Security for Costs at the UPC: Who Gets Ordered to Pay, and How Much

Across the Unified Patent Court cases our platform has classified, 49 carry a quantified security-for-costs order — a specific sum the court told one side to put up to cover the other side's potential cost award. Those orders run from €10,000 to €3,000,000, with a median of roughly €200,000. That range is the single most useful number a litigation underwriter or funder can know before backing a UPC case, because security for costs is collateral they may have to size in cash — sometimes within four weeks of an adverse order.

A note on precision up front, because the gap matters. Around 147 cases in our corpus reference security for costs at all — citing Rule 158 of the Rules of Procedure, Article 69(4) UPCA, or the German-law shorthand Prozesskostensicherheit somewhere in the record. But referencing the mechanism is not the same as being hit by it. Only 49 carry a quantified order — an actual figure the court fixed. The rest are applications refused, withdrawn, settled, or resolved on principle without a number ever being set. So when you read "the UPC orders security in X cases," check whether the writer means applications filed or orders with a euro figure attached. We mean the latter.

What "UPC security for costs" actually is — the Rule 158 mechanism

Security for costs is a defendant's insurance against an empty-handed opponent. The legal basis is Article 69(4) UPCA, implemented by Rule 158 of the Rules of Procedure: at the request of the defendant, the Court may order the claimant to provide adequate security for the legal costs and other expenses the defendant incurs and which the claimant may later be ordered to bear. The security is lodged as a deposit of money or a bank guarantee. If the claimant ultimately loses and is ordered to pay costs but cannot or will not, the defendant draws on the security instead of chasing an unenforceable judgment across borders.

Three features of the mechanism drive everything in our data:

• It is the defendant's tool. In June 2025 the Court of Appeal confirmed that, because Article 69(4) deliberately limits the right to the defendant, Rule 158 cannot extend it to claimants — and a defendant who files a revocation counterclaim does not thereby become a "claimant" who can be ordered to post security (Bird & Bird; EPLAW). That single ruling reshapes who can even ask.
• The test is enforceability, not merits. The court does not weigh who is likely to win on infringement or validity. It asks whether the claimant's financial position gives rise to "a legitimate and real concern that a possible order for costs may not be recoverable and/or … not, or only in an unduly burdensome way, be enforceable" (Taylor Wessing). The burden sits on the defendant to substantiate that concern.
• Cross-border enforcement is the usual trigger. A claimant with no assets inside the EU/EEA — or one based where a UPC cost order is slow or costly to enforce, such as the US or China — is the classic candidate. Divisions have ordered security where enforcement against a Chinese or US claimant was shown to be unduly burdensome, while refusing it where the defendant simply asserted difficulty without proving the foreign law (Taylor Wessing; Pinsent Masons).

That is why this topic is a wedge for insurers and funders rather than a procedural footnote. The whole point of the order is to neutralise an under-capitalised or offshore claimant — which describes a large share of the entities that funders and after-the-event insurers back.

Who gets ordered to post — claimants, frequently NPEs

The posting party in our 49 orders breaks down sharply: claimants in 27, respondents in 10, and unspecified in our record for the remaining 12. The claimant skew is exactly what the mechanism predicts: where a posting party is identified, claimants are ordered to post security nearly three times as often as respondents. And the claimants who draw these orders are disproportionately non-practising entities — patent-holding vehicles whose only material asset is the patent in suit, often domiciled outside the EU. As Taylor Wessing puts it, an NPE "may be ordered to provide security for costs, especially if their only assets are the patents in question." A US-based LLC with no tangible assets and cost recovery in Texas described as "unduly burdensome" is, in our records, a textbook fact pattern for an order.

The 10 orders against respondents are the smaller, more situational bucket — typically security tied to a specific procedural step (for example, attached to a provisional-measures order under Rule 196.6 RoP) rather than the headline Article 69(4) "post security for my costs" application. The signal is clean: if you fund or insure a claimant without local assets, price in the real chance of a Rule 158 order; if you act for one, assume the application is coming and prepare the financials in advance.

Where these fights happen — and why the Court of Appeal leads

The divisional distribution of the 49 quantified orders is rarely published anywhere — and it carries a strategic surprise.

The headline is that the Court of Appeal in Luxembourg accounts for the single largest share — 18 of the 49 — not as a first-instance forum, but because security disputes are fought on appeal more than almost any other procedural question. A first-instance order to post €600,000, or a refusal to order anything, is exactly the kind of ruling each side will take up: the amount is concrete, the cash consequences are immediate, and the governing principles (who can apply, what the test is, what counts as security) were still being settled across 2024–2025. Among the first-instance divisions, Düsseldorf leads at 9, with Hamburg and Munich tied at 4 — broadly tracking those divisions' overall caseload weight.

For a litigant, the practical reading is: an unfavourable security ruling is rarely the end of the matter. For a funder, it means the collateral question can stay live — and the number can move — well after the first-instance order.

