For the first time in over four decades of C++ standardization, an ISO national body ballot has come back disapproved. C++26 — carrying four of the most anticipated language upgrades in years — is now in limbo, its ratification blocked by a coordinated minority campaign over a single feature that has divided the committee since 2019: Contracts.
What’s Actually on the Line
C++26 is not a minor maintenance release. Herb Sutter, the ISO C++ committee chair, called it “the most compelling release since C++11” — and that assessment is well-supported. The standard packages four transformative features that have been years in development, now collectively stalled.
Reflection is the flagship. Using the new ^^ operator and std::meta::info type, C++ gains compile-time introspection with zero runtime overhead. No reflection metadata bloating your binary. Instead, the compiler uses it to generate specialized code and discards the scaffolding. Serialization libraries, logging frameworks, binding generators — entire categories of macro-dependent code become obsolete. GCC 16.1 already supports most of the feature experimentally behind -std=c++26.
std::execution brings a unified asynchronous and parallel execution framework to the standard library, built on Meta’s libunifex and NVIDIA’s stdexec. Memory safety improvements harden the standard library with bounds checking and eliminate undefined behavior from reading uninitialized locals. And then there’s Contracts — the one that broke everything.
The C++26 Contracts Dispute
Contracts add language-level preconditions, postconditions, and assertions to C++. The syntax introduces pre, post, and contract_assert keywords as structured alternatives to C’s blunt assert macro. In principle, it is a solid idea — design-by-contract has a strong track record in Ada and Eiffel.
In practice, contracts have been a C++ committee flashpoint for the better part of a decade. They were removed from C++20 in 2019 under nearly identical circumstances. This time, WG21 voted 100 to 14 to include them and 114 to 12 to send the full draft to ISO for ratification. C++ inventor Bjarne Stroustrup voted against, calling the feature “not minimal, not viable” and arguing it “changes the meaning of code depending on where it is.”
Opponents organized. More than 20 national body comments targeting contracts were filed during the DIS ballot period. A coordinated campaign persuaded enough national body representatives — who may lack the deep technical context of the WG21 working group — to vote No. The outcome: a majority of the world’s C++ practitioners, compiler vendors, and library authors were overruled by a process their technical work is supposed to govern.
How a Minority Defeats a Supermajority
The ISO standardization process divides authority between WG21 — the technical committee of working engineers — and national standards bodies, the ISO member organizations from each participating country. WG21 does the technical work. National bodies ratify it. They exist to catch issues WG21 missed, not to relitigate settled votes.
When a DIS ballot fails, ISO procedure requires WG21 to review the national body comments before recommending any course of action to the SC22 chair. Community discussions on Lobsters and the broader developer ecosystem suggest WG21 is leaning toward stripping contracts and resubmitting — an 8-week DIS ballot that could put C++26 in print by early 2027.
The alternatives are considerably worse. A full CD revision takes months and reopens the entire feature set to new comments. Technical Specