The Restraint Without a Present Tense

Introduction

A software contract is executed on a date, but dependence on the software develops over time. As data accumulate, employees learn the system’s conventions, internal processes are built around it, and other products are made to communicate with it. The expense of leaving therefore cannot be reduced to a price stated in the original agreement; it is also the product of everything that happens after the agreement is signed.

D.E. Shaw had used Geneva, a portfolio-accounting platform, for more than two decades. Its agreements allowed outside hosting, but not by a provider that competed with an SS&C portfolio-accounting product. After D.E. Shaw’s host, Arcesium, launched a competing product, D.E. Shaw changed hosting providers and later alleged that the restriction, together with SS&C’s control over maintenance and asserted rights in Geneva-generated data, violated federal antitrust law. By then, however, the agreements from which most of those claims arose had been signed in 2003 or 2014.

The Southern District of New York never decided whether D.E. Shaw had identified valid antitrust markets, whether SS&C possessed monopoly power, or whether the restrictions were unlawful. It dismissed most of the federal counterclaims because the four-year limitations period had begun when the contracts were executed, and it dismissed the remaining theory because that claim rested on litigation conduct protected by the Noerr-Pennington doctrine. The result also reflected shortcomings in D.E. Shaw’s pleading: the counterclaims did not adequately connect a later contract nonrenewal to D.E. Shaw’s migration costs, explain why those costs could not have been estimated earlier, or invoke an exception to Noerr-Pennington.

The opinion joins two bodies of doctrine that assign different dates to the same commercial relationship. Substantive aftermarket doctrine recognizes that switching costs and imperfect information may allow market power to arise after an initial purchase. Accrual doctrine may nevertheless locate the plaintiff’s injury at the earlier moment of contracting, so long as later costs were calculable. The evidence that makes lock-in economically plausible may therefore help make the resulting claim legally late.

A second issue arose from D.E. Shaw’s attempt to identify present conduct. When D.E. Shaw located the alleged restraint in the old agreements, its theory was assigned to the past; when it located the restraint in SS&C’s present assertion of proprietary rights in court, the lawsuit was assigned to the First Amendment. Between them remained an alleged commercial dependence for which the law could identify no actionable present act.

Contract-based antitrust law needs a more precise account of that interval. Existing doctrine properly distinguishes an original act from continuing effects and protects the repose supplied by a limitations period. It should also separate routine performance of an old agreement from a discretionary application of a contingent restriction to a new competitive condition. The further question is whether gradual dependence made an antitrust injury provable at the outset, apart from the ability to estimate exit costs. D.E. Shaw’s pleading may still have failed under that inquiry. Courts would nevertheless have a more precise way to determine when an old contract begins to govern a market that did not yet exist in the same form when signed.

I. The Contract and the Clock

The legal result turned on chronology, and that chronology began in 2003.

D.E. Shaw had licensed Geneva since 2003 under a Software License Support Agreement that granted a perpetual license and separately governed maintenance services, including updates, bug fixes, and technical support, for which D.E. Shaw paid annually. D.E. Shaw later alleged that this arrangement bundled support that third parties could have supplied, such as telephone consultation, with services that only SS&C could provide, such as critical updates and bug fixes.

In 2014, the parties amended the agreement to permit third-party hosting, but only on the condition that D.E. Shaw not use a provider that competed with an SS&C Advent portfolio-accounting product. D.E. Shaw later selected Arcesium as its host, and in 2015 SS&C and Arcesium entered a separate Hosted Reseller Agreement. After Arcesium launched its own portfolio-accounting product, UBOR, in 2020, SS&C declined to renew that agreement and, according to D.E. Shaw’s allegations, prohibited Arcesium from continuing to host D.E. Shaw’s Geneva instance. D.E. Shaw changed hosting providers between late 2021 and late 2022, and SS&C terminated its access to Geneva maintenance in 2023.

By the time the court addressed the counterclaims, the suit that produced them had already changed shape. SS&C had filed the original action, alleging trade-secret misappropriation and breach of contract, but the court dismissed its amended complaint on the pleadings in February 2026. D.E. Shaw’s counterclaims were what remained of the federal case.

Those events furnished the basis for counterclaims of monopolization, attempted monopolization, and unlawful tying, each pleaded under section 2 of the Sherman Act. D.E. Shaw alleged exclusive dealing in a market for portfolio-accounting software serving large, high-volume “Complex Funds,” tying in a Geneva-hosting aftermarket, and tying in a Geneva-maintenance aftermarket. It also advanced what the court called an “imprecise” theory of de facto exclusive dealing based on SS&C’s assertion of proprietary rights in data produced through Geneva.

The Clayton Act requires a private antitrust action to be brought within four years after accrual. Because D.E. Shaw’s counterclaims were compulsory, the court measured that period backward from October 18, 2023, when SS&C filed its complaint, making October 18, 2019, the dispositive date.

Although that conclusion gave D.E. Shaw the benefit of an earlier filing date, it did not solve the larger problem. Under US Airways, Inc. v. Sabre Holdings Corp., a contract-based antitrust claim generally accrues when the parties enter the agreement. The Second Circuit expressed the premise in temporal terms: “A contract is a vehicle for determining at the time of contracting what should happen at some time thereafter.” Later performance, on that account, manifests the original contracting decision rather than supplying a new overt act. Performance implements the bargain instead of repeatedly recreating it.

The rule does not render a long-term agreement categorically immune. A new and independent overt act that causes new and accumulating injury begins a new limitations period, and when damages were genuinely too speculative to calculate at the time of the violation, a claim for those injuries may accrue when they materialize. Neither exception, however, is satisfied merely because an old contract remains costly.

Applying that framework, the district court treated the 2014 amendment as the source of the exclusive-dealing and hosting-tying theories, while locating the maintenance tie in the 2003 agreement. Both therefore fell outside the limitations period.

D.E. Shaw identified two later events. The first was SS&C’s 2020 decision not to renew the Hosted Reseller Agreement with Arcesium, but that decision did not restart the period because D.E. Shaw was not a party to the agreement and did not derive its Geneva license or hosting rights from it. More importantly, the counterclaims failed to explain how the nonrenewal forced D.E. Shaw to migrate. The 2014 amendment, rather than the reseller agreement, governed D.E. Shaw’s use of third-party hosts, and whatever causal account might have connected SS&C’s later conduct to D.E. Shaw’s migration was absent from the pleading.

The second event was SS&C’s 2023 termination of maintenance. Although new in an ordinary chronological sense, the termination was not new in the sense relevant to D.E. Shaw’s tying theory. The alleged tie consisted of requiring the purchase of optional support in order to obtain necessary updates and fixes; termination did not impose that bundle again, but instead ended access to both of its components. The court accordingly treated the annual payments made before termination as performance of the 2003 agreement and the termination itself as conduct distinct from the alleged tie.

The speculative-damages argument failed for a related reason. D.E. Shaw pointed to migration expenses incurred in 2022 and operational risks that became concrete after maintenance ended in 2023, but the court asked not when the precise figures became known, but whether the injuries could have been calculated earlier. It concluded that they could. The 2014 restriction made the possibility of migration apparent if a host became a competitor, and D.E. Shaw, which had previously hosted Geneva internally, possessed information relevant to estimating the costs. Its assertion that those damages had been incapable of reasonable estimation was therefore conclusory.

The court’s reliance on Chalmers v. NCAA and SL-x IP S.à.r.l. v. Merrill Lynch sharpened the point: uncertainty in amount is not the same as legal speculation, and a future consequence may be difficult to price without being impossible to value.

The court identified the challenged provisions, assigned each claim to its contractual source, tested the alleged later acts for independence and causation, and required facts supporting delayed calculability. D.E. Shaw’s theory asked the court to move the clock forward, but its pleading did not explain with sufficient precision why the clock should move.

The more difficult question appears only when that temporal analysis is placed beside antitrust’s substantive treatment of lock-in.

II. The Aftermarket Paradox

In Eastman Kodak Co. v. Image Technical Services, Inc., the Supreme Court rejected the categorical assumption that competition in an initial equipment market necessarily prevents power in a market for parts or service. Customers may lack the information required to calculate a product’s full lifecycle cost at the time of purchase, and once they have invested in the equipment and in the materials or processes that support it, switching may become expensive. Together, those conditions can weaken the discipline that the primary market would otherwise impose on the aftermarket.

