The Borderless-Internet Defense Just Died in the Ninth Circuit

By Brandon Orewyler | July 2026

A California investor puts more than $750,000 into a digital-art venture run by a United Kingdom resident through a Hong Kong entity. The money, he alleges, does not come back. When he sues in California, the district court dismisses for lack of personal jurisdiction, accepting the defense that the defendants are foreign, the project was global, and California cannot reach them. In April 2026, the Ninth Circuit reversed, unanimously.¹ I briefed and argued that appeal, Gelasio v. Zafar, and the reversal was not a surprise. It was the predictable application of a rule the Ninth Circuit announced a year earlier, sitting en banc (the full court, not a three-judge panel), in Briskin v. Shopify, Inc.²

For two decades, an internet defendant’s surest exit from an American courtroom was personal jurisdiction (the constitutional requirement that a court have authority over the defendant, not just the dispute). The argument was always the same: our product works identically everywhere, so we never targeted any particular state. Courts accepted it, and cross-border operators built it into their risk models. Briskin rejected the premise outright, and Gelasio showed the rule reaching a foreign crypto operation on ordinary facts. The Supreme Court then declined an invitation to revisit the question, denying certiorari in a parallel internet-jurisdiction case, Meta Platforms, Inc. v. Vermont, on May 26, 2026.³ The borderless-internet defense is dead in the Ninth Circuit, and no rescue is scheduled.

One caveat up front: the doctrine still has a seam. Known-resident contacts count particularly where the defendant conducts regular business in the forum (the Gelasio panel’s reading of Briskin), and a defendant with genuinely thin forum contacts will press that clause hard. The rest: what changed, what the new rule looks like applied, how far it reaches beyond crypto, and where the fight moves next.

Key Takeaways

• The old shield is gone. In the Ninth Circuit, express aiming no longer requires “differential targeting” of the forum state or a forum-specific focus; a platform that treats every state the same can still be sued where its known counterparties live.²
• Knowing your counterparty’s location is now a jurisdictional fact. Deliberate dealings with a person the defendant knows is in California count toward express aiming.²
• Foreign incorporation is not an answer. Gelasio applied the rule to a UK resident and a Hong Kong entity behind an NFT project, on contacts as ordinary as loan requests, a recorded call, and promoter solicitations aimed at Californians.¹
• The logic is not crypto-specific. Any operator whose defense was “same product everywhere” (AI tools, SaaS, marketplaces, payment rails) should re-run its exposure analysis.
• The Supreme Court passed. The May 2026 cert denial in Meta v. Vermont leaves Briskin standing; the live limits are doctrinal, not appellate.³

Why “We Never Targeted California” Used to Win

Until April 2025, a globally accessible website was close to a jurisdictional safe harbor in the Ninth Circuit. The court’s cases required a plaintiff invoking the Calder effects test to show the defendant “expressly aimed” its conduct at the forum state, and the court had glossed that requirement to demand differential targeting: something the defendant did toward California specifically, beyond operating a site Californians could use like anyone else.⁴

Three decisions built the shield. AMA Multimedia, LLC v. Wanat held a foreign-based site with a large American audience was not subject to jurisdiction because it did not single out the forum market. Will Co. v. Lee and Doe v. WebGroup Czech Republic, a.s. ran the same logic for other foreign operators.⁵ The practical effect was perverse and well understood by defense counsel: the more uniformly global a platform was, the safer it was. An operator that “targeted everyone” had, on this logic, targeted no one. For cross-border crypto ventures (which are global by architecture and often foreign-incorporated by design), the shield was close to purpose-built.

If your jurisdictional defense has always been “we serve every state the same way,” understand that in the Ninth Circuit that sentence now cuts the other way.

What Personal Jurisdiction Requires Today

Specific personal jurisdiction in California federal court still runs on the familiar constitutional architecture: California’s long-arm statute extends to the limits of due process, so the whole analysis is the federal minimum-contacts test.⁶ The Ninth Circuit applies a three-part test: the defendant must purposefully direct its activities at the forum or purposefully avail itself of forum benefits; the claims must arise out of or relate to those forum contacts; and the exercise of jurisdiction must be reasonable. The plaintiff carries the first two prongs; the burden then shifts to the defendant to make a compelling case that jurisdiction would be unreasonable.⁷

For intentional torts, purposeful direction is measured by the Calder effects test: an intentional act, expressly aimed at the forum state, causing harm the defendant knows is likely to be suffered there.⁸ Two Supreme Court guardrails frame everything that follows. Walden v. Fiore requires that the contacts be the defendant’s own conduct connecting it to the forum, not merely the plaintiff’s residence there.⁹ And general jurisdiction (suing a company anywhere for anything) remains confined to where a corporation is at home, so a Hong Kong entity is not generally suable in California no matter how the specific-jurisdiction fight comes out.¹⁰ Nothing in Briskin touched either principle. What Briskin changed is what counts as the defendant’s own forum-directed conduct.

