OUR SERVICES
What happens when a software license agreement, a development project, or a distribution deal breaks down, and how to protect your position on either side.
Most software disputes start as contract disputes. A customer deploys past its seat count. A development project blows through its milestones and nobody can agree what “done” means. A reseller sells outside its territory. The agreement everyone skimmed at signing becomes the most important document in the company, and the fight turns on license terms nobody remembers negotiating.
Software licensing disputes are, at bottom, fights over what a license actually permits: who may use the software, how much, where, and on what terms. The stakes are usually bigger than they first appear, because a software license agreement is not an ordinary contract. Software companies deliberately write their license agreements so that exceeding the license is copyright infringement in addition to breach of contract. That changes the available remedies, the leverage, and the size of the check. Here is how these disputes arise, how they work, and what decides them.
When you buy most goods, you own them, and the seller’s rights end at the sale. Software is built differently on purpose. Nearly every commercial software license agreement states that the software is licensed, not sold, restricts your ability to transfer it, and limits how you can use it. Under the leading test, from Vernor v. Autodesk, exactly those three features make the user a licensee rather than an owner of the copy. That status cuts off the first-sale and other ownership rights an owner would have under the Copyright Act. Not every court runs the same analysis. Some ask instead whether the customer exercises enough incidents of ownership over its copy, and under either approach the label alone never decides it. But the drafting pattern is national, and so is its purpose.
The second piece of deliberate architecture is the condition versus covenant distinction. A well-drafted license conditions the grant itself on compliance with its terms. Use the software outside the scope of the grant, and the licensor’s position is not just that you broke a promise. It is that your use was never licensed at all, so every copy the software made along the way was copyright infringement. That framing preserves two tracks of remedies at once: contract damages under the agreement, and copyright remedies. The copyright remedies can include injunctions and, where the copyright was registered in time, statutory damages and attorneys’ fees.
Courts police the theory. Under cases like MDY Industries v. Blizzard, a restriction supports an infringement claim only if it actually limits the scope of the license and connects to one of copyright’s exclusive rights. Ambiguous terms get read as covenants. A payment obligation or housekeeping rule does not become an infringement trigger just because the drafter labels it a condition. The same architecture powers open source licensing. Attribution and share-alike terms drafted as conditions have supported infringement claims, not just contract claims, since Jacobsen v. Katzer.
Which side of that line a given restriction falls on is a fact-specific question decided under state contract law. That question is often the whole ballgame in a licensing dispute. It is worth understanding before the dispute, not after.
A handful of provisions trigger most software licensing disputes and the litigation that follows:
If a vendor is invoking its audit clause against you, a software license audit is its own discipline with its own playbook. The Software License Audit Defense page covers it in depth.
Failed development projects and enterprise software implementations are among the most common software disputes, and they are won or lost on the project record. Four contract terms do most of the work for software developers and their customers:
When these projects end up in litigation, the record of requirements, change requests, test results, and the software’s actual technical behavior decides who is responsible for the failure. Building and dismantling that record is technical work as much as legal work.
Software sits awkwardly in commercial law. Article 2 of the Uniform Commercial Code governs sales of goods. It is generally understood to cover tangible products, but many courts have applied its principles to software transactions. That includes appellate decisions treating deals labeled licenses as sales of goods for all practical purposes. There is no clean national consensus. The answer can turn on whether title to a copy actually passes, and cloud subscriptions and SaaS contracts are frequently treated as services outside Article 2 entirely. The uncertainty is exactly why vendors disclaim the implied warranties of merchantability and fitness in conspicuous bold capitals, and why those disclaimers usually hold.
What a customer typically gets instead is a performance warranty: the software will conform to its specifications for 30 to 90 days, with remedies limited to repair, replacement, or refund. Two other risk-allocation terms deserve attention in every deal, because they set the boundaries of any later dispute. IP indemnification should obligate the vendor to defend infringement claims and to fix the problem. Fixing the problem means modifying the software, obtaining rights, or replacing it, with a refund if none of that works. The limitation of liability typically excludes consequential damages and caps exposure at the fees paid, subject to negotiated carve-outs for things like indemnity obligations and willful misconduct. These provisions interlock. In a dispute, they are read together, and a concession in one changes the value of the others.
Distribution relationships fail in their own ways. A value-added reseller or OEM agreement is a hybrid of license, distribution, and development agreement, and the intermediary wears all of those hats at once. That multiplies the ways the relationship can break. Recurring fights include resellers selling outside their territory or channel, disputes over minimum commitments and exclusivity, unbundling OEM software for stand-alone resale, and terminated partners who keep selling.
Three traps are worth flagging for software providers building a channel. First, resale pricing: since the Supreme Court’s Leegin decision, minimum resale price maintenance is judged federally under the flexible rule of reason rather than condemned automatically. But several states still treat it more harshly, which is why careful vendors use unilateral pricing policies instead of price agreements. Second, a trademark license plus a required fee plus significant control over the partner’s business can add up to an accidental franchise under the FTC Franchise Rule and state franchise laws. The consequences can include rescission and personal liability. A true buy-for-resale arrangement at bona fide wholesale prices avoids the required-fee element, so structure matters. Third, standing: a non-exclusive distributor generally cannot sue infringers on your intellectual property rights, so enforcement obligations and step-in rights belong in the agreement.
Some licensing disputes are stranger than a channel fight. Courts have seen customers go as far as reverse engineering a vendor’s product, decompiling and recreating it to avoid licensing it at all. The vendor’s rights in that situation come from copyright law, trade secret law, and the license together. That is exactly why the layers are set up the way they are. The boundary runs both directions, though. A third-party product is not an infringing derivative work merely because it works only with the licensed software, as the Ninth Circuit confirmed in Oracle v. Rimini Street.
Remedies in these cases run wider than ordinary contract damages. Licensors routinely seek injunctions on the theory that breach threatens irreparable harm to their proprietary rights, and license agreements often recite as much. The recital helps, but it does not control. Since the Supreme Court’s eBay decision, irreparable harm must be proven with evidence, not presumed. The dual contract-and-copyright track can add statutory damages and attorneys’ fees where the registration timing supports them, with one trap worth knowing. Infringement “commences” for those purposes at the first act, so a registration that arrives after the trouble started usually forfeits statutory damages even for what happens afterward. On the other side, licensees negotiate perpetual and irrevocable grants precisely to keep the software running during a fight and limit the vendor to money damages.
Many software agreements also route disputes through a tiered dispute resolution process, usually called an escalation clause: senior-executive negotiation first, then mediation, then binding arbitration. Courts enforce those clauses as written, including the sequence, so skipping the negotiation or mediation step can block a party from compelling arbitration until it complies. Two drafting notes keep the mechanism honest. First, arbitration is only as confidential as the clause makes it. Second, click-through terms and end user license agreements (EULAs) hold up only when the assent design is clear, since courts refuse to enforce terms hidden behind inconspicuous links. Done right, the machinery often resolves a licensing dispute in months rather than years.
Whether you are the vendor or the customer, the pattern in these cases is consistent. The side that understands the agreement, the deployment facts, and the technology first usually controls the dispute. If a software license or development relationship is starting to strain, the time to get counsel who can read both the contract and the code is before positions harden.
It can be both. If the agreement conditions the license grant on compliance, use beyond the scope of the grant supports a copyright infringement claim alongside the contract claim, which adds remedies like injunctions and, with timely registration, statutory damages and attorneys’ fees. Courts also require the violated term to connect to one of copyright’s exclusive rights rather than an ordinary business obligation. If the restriction is a mere covenant, or the transaction is treated as a sale, the claim usually sounds in contract alone. The distinction is technical and heavily litigated.
Not automatically. A standalone program is not among the categories that qualify as a specially commissioned work made for hire, so an independent contractor keeps the copyright unless a signed agreement assigns it to you. Employee-created code generally belongs to the employer, but IP ownership of contractor code requires paper. If you never got the assignment, fix it now rather than during a dispute or a financing.
Start with the agreement. Acceptance-testing provisions and performance warranties define your remedies, which typically run from repair or replacement through refund, and payment obligations may be tied to acceptance. Claims outside the warranty run into the disclaimer and limitation-of-liability provisions, so the analysis is usually about what survives those clauses. The project record, including specifications, test results, and change requests, will matter as much as the contract.
Generally not, if your agreement is drafted the standard way. Where the agreement grants a license, significantly restricts transfer, and imposes notable use restrictions, courts treat the customer as a licensee rather than an owner. In that case, the first-sale doctrine that permits resale of owned copies does not apply. The analysis is not uniform across courts, and the label alone does not decide it, so the drafting details matter.
If the agreement provides for it, yes: typically for uncured material breach, with shortened or no cure for serious violations like unpaid fees or out-of-scope use, or by declining renewal. Expect to deal with wind-down terms, existing end-user licenses that survive, inventory, and the return of confidential information. Check one more thing first: if the relationship combined your trademark, required payments, and significant control, state franchise laws can impose good-cause and notice requirements that override the contract. Terminating outside the agreement’s mechanics is how the terminating party becomes the defendant.