: under SBIR, the small business must perform at least two thirds of the Phase I work and at least half of Phase II, so a lab can take up to a third of a Phase I budget as a subcontractor. Under STTR, a federally funded research and development center qualifies as the research institution, the small business performs at least 40 percent, and the research institution performs at least 30 percent. A DOE national lab can be your STTR partner, and the arrangement carries no cost share.
An STTR with a lab does something a subcontract cannot. It puts a named lab scientist on your team, gives you a documented research relationship, and produces results both parties can point to. Our team has structured research-institution partnerships across university and federal-lab settings, and the pattern that works is always the same: bring the lab a question their people already want answered, not a request for validation of something you already finished.
User facilities and compute, the underused door
The Office of Science operates a set of scientific user facilities open to researchers from any sector. The access rule is the part companies miss: use is free of charge for non-proprietary work whose results are published openly, and available at full cost recovery for proprietary work. A company willing to publish gets access to instruments it could never buy.
For AI and data work the relevant facilities are computational. The Oak Ridge Leadership Computing Facility runs Frontier, the Argonne Leadership Computing Facility runs Aurora, and NERSC at Berkeley runs Perlmutter. Allocations come through INCITE, which is jointly managed by the two leadership computing facilities and awards large blocks through an annual peer-reviewed call; through the ASCR Leadership Computing Challenge for shorter high-impact projects; and through director's discretionary awards, which are small, fast, and the right first ask. A discretionary allocation is the lowest-friction way to start a real technical relationship with a lab, and the review turnaround is measured in weeks.
Getting in the door
From cold to first signed agreement
Step three is where companies either succeed or stall. Labs are organized around research problems and the people who own them. A message to a general procurement inbox describing your capabilities goes nowhere. A message to a principal investigator that engages a specific paper, names the limitation you can address, and proposes a concrete piece of work gets a reply, because that person has problems and a limited number of people to hand them to.
Mentor-protégé sits alongside all of this
DOE operates a mentor-protégé program in which prime and M&O contractors develop small business suppliers. It does not award contracts by itself, and it is not a substitute for winning work. What it does is create a structured relationship with a large contractor that already holds the lab's business, which is a durable path to subcontract flow.
Selling to a lab and partnering with a lab are different businesses
The clearest way to think about all six instruments is to ask which direction the money moves and who owns the result.
Selling. The lab has a budget, a scope, and a need. You compete, you deliver, you invoice. The lab owns the deliverable. Your gain is revenue, a technical reference, and a person inside the building who now knows what your team can do. This is the fastest path to a first transaction and the one most firms should start with.
Partnering. Both parties bring something and both parties want something out. Under a CRADA you contribute staff and possibly funds and take an IP position. Under a license you take the lab's finished work into a market it cannot reach. Under an STTR a federal agency funds both of you to answer a question neither would fund alone. Your gain is a product position rather than a receipt.
The failure mode is asking one instrument to do the other's job. A company that wants development labor paid for should not be writing a CRADA. A company that wants an exclusive IP position should not take a fixed-price subcontract to build the thing, because a subcontract deliverable generally belongs to the lab. Pick the instrument to match the outcome, and say plainly at the first meeting which outcome you are after. Lab tech transfer offices absorb a great deal of ambiguity from companies, and clarity is genuinely differentiating.
Where our team fits
Precision Federal builds AI, data, and software systems for federal, state, and commercial customers, and lab work sits squarely in that scope: data pipelines around instrument output, machine learning on scientific and operational datasets, model validation with documented provenance, and cloud and HPC engineering that has to survive review by people who know the science. We hold JCP / DD-2345 certification and CAGE 1AYQ0, we field named engineers and licensed professional engineers across the domains this work touches, and we work as prime or as subcontractor.
If your organization is an M&O contractor looking for a small business supplier on a data or AI scope, a lab researcher who needs a software partner on an STTR, or a company trying to decide which of these six instruments fits your situation, that conversation is a short one and we are glad to have it.
Frequently asked questions
No. 15 U.S.C. § 3710a(d)(1) specifically excludes providing funds to the non-federal participant. The lab contributes staff, facilities, equipment, and intellectual property, and the participant may contribute funds to the lab. The value a company receives is access and IP position, including an option on an exclusive license to lab inventions in a negotiated field of use.
They do not produce a CPARS record, because CPARS covers prime contracts. They are still citable: FAR 15.305(a)(2)(iii) directs evaluators to consider the past performance of subcontractors performing major or critical portions of a requirement. Get a written performance letter from the technical monitor at closeout and use it as the evidence.
They are the same mechanism under different names. DOE renamed Work For Others to Strategic Partnership Projects in 2015, and the governing document is DOE Order 481.1. Non-federal sponsors pay full cost recovery, generally in advance, plus a DOE administrative charge, and the project cannot place the lab in competition with the domestic private sector.
Yes. Federally funded research and development centers qualify as research institutions under the STTR program, so a national lab can hold the research-institution role. The small business must perform at least 40 percent of the work and the research institution at least 30 percent, and no cost share is required.
A straightforward subcontract can move in weeks once you are registered in the lab's supplier system. CRADAs and exclusive licenses commonly take two to six months because IP terms, field-of-use definitions, and internal approvals all have to clear. Director's discretionary compute allocations at the leadership computing facilities are among the fastest first steps, often turning around in weeks.