How much — the range underwriters must size

Back to the amounts, because this is where collateral gets priced. Across the 49 orders:

• Minimum: €10,000. A nominal figure, usually a procedural step rather than a full costs-shield.
• Median: roughly €200,000. The realistic centre of gravity. Düsseldorf's judge-rapporteur ordered €200,000 against AorticLab on the strength of the claimant's own admission of financial instability in its Defence — despite that party having raised USD 10.5m (Pinsent Masons). Reported sister orders cluster nearby: around €300,000 (Microsoft v Suinno) and €400,000 (ICPillar v ARM) (Taylor Wessing).
• Maximum: €3,000,000. The outlier, reserved for the largest disputes. Amounts are typically anchored to the value of the proceedings and the recoverable-cost ceiling — in a dispute valued at €4–8m, the recoverable ceiling is around €600,000, and divisions have set security at roughly 50% of that ceiling, payable into the UPC's dedicated security-deposit account within about a month.

The distribution is the takeaway. A funder sizing exposure on a typical UPC matter should budget for a mid-six-figure security demand as the base case, treat sub-€100,000 outcomes as the lighter procedural tail, and reserve seven-figure provisioning only for the highest-value disputes. Because the order can land early and demands cash or a bank guarantee on a tight clock, it is a liquidity problem, not just a balance-sheet one.

What this means for funders and insurers

The open question underwriters should watch most closely is what form the security must take. The Rules contemplate security provided by deposit or by bank guarantee, but they do not settle whether a robust after-the-event (ATE) litigation-insurance policy can stand in for cash. As more funded and insured claimants — many of them NPEs without EU assets — appear at the UPC, that question is moving to the centre of the security debate, and it is being pressed on appeal. The answer will decide whether a security order is a genuine liquidity wall or, for a well-insured claimant, largely a drafting exercise.

If the divisions do come to accept insurance as adequate security, the commentary already points to the terms that would matter: an EU-regulated (Solvency II) insurer, an anti-avoidance endorsement making the policy non-voidable and non-cancellable, a direct route for the defendant to claim against it, a coverage limit adequate to the defendant's recoverable costs, and a termination moratorium giving the defendant time to seek alternative security. Get those terms right and a well-drafted ATE policy could one day substitute for a multi-million-euro cash deposit; leave them loose and a court will keep demanding cash.

For the insurer/funder, three things follow. There is a market opening — a properly structured ATE product could become the recognised answer to the biggest cash drag on funding a UPC claimant, if the case law allows it. Policy drafting is the whole game — the non-avoidance, direct-claim, and coverage-adequacy terms are what would convert a policy from "considered" into "sufficient." And the base case still assumes a deposit — until the appellate position is settled, the order means roughly €200,000 you must be able to produce on demand.

What to watch

• Will insurance count as security? Watch whether the Court of Appeal and the first-instance divisions come to accept a strong ATE policy in place of cash — and, if they do, whether they settle on a checklist of policy terms. A clear standard would let funders standardise an insurable product; continued case-by-case treatment keeps cash deposits the safe default. We track where this lands in our Syntorr v. Arthrex analysis.
• The non-EU enforcement question. With UK long-arm rulings and the post-BSH Hausgeräte expansion drawing more non-EU claimants into the UPC, expect more security applications premised on enforcement difficulty in the US and China — and more appellate fights over what evidence of foreign law a defendant must put in.
• Who can apply, post-counterclaim. The Court of Appeal has shut the door on ordering revocation counterclaimants to post security. Watch for attempts to reopen it on different procedural facts — the boundary is not fully mapped.
• Median drift. As the case law matures and ATE coverage spreads, the realistic question is whether the ~€200,000 median holds, drifts up with case values, or splits into a two-tier pattern — light procedural security versus heavy enforcement-risk orders.

For funders and insurers, security for costs has gone from a procedural afterthought to a sizeable line item — and possibly, soon, an insurable one. The number to carry into every UPC underwriting conversation is the one our data fixes: about €200,000 at the median, €10,000 to €3,000,000 across the range, and a claimant skew that puts NPEs squarely in the frame.

Want the security-for-costs orders behind these figures — by division, amount, and posting party? Explore the platform to see the case-level data.

Related reading

• Litigation Insurance as Security: The Syntorr v. Arthrex Turn
• The 20 Most Active Litigants at the UPC
• The UPC at Three Years: Outcome Base Rates Every Litigant Should Know

Sources

• UPClytics.com — internal dataset of classified UPC decisions (snapshot June 2026 build).
• Rule 158 RoP — Security for costs of a party (IPPT / IP-PorTal)
• Bird & Bird — UPC Court of Appeal rules only defendants can request security for costs
• Taylor Wessing — Security for costs before the Unified Patent Court: current case law and strategic implications
• Pinsent Masons (Out-Law) — Security for costs
• EPLAW — UPC: security for costs in counterclaims, insights from a recent UPC case