This does not mean that every installed base creates a monopoly. It means that present alternatives must be evaluated as a matter of commercial reality, because a customer who could have selected another system years earlier may have no practical substitute now.

Although Sabre involved a single-brand market for Sabre’s platform rather than a derivative aftermarket, the decision contains both sides of the temporal problem. In addressing market definition, the court drew on Kodak to hold that US Airways had adequately pleaded a market limited to Sabre’s own platform because travel agents allegedly faced prohibitively high switching costs and lacked viable substitutes. Yet in addressing limitations, the court treated each supracompetitive charge under an old agreement as a manifestation of the original contract. The same opinion therefore looked to present lock-in when asking what the market was, but to the original commitment when asking when the injury occurred.

The inquiries are coherent when considered separately: market definition concerns substitution in the present, whereas accrual concerns the act that caused injury. When market power develops through the relationship itself, however, the two inquiries can point in opposite directions.

SS&C inherits that tension. D.E. Shaw alleged two Geneva-specific aftermarkets, one for hosting the software and another for maintaining it. The court summarized those allegations and cited Kodak for the general proposition that a firm may exercise power in a secondary market for services specific to a primary product. It never decided whether the alleged aftermarkets were legally cognizable, however, because limitations ended the inquiry first.

That sequence produces a paradox. Switching costs can help establish why customers are locked into an aftermarket, while the foreseeability of those same costs can establish that damages were calculable when the original contract was signed. The stronger the allegation that departure is costly, the easier it may be to say that the plaintiff should have valued the cost at the beginning. Evidence of dependence can thus make the substantive theory more plausible while making the procedural theory less timely.

Substantive doctrine has developed its own version of this discipline. Post-Kodak decisions draw a line between aftermarket power arising from market imperfections or undisclosed post-purchase conduct and power arising solely from contractual rights knowingly granted at the outset. A buyer subject to a disclosed restraint may be constrained by its bargain rather than exploited through the informational failures and switching costs that animated Kodak.

At contracting, substantive aftermarket analysis considers whether the asserted power arose from the bargain or from later imperfections in the relationship. Accrual analysis considers whether the resulting injury could then be valued. Current doctrine supplies no general rule for determining when the later antitrust injury came into existence.

One way to see the resulting difficulty is to separate the relationship into three clocks. The contract clock begins when the parties accept the restriction, furnishing a definite date and serving the interest in repose. The dependence clock measures the gradual accumulation of data, integrations, training, and operational reliance, and therefore has no single starting point. The competitive-injury clock begins when the supplier’s position, the customer’s diminished alternatives, and the challenged conduct combine to produce a cognizable antitrust injury—a moment that may coincide with the contract, but need not.

Contract accrual tends to align all three clocks. If an agreement states the condition and the cost of exit can be estimated, later consequences are attributed to the original decision. Yet a calculable cost is not necessarily a present antitrust injury. A firm may be able to estimate the expense of changing vendors while lacking any basis to allege monopoly power, substantial foreclosure, or even the existence of the rival whose entry will activate the restriction.

The chronology alleged in SS&C illustrates the distinction. The hosting restriction was adopted in 2014, before Arcesium became the host and six years before it launched the allegedly competing product. The contractual condition therefore preceded the competitive event that gave the condition its alleged significance as applied to Arcesium. Migration costs might have been estimable in 2014, but whether the restriction would later operate against an actual host that had become a rival was a different question.

The problem is not confined to hedge funds or portfolio-accounting software. Any organization that builds years of data, employee practices, regulatory processes, and third-party integrations around an enterprise platform may confront the same temporal divide. The contract fixes formal rights at the outset, while the economic character of the relationship changes through use. By the time dependence becomes visible as a market condition rather than an ordinary cost of doing business, the legal challenge to the clause that helped shape it may already be old.

That observation should not be converted into a disguised discovery rule. Antitrust accrual does not ordinarily wait until a plaintiff has assembled enough evidence to win, nor should a sophisticated firm preserve a claim indefinitely by asserting that an express restriction became objectionable only when compliance grew painful. A limitations period is meant to force some disputes to be brought early or lost.

