Babak Siavoshy joined me on the Acquisition Talk podcast to discuss opportunities for improving intellectual property and software data rights practices in defense contracts. He is Vice President and General Counsel of Anduril Industries, and has recently written an excellent article on one of the least understood aspects of the acquisition world.
Babak extends the argument that software is “eating the world,” or comprising an increasingly larger share of the value being generated by firms today. Even for the physical products Anduril is developing like UAVs, sensor towers, drone interceptors, and submarines, the software backbone is what makes them game changing. Government, however, has struggled to value software and compensate for private investment. These issues show up in its templatized data rights practices.
Download the full-text transcripts
Commercial Rights
Commercial data rights in theory fit the need. Government doesn’t need to determine the license terms or even the pricing. The commercial markets provide the pressures to assure government is getting a fair deal. But as Babak points out, commercial data rights “are a somewhat awkward fit for military technology.”
Consider the fact that commercial items aren’t just those sold to commercial markets, like Apple laptops or automotive parts, but those items based on commercial technologies but tailored for DoD use. This has led to many circumstances where firms offer a “commercial item” that is heavily modified in product details, causing government officials to push back – and for good reasons. “The government I think is correct,” Babak argues. “Don’t you think it’s weird that your missile targeting software has a commercial license attached to it?”
Government Purpose Rights
Government Purpose Rights (GPR) are another major category and the default template whenever there is mixed funding in the product, some government RDT&E funding and some private funding. This allows government to take possession of software code, share it across agencies, and contract with other firms to modify it so long as the code is exclusively used for government.
One problem is that GPR incentivizes firms to build bespoke software, or even spaghetti code, so that it can only work for that one requirement. Babak recommends narrowly scoping licenses to a program, location, or user because that incentivizes firms to make highly interoperable products so that they can sell the same thing to other agencies. This double-charging seems inefficient, but is actually highly effective in preventing multiple programs rebuilding the same thing. “The reason the software titans of our age have become so valuable is largely because they have one shared code base that they’re selling across many different customers… It’s a lot of work to maintain a shared code base.”
Another problem of GPR is that it makes firms unwilling to take a small prototype contract for fear that their product ideas or even the code base itself will be competed out for the production contract. The innovating company might then lose out to a large prime who bids low-priced labor. “What scares them away is that they’re being asked to sell the farm at the very early stages.”
Babak recommends that government not ask for data rights to experiments and demonstrators when what they really want is to understand what is possible. But when moving into a production contract, Babak says that Anduril and other firms would be willing to provide greater data rights to government than is normal in the commercial sector. There could be a sliding scale, but until government commits to $20 million or $50 million, data rights should be favorable to the contractor.
Planning for data rights early
The reason defense policy has been pushing for ordering data rights of GPR or more so early in the acquisition is because officials have found themselves locked into sole source vendors for the class of programs started in the 1990s and 2000s. The pain from “swiss cheese” data rights is real. Arguments over sustainment data rights on the F-35, for example, have gone on for years and the price tag has recently been put at half-a-billion for DoD to organically sustain and compete out spares and repairs.
Renewed policy focuses on having contractors separately price data rights in their early-stage proposals so that competition makes sure their terms and pricing are reasonable. But so long as government hasn’t entered a de facto sole source situation, such as awarding a single firm a major Engineering and Manufacturing Design contract for a post-Milestone B program of record, the requirements for data rights can be deferred.
Let’s do it today!
Perhaps the most important aspect of Babak’s recommendations are that they can be achieved fully within the today’s acquisition rules. The FAR allows for specifically negotiated rights. “There are all the tools that we need to have a flexible data rights regime without any changes.” He continues:
We can already do this. You just need a smart contract officer who is willing to be nimble, has top cover, and can negotiate these specific rights. But that is not how things work in practice. In practice, there’s an incredible amount of inertia behind these existing categories. And the proposal is, let’s pilot another templatized category of rights that is more flexible and more accommodating of the needs of this software-defined world that everyone agrees we are moving towards.
Listen to the whole thing for more.
Thanks Babak Siavoshy!
I’d like to thank Babak Siavoshy for joining me on the Acquisition Talk podcast. Be sure to read his full article on software data rights – a lot more is there. He has another article on software-defined approaches to acquisition, and was featured by Business Journal in their General Counsel Awards. Babak has also been on the Execs podcast with Erik Torenberg as well as the Operators podcast with Delian Asparouhov. ICYMI, Mason GovCon’s Acquisition Next playbook has a section on intellectual property that rhymes with Babak’s.
Leave a Reply