An Unwarranted Free Pass for Extra-EU ISDS?

On 31 July 2025, the German Constitutional Court (Bundesverfassungsgericht or ‘BVerfG’) delivered a decision that revived interest in the German Federal Court of Justice’s (Bundesgerichtshof or ‘BGH’) ruling of 12 October 2023 on the compatibility of investor-State dispute settlement (‘ISDS’) in bilateral investment treaties concluded between an EU Member State and a third country (‘extra-EU BITs’) with EU law. In particular, the BVerfG shed light on a still under-explored facet of the BGH’s judgment, namely its refusal to apply the European Court of Justice (‘ECJ’)’s Opinion 1/17 to extra-EU BITs – which may carry far‑reaching consequences for the future relationship between EU law and international investment law.

As a brief reminder, in its Achmea (2018) and Komstroy (2021) rulings, the ECJ held that intra-EU ISDS is incompatible with EU law in light of the principles of mutual trust and autonomy of the EU legal order. In Opinion 1/17 (2019) on the compatibility of the ISDS mechanism envisaged in CETA with EU law, the ECJ outlined the application of the autonomy principle in extra-EU configurations involving a treaty concluded by the EU.

In 2023, the question whether Achmea/Komstroy could be transposed wholesale to ISDS disputes under extra-EU BITs was raised by India before the BGH, in the context of enforcement proceedings of the Deutsche Telekom v India award rendered under the Germany-India BIT. The BGH’s answer was unequivocal: the prohibition affirmed in Achmea does not extend to extra-EU ISDS. Most commentators focused on that finding (see here and here).

However, the BGH did not merely confirm that EU law does not prohibit extra-EU ISDS; it went so far as to dismiss any autonomy concerns regarding extra-EU ISDS clauses. Since Opinion 1/17, the ECJ has not had the opportunity to rule expressly on how these principles would operate in the context of extra-EU BITs concluded by Member States. Yet, remarkably, the BGH decided to settle this most contested question on its own. The BGH dismissed the relevance of Opinion 1/17 for extra-EU BITs and declined to refer the matter to the ECJ, based – in our view – on an erroneous reading of the ECJ’s Komstroy ruling.

The BVerfG’s 2025 decision, which rejected India’s challenge against the BGH’s failure to refer to the ECJ, offers an opportunity to take a fresh look at the BGH ruling.

The BGH’s erroneous interpretation of Komstroy 

In its ruling, the BGH appears to have conflated two separate legal questions, namely whether, in light of ECJ case law, (i) EU law per se precludes extra-EU ISDS, and (ii) ISDS clauses contained in certain extra-EU BITs can be contrary to EU law and in particular the principle of autonomy.

The BGH simultaneously addressed both questions without distinguishing them, confining itself to the conclusion that Achmea is “not transferrable” to extra-EU BITs and that “arbitration clauses in bilateral investment protection agreements between Member States of the European Union and third countries do not conflict with EU law” (para. 27).

Admittedly, it cannot be inferred from the ECJ’s case law that Achmea extends to extra-EU ISDS disputes. However, the inapplicability of the Achmea decision does not dispose of the autonomy question altogether. The BGH entirely disregarded this dimension, going so far as to read paragraph 65 of Komstroy as a general, non-rebuttable clearance for extra-EU ISDS clauses under EU law. That interpretation is, in our view, fundamentally flawed. Paragraph 65 reads:

“[…] although the ECT may require Member States to comply with the arbitral mechanisms for which it provides in their relations with investors from third States who are also Contracting Parties to that treaty as regards investments made by the latter in those Member States, preservation of the autonomy and of the particular nature of EU law precludes the same obligations under the ECT from being imposed on Member States as between themselves.”

That paragraph does no more than to confirm the absence of a general prohibition of ISDS clauses in extra-EU relations, in contrast to intra-EU settings (see also here, p. 27, and here). It does not follow that such clauses are immune from scrutiny, nor that they cannot infringe EU law – including the principle of autonomy – on a case-by-case basis.

Refusal to safeguard the autonomy of the EU legal order

In Opinion 1/17, the ECJ provided a clear analytical framework to assess whether ISDS clauses in agreements concluded by the EU with third States comply with EU law and the principle of autonomy (para. 119). That framework requires a contextual assessment of the specific ISDS mechanism at issue. The key is that the agreement includes safeguards preventing arbitral tribunals from interpreting and applying EU law (paras 120-136) and from interfering with the level of public interest protection sought by the EU (general exceptions, right to regulate, limitations to investment protection standards) (paras 137-161).

The BGH refused to apply that framework to extra-EU BITs. It argued that, had the ECJ intended Opinion 1/17 to apply in that context, it would have said so explicitly in paragraph 65 of Komstroy. It is difficult to see how a passing reference, buried in an obiter dictum, could have been intended to resolve a question of such systemic importance, which could potentially affect hundreds of BITs concluded by Member States.

The BGH further held that Opinion 1/17 was inapplicable because CETA was concluded by the EU itself and not by Member States. However, this argument rests on a misconstruction of paragraph 127 of Opinion 1/17, which the BGH read as drawing a distinction between agreements concluded by Member States and those concluded by the EU. No such distinction can be derived from the opinion. Paragraph 127 instead differentiates between “an agreement between Member States” and an agreement such as CETA which is concluded between the EU and a third country, on the ground that the mutual trust principle “is not applicable” between the Union and third States (para. 129). The ECJ merely intended to distinguish intra-EU situations – addressed in Achmea – from extra-EU situations. If anything, paragraph 127 confirms that a different analytical framework applies to extra-EU agreements, and that such treaties are not released from the obligation to respect EU law autonomy. This understanding is also consistent with the position adopted by Advocate General Bot in his Opinion under Opinion 1/17, where he expressly emphasised that requirement (para. 113).

It is also settled ECJ case law that any international agreement, whether concluded by the EU itself or by its Member States, cannot affect the autonomy of the EU legal system (Achmea, para. 32; Opinion 2/13, para. 201). In fact, in 2009 the ECJ had already stated that pre-accession BITs concluded by Member States must not prevent them from exercising their rights and fulfilling their obligations as EU members.

In addition, according to ECJ standards, the same autonomy concerns may arise regardless of whether the agreement is concluded by the EU or a Member State – as already noted by several commentators (see also here, here and here). This is so, in particular, where arbitral tribunals are called upon to hear challenges against domestic measures implementing EU obligations (Opinion 1/17, para. 150). Many extra-EU BITs indeed lack the procedural safeguards identified in Opinion 1/17.

Despite this, the BGH proceeded as though autonomy concerns simply vanish in extra-EU disputes. In this respect, the ECJ held in its 2024 ruling in the Micula saga that “it cannot be accepted that a national court or tribunal, still less so a court [of last instance], […] may adopt an erroneous interpretation of EU law” that leads “to exclude the application of EU law in its entirety” (paras 85-86). On that basis, it found that the UK Supreme Court had “seriously compromised the EU legal order”. The BGH’s ruling appears equally problematic, since its dubious reading of EU law effectively treats extra-EU BITs as immune from the requirement to respect EU law autonomy.

The BGH’s failure to refer to the ECJ

The BGH failed to refer the compatibility issue to the ECJ. In accordance with settled case law, the BGH is, as a court of last instance, required to refer questions concerning the interpretation of EU law to the ECJ pursuant to Article 267(3) TFEU. Exceptionally, a court is not obliged to refer when the question of EU law raised is irrelevant to the outcome of the case at hand; that question has already been resolved by the ECJ (acte éclairé); or the correct application of EU law leaves no scope for reasonable doubt (acte clair).

In the present case, the BGH justified its refusal to refer by relying on its reading of paragraph 65 of Komstroy. However, as stated above, that paragraph does not settle all autonomy concerns regarding ISDS clauses in extra-EU BITs. The issue whether ISDS clauses in certain BITs may impair the autonomy of EU law remains unresolved by the ECJ (see also van der Beck, Europäische Zeitschrift für Wirtschaftsrecht 2024, p. 252).

That question could have had a bearing on the case at hand, as required by Article 267(3) TFEU. Applying Opinion 1/17 would likely have led the BGH to find that the ISDS clause contained in the Germany-India BIT is incompatible with EU law: Article 13(1) of the BIT expressly provided for the application of domestic law, and the BIT included no CETA-like safeguards.

The BVerfG’s assessment

When reviewing the BGH’s failure to refer, the BVerfG engaged more substantively with autonomy concerns – albeit merely alluding to (rather than directly applying) the compatibility criteria laid out in Opinion 1/17.

According to the BVerfG, the constitutional complaint ought to have demonstrated a “genuine conflict” (“echter Konflikt”) with EU law in the dispute at hand (para. 39). The BVerfG added that no such conflict arose in the underlying dispute because it concerned an investment made in India and EU law was not applicable. This reasoning appears questionable. An ECJ référendaire recently criticised the BVerfG’s approach, stressing that according to the ECJ, the “mere possibility” of an encroachment on EU law autonomy is sufficient to make an external dispute resolution system incompatible with EU law (Kuplewatzky, Europäische Zeitschrift für Wirtschaftsrecht 2025, p. 1220). The fact that ECJ case law concerns the incompatibility of the ISDS clause itself – rather than the BIT’s substantive provisions – may also imply the existence of an “objective” incompatibility that arises irrespective of whether the application of EU law is concretely at issue in the underlying dispute. At the very least, this calls for clarification by the ECJ, as noted elsewhere.

The BVerfG further held that no real conflict could arise because EU law does not regulate the entirety of the relationships governed by extra-EU BITs. This seems inconsistent with ECJ case law indicating that, even where an agreement concluded by a Member State and a third country falls outside the scope of the Treaties, measures adopted pursuant to this agreement may impede the application of EU law.

Against this background, the view taken by the BHG and upheld by the BVerfG that no reference to the ECJ was warranted is difficult to sustain, as a preliminary ruling could have provided much-needed clarification on key questions of EU law.

Conclusion

Despite the absence of conclusive ECJ case law specifically addressing the implications of Opinion 1/17 for extra-EU BITs, the BGH viewed this question as settled based on an incorrect reading of Komstroy. The BGH’s general clearance of extra-EU ISDS clauses under the EU law autonomy principle – a finding the BVerfG largely left unchallenged – carries potentially far-reaching implications for enforcement proceedings across the EU, and more broadly, for the effectiveness of EU law. This stance may also reflect a reluctance on the part of the German courts to place themselves in conflict with Germany’s public international law obligations (Ebert, Recht der Internationalen Wirtschaft 2024, p. 193), and speaks to a broader tension between Member States’ international obligations and the principle of autonomy (see the AG’s opinion in C-641/18 and here).

That reluctance, however, is precarious. Should the ECJ formally acknowledge autonomy concerns in this context, the restraint shown by the BGH (and possibly by other national courts in EU Member States) would likely dissipate – with potentially significant consequences for the continued application of numerous old BITs within the EU legal order. In that respect, and in light of the Micula precedent, one can only speculate what the ECJ would make of the BGH’s decision, which insulates all extra-EU awards from judicial review on EU law autonomy grounds.