The speculative-damages inquiry should remain tied to the antitrust injury alleged. Estimating a hypothetical migration project does not establish that an exclusionary injury in a defined market was then provable. Courts should ask what antitrust injury existed at contracting and whether it could be valued then.

Contingent restrictions sharpen the distinction. A provision may prescribe what will occur if a future business partner becomes a competitor, if a new technology is introduced, or if interoperability develops in a way that threatens the incumbent product. The provision exists at execution; its competitive application may arise only when the condition occurs. A court treating a later invocation as ordinary performance should identify what the earlier agreement had already done to competition and how the triggering event changed the relationship.

SS&C did not need to resolve that distinction because D.E. Shaw failed to plead the bridge between the old provision and a legally adequate later act causing new injury. The pleading did not supply that bridge.

III. The Protected Present

D.E. Shaw’s de facto exclusive-dealing theory presented a different problem. It alleged that SS&C asserted proprietary and trade-secret rights in the organization of data produced through Geneva, thereby restricting customers’ ability to move to competing systems. If the theory rested on the data-ownership provisions of the 2003 agreement, it was untimely for the same reason as the other contract-based claims. D.E. Shaw also pointed to SS&C’s contracts with two other customers, but the court found that those agreements showed only that SS&C had reserved rights against parties other than D.E. Shaw. The pleading, however, did not clearly rest on contract alone; its operative allegation was that SS&C presently “asserts” those rights, principally through the trade-secret claims it had brought against D.E. Shaw.

That choice of verb moved the analysis from limitations to Noerr-Pennington. The court took the word seriously enough to consult two dictionaries and concluded that “asserts” described a legal position advanced in litigation rather than a restriction imposed by contract. Subject to established exceptions, Noerr-Pennington protects efforts to petition the government, including access to courts, from Sherman Act liability. The doctrine does not declare the underlying legal position correct; it prevents antitrust liability from attaching merely because a party seeks governmental action, even when the petitioning is alleged to have an anticompetitive purpose.

Because D.E. Shaw did not argue that an exception applied, the court treated SS&C’s litigation position as protected conduct and dismissed the remaining federal theory.

The petitioning protected by Noerr-Pennington had already failed: five months earlier, the same court had dismissed SS&C’s amended complaint, including the trade-secret claims on which D.E. Shaw’s de facto theory rested, on the pleadings. That history properly changed nothing, because immunity does not depend on success and a claim is not objectively baseless within the meaning of the sham exception merely because it fails. Noerr-Pennington protects the act of asking a court to adopt a legal position, not only successful petitions.

Together, the holdings assigned the contractual restrictions to past acts and the present assertion of rights to protected petitioning. Limitations doctrine reclassified the contractual restrictions as past acts, while Noerr-Pennington reclassified the present assertion of rights as petitioning. The first doctrine treated the allegedly restrictive conduct as something that had already happened; the second treated the conduct happening now as something other than the kind of market behavior the Sherman Act could condemn on the theory pleaded.

The old contractual theories were time-barred, and the present litigation theory was protected petitioning. The court left the legality of the alleged restraint unresolved. The continuing commercial relationship therefore remained outside the antitrust claims as pleaded.

There are sound reasons for both rules. Without a demanding overt-act requirement, every payment or refusal under a long-term agreement could renew liability and deprive the limitations statute of practical effect; without protection for petitioning, antitrust law could make resort to courts a source of treble damages whenever litigation affected a competitor. The difficulty arises from the rules’ combination, because each doctrine sees only part of the course of conduct and places that part outside the claim for a different reason.

A plaintiff cannot establish timeliness merely by narrating a long relationship as one continuous wrong; it must identify the present act with precision. That precision is more than a pleading demand, because it determines which body of law receives the act. Characterize the conduct as performance, and Sabre may assign it to the original contract; characterize it as petitioning, and Noerr-Pennington may immunize it; characterize it as a discretionary application of an old restriction to a new competitive condition, and an overt-act inquiry becomes possible. The labels are not interchangeable, and neither are their consequences.

IV. The Act That Counts

The tension exposed by SS&C invites three competing instincts. One is to anchor accrual to the contract, gaining certainty and repose but risking a challenge before a rival, market, or distinct injury exists. Another is to restart the clock when a contingent restriction is invoked against a new competitor or technology, which captures a present commercial choice but risks redescribing routine enforcement as a fresh violation. A third is to wait until lock-in or foreclosure becomes a cognizable antitrust injury, an approach attentive to commercial reality but sufficiently indeterminate to resemble a discovery rule that Congress did not enact. None is satisfactory as a general rule.

A narrower approach can preserve the advantages of contract accrual without treating every later development as part of the original bargain. Courts should continue to regard ordinary performance of a settled obligation as a manifestation of the agreement and should require any later overt act to be both independent and causally connected to new injury. Within that framework, however, they should distinguish among performance, application, and maturation.

Performance occurs when the parties carry out what the old agreement already and automatically required. Application occurs when a party must make a present, discretionary determination that a new condition falls within the restriction—for example, that a host has become a competitor or that a new use violates a proprietary-rights clause. Maturation describes the gradual deepening of the relationship’s economic consequences without a distinct decision by the defendant.

Ordinary performance should not restart the clock. Application may qualify as an overt act when it changes the plaintiff’s competitive choices and causes new injury, provided the plaintiff pleads the act, the contractual instrument, and the resulting injury in alignment. Maturation alone should not create a perpetually renewable claim, but it should bear on whether an antitrust injury—as opposed to an abstract or hypothetical exit expense—was reasonably calculable when the contract was executed.

The inquiry should therefore focus on four connected matters: the precise conduct alleged to be anticompetitive, whether the later conduct followed automatically from the old agreement or required a new decision, and what injury that decision caused. It should then ask whether that antitrust injury—rather than merely the possibility of future expense—could have been proved and valued at contracting. This formulation preserves Sabre‘s central insight that continuing performance cannot erase a limitations period, while preventing “calculable” from becoming a synonym for “imaginable.”

Applied to SS&C, the distinction confirms much of the opinion. The 2023 termination of maintenance did not apply the alleged maintenance tie anew; it ended the services said to be tied. The 2020 nonrenewal of the reseller agreement was not adequately connected to an injury suffered by D.E. Shaw, and the counterclaims did not explain why the relevant damages were incapable of earlier calculation. On those pleadings, dismissal was unsurprising.

The framework nevertheless identifies the question the pleading left undeveloped: whether a present decision applying the 2014 restriction after a host became a portfolio-accounting competitor was analytically different from routine performance of the agreement. Neither the contract date nor the later migration answers that question by itself. What D.E. Shaw failed to provide was the causal and temporal account connecting the two.

A viable claim would therefore have needed to identify the later act and connect it to new injury.

Conclusion

SS&C Technologies v. D.E. Shaw did not determine that SS&C monopolized a market, unlawfully tied services, or foreclosed a rival. The court dismissed the counterclaims under existing Second Circuit law without deciding the alleged markets, monopoly power, or legality of the restrictions. The unresolved issue is how accrual should operate when dependence and competitive conditions develop after contracting.

Antitrust doctrine knows how to identify the date of a contract, and it also recognizes that an aftermarket may become legally cognizable because customers become locked in after the initial purchase. What it does not yet describe with equal clarity is the relationship between those propositions. A contract may precede dependence; dependence may precede the emergence of market power; and market power, if it arises at all, may precede the later act that turns power into injury.

The law need not leave every old agreement open to perpetual challenge, because repose is not an embarrassment to be reasoned away. But neither should the date of a clause silently become the date of every market condition the clause might later govern.

A customer can foresee that leaving will be costly without yet possessing an antitrust claim, just as a contract can anticipate a rival without creating one. And the later application of an old restriction may be neither automatic performance nor a newly invented agreement, but a distinct act whose legal significance depends upon what changed, what injury followed, and what could have been known before.

An antitrust system capable of recognizing an aftermarket should also explain when injury in that market begins. Courts therefore need a rule that identifies when later application of an old contractual restriction causes a new antitrust injury.