Briskin v. Shopify: The En Banc Decision That Broke the Shield

Sitting en banc in April 2025, the Ninth Circuit held that express aiming does not require “differential targeting” of the forum state or a “forum-specific focus,” and it expressly overruled the AMA Multimedia line to the extent it demanded one.² Contacts with a plaintiff the defendant knows is in the forum count toward express aiming (particularly, as the Gelasio panel would later put it, where the defendant conducts regular business there). Shopify, which was alleged to have extracted and monetized a California consumer’s data through its payment infrastructure while knowing his location, could be sued in California even though its platform operates identically everywhere.²

The clause that matters most going forward ties known-resident contacts to a defendant that conducts regular business in the forum. That formulation is the Gelasio panel’s reading of Briskin at 758-59; the en banc holding itself is that an interactive platform expressly aims its wrongful conduct at a forum state when its contacts are its “own choice and not ‘random, isolated, or fortuitous.’”² Read one way, regular in-forum business is a required element, and a defendant with thin forum contacts escapes. Read the other way, it is an intensifier: knowing contacts with a forum resident suffice, and regular business simply makes the case easier. A thin-contacts defendant will press the first reading at full strength, and it is the strongest argument left on the defense side. My read (and the way Gelasio was briefed and decided) is that the en banc reasoning points to the intensifying reading. The holding language is choice-based, not volume-based. And the court’s rejection of differential targeting makes little sense if a separate regular-business element quietly reimposes it.² How far the clause reaches is the seam future panels will litigate.

Two points defense counsel will raise, so plaintiffs should raise them first. On remand, Briskin itself stalled on the merits in January 2026: the district court dismissed every claim as pleaded, all but one with leave to amend.¹¹ The jurisdictional holding stands. Winning jurisdiction is necessary, not sufficient. And Judge Callahan’s dissent warned that attaching jurisdiction to wherever the plaintiff happens to be risks the jurisdiction-follows-the-plaintiff problem Walden forbids; that critique remains the sharpest doctrinal objection to the new rule, and district courts will be policing the line.¹²

If you knew your counterparty was in California when you dealt with them, that knowledge is now a jurisdictional fact, not a neutral detail.

Gelasio v. Zafar: What the New Rule Looks Like Applied

Gelasio is the proof of concept. The defendants were a UK-resident founder and a Hong Kong entity behind an NFT venture called Project Ether; the plaintiff was a California investor who had put more than $750,000 in. The district court dismissed for lack of personal jurisdiction before Briskin came down. Applying Briskin, the Ninth Circuit panel reversed, unanimously.¹

The contacts the panel found dispositive were ordinary business conduct, not exotic crypto facts: repeated loan requests directed to a plaintiff the defendants knew was in California; the surreptitious recording of his conference call while he was in California; a California-based software engineer and a California attorney working on the project; solicitation of people they knew or should have known were California residents to promote the venture in exchange for NFT-presale access; and offers to sell NFTs to Californians.¹ Express aiming was satisfied on those facts, and the defendants conceded the other two effects-test prongs (an intentional act, and harm they knew would be felt in California).

On relatedness (the second prong), the panel took a hub-based path worth understanding precisely, because it is a model for pleading. The claims arose out of and related to the California contacts: the in-forum solicitation and recording took place while the plaintiff was in California, and the other California activities were related to Project Ether, the venture from which all of the claims arose.¹ The panel found the claims arise out of and relate to the contacts without demanding claim-by-claim causation from each contact. For a plaintiff structuring a complaint against a foreign operator, that hub approach (tie every pleaded contact to the venture the claims arise from) is the practical takeaway.

Reasonableness did not save the defendants either. Under the Ninth Circuit’s seven-factor test, only one factor (the availability of an alternative forum) weighed against jurisdiction (the rest were neutral, leaned only slightly against, or favored California), and one factor is not the compelling case the burden-shifted defendant must make.¹³

One discipline point: Gelasio is a memorandum disposition, unpublished and not precedent under Ninth Circuit Rule 36-3.¹⁴ Briskin is the binding authority. Gelasio’s value is evidentiary, not precedential: it shows a panel applying the en banc rule, with no judge writing separately, to reach a foreign crypto operation on ordinary facts, which is exactly the fact pattern cross-border operators told themselves was unreachable.

What “Expressly Aimed at California” Looks Like Now

If a foreign operator dealt with you knowing you were in California, the Gelasio contact categories are the pleading checklist.

How Far Beyond Crypto Does Briskin Reach?

The engine of Briskin is not crypto-specific. It is the combination of an interactive service accessible in the forum with knowing, deliberate dealings with forum residents. That description fits a large share of the modern technology economy.

An AI-tool or API provider that onboards customers it knows are in California, bills them there, and supports them there is running Gelasio-grade contacts at scale. A court applying Briskin could find express aiming on those facts; the analogy weakens where the provider genuinely does not know customer locations. A SaaS platform whose defense to date has been “same functionality in every state” is squarely the defendant Briskin describes; its remaining argument is the thin-contacts seam, not the old global-neutrality shield. Marketplaces and payment rails sit closest to Briskin itself, which involved payment infrastructure knowingly processing a Californian’s transactions.²

Three limits remain. Walden still bars jurisdiction built only on the plaintiff’s own forum ties, and Daimler’s at-home rule still confines general jurisdiction (both untouched, as noted above).⁹ The regular-business clause is the third and the only live one: a foreign defendant whose only California contact is one known counterparty, with no in-forum personnel, solicitation, or recurring business, presents the case Briskin did not decide. Those limits are real, and pretending otherwise is how plaintiffs draw sanctions and operators get bad advice.

If you operate from abroad and knowingly take money from California users, “we’re incorporated overseas” is no longer a jurisdictional answer.

What Plaintiffs and Operators Should Each Do Now

For plaintiffs’ counsel chasing a foreign or offshore operator, the threshold work is now contact development, structured around the Gelasio contact categories (known-resident dealings, in-forum conduct, forum personnel, solicitation of forum residents, offers to forum residents). Plead the defendant’s knowledge of the plaintiff’s California location and the dealings that followed it; plead in-forum conduct, forum personnel, and solicitation of forum residents; and tie every contact to the venture the claims arise from, the hub approach the Gelasio panel accepted. Plead the contacts separately for each defendant; the Gelasio panel analyzed the defendants jointly only because the parties briefed it that way. Where the record is thin, seek jurisdictional discovery, which should ordinarily be granted where pertinent jurisdictional facts are controverted or a more satisfactory showing is necessary.¹⁵

For operators and their general counsel, the exposure analysis inverts. Knowing customer location (which compliance programs increasingly require) now cuts both ways: it satisfies regulators and supplies plaintiffs a jurisdictional fact. The immediate work is an inventory: where are the counterparties you know about, what dealings follow that knowledge, and which of the Gelasio contact categories does your operation light up. Foreign incorporation, offshore servers, and uniform global products no longer answer the question. Contractual forum-selection and arbitration clauses do real work here precisely because the default forum rules no longer protect you; that is a drafting project, not a litigation position.

Could the Supreme Court Take This Away?

Not soon, on the current record. A clean vehicle came and went: the petition in Meta Platforms, Inc. v. Vermont, No. 25-909, presented the internet personal-jurisdiction question, drew a call for response, full briefing, and a NetChoice amicus, and was denied on May 26, 2026.³ A denial is not an endorsement, but it leaves Briskin undisturbed, and no petition from Briskin itself has surfaced.

Meanwhile the trend line runs the other way. The First Circuit in Stokinger v. Armslist and the Nevada Supreme Court, in an unpublished order in Snap v. Eighth Judicial District Court, reached functionally parallel results in 2026 without citing Briskin, which suggests convergence; no other circuit has squarely adopted or rejected Briskin’s rule, and no clean circuit split has formed.¹⁶ What would actually move the doctrine is specifiable: a future cert grant on a thin-contacts vehicle, or a panel decision reading the regular-business clause as a required element. Watch those two events; nothing else on the horizon changes the analysis. Judge Callahan’s Walden-based critique is the intellectual seed of any future reversal, and defense counsel will keep planting it.¹²

The Shield Is Gone; the Limits Are Not

The live questions are about edges, not existence: watch for a thin-contacts vehicle and for how panels read the regular-business clause, because those are the two events that could move the doctrine. Until one does, cross-border jurisdictional exposure is a question worth answering deliberately, before someone else answers it for you in a complaint.

Related Resources

• When the SEC Won’t Act: Private Crypto Litigation
• DAO Liability After Lido: The Legal Wrapper Guide
• Smart Contract Enforceability: Code Isn’t Law
• Crypto Bankruptcy: Recovering Assets After FTX

This article provides general information for educational purposes only and does not constitute legal advice. Personal-jurisdiction doctrine is evolving rapidly. Consult qualified legal counsel for advice on your specific situation. Attorney Advertising.

Footnotes

1. Gelasio v. Zafar, No. 24-7277 (9th Cir. Apr. 29, 2026) (mem.) (unpublished memorandum disposition; reversing dismissal for lack of personal jurisdiction). The author briefed and argued the appeal for the plaintiff-appellant. ↩ ↩² ↩³ ↩⁴ ↩⁵

2. Briskin v. Shopify, Inc., 135 F.4th 739, 757-59 (9th Cir. 2025) (en banc) (overruling the “differential targeting” requirement at 757; holding at 758 that an interactive platform “expressly aims” its wrongful conduct at a forum state when its contacts are its “own choice and not ‘random, isolated, or fortuitous’”). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

3. Meta Platforms, Inc. v. Vermont, No. 25-909 (U.S.) (petition for certiorari filed Jan. 26, 2026; response requested Feb. 18, 2026; reply filed May 1, 2026; petition denied May 26, 2026). ↩ ↩² ↩³

4. Calder v. Jones, 465 U.S. 783 (1984) (effects test); see Schwarzenegger v. Fred Martin Motor Co., 374 F.3d 797 (9th Cir. 2004) (three-part specific-jurisdiction framework). ↩

5. AMA Multimedia, LLC v. Wanat, 970 F.3d 1201 (9th Cir. 2020); Will Co., Ltd. v. Lee, 47 F.4th 917 (9th Cir. 2022); Doe v. WebGroup Czech Republic, a.s., 93 F.4th 442 (9th Cir. 2024), overruled in relevant part by Briskin v. Shopify, Inc., 135 F.4th 739 (9th Cir. 2025) (en banc). ↩

6. Cal. Code Civ. Proc. § 410.10. Where no single state’s courts would have jurisdiction and the claim arises under federal law, Federal Rule of Civil Procedure 4(k)(2) supplies a federal long-arm analysis against foreign defendants. ↩

7. Schwarzenegger v. Fred Martin Motor Co., 374 F.3d 797, 802 (9th Cir. 2004); see Briskin, 135 F.4th at 761 (quoting Schwarzenegger on the defendant’s burden to present a compelling case); Roth v. Garcia Marquez, 942 F.2d 617, 625 (9th Cir. 1991). ↩

8. Calder v. Jones, 465 U.S. 783 (1984). ↩

9. Walden v. Fiore, 571 U.S. 277 (2014). ↩ ↩²

10. Daimler AG v. Bauman, 571 U.S. 117, 137-39 (2014). ↩

11. Briskin v. Shopify, Inc., No. 21-cv-06269-PJH, 2026 WL 161441 (N.D. Cal. Jan. 21, 2026) (on remand, dismissing all claims as pleaded under Rule 12(b)(6), all but one with leave to amend, while rejecting several other defense grounds). ↩

12. Briskin, 135 F.4th 739, 773 (Callahan, J., dissenting) (warning that keying jurisdiction to the plaintiff’s location risks conflict with Walden v. Fiore, 571 U.S. 277 (2014)). ↩ ↩²

13. Herbal Brands, Inc. v. Photoplaza, Inc., 72 F.4th 1085 (9th Cir. 2023) (seven-factor reasonableness test); Roth v. Garcia Marquez, 942 F.2d 617 (9th Cir. 1991). ↩

14. 9th Cir. R. 36-3 (unpublished dispositions are not precedent, citable per Fed. R. App. P. 32.1). ↩

15. Butcher’s Union Local No. 498 v. SDC Investment, Inc., 788 F.2d 535, 540 (9th Cir. 1986) (quoting Data Disc, Inc. v. Systems Technology Associates, Inc., 557 F.2d 1280, 1285 n.1 (9th Cir. 1977)) (discovery “should ordinarily be granted where ‘pertinent facts bearing on the question of jurisdiction are controverted or where a more satisfactory showing of the facts is necessary’”); cf. Pebble Beach Co. v. Caddy, 453 F.3d 1151, 1160 (9th Cir. 2006) (discovery properly denied where the jurisdictional claim is attenuated and rests on bare allegations met by specific denials). ↩

16. Stokinger v. Armslist, LLC, 166 F.4th 229, 240-41 (1st Cir. 2026) (vacating in part a jurisdictional dismissal; prima facie purposeful availment shown by the platform’s own forum-directed conduct, with relatedness left open on remand); Snap, Inc. v. Eighth Judicial District Court, No. 90276, 2026 Nev. Unpub. LEXIS 152, 2026 WL 501564 (Nev. Feb. 23, 2026) (unpublished disposition; cited per Nev. R. App. P. 36(c)(3)) (sustaining specific jurisdiction over Snap on its Nevada user contracts, data collection, and forum-targeted advertising). Both are cited for the parallel-result trend, not for any quoted language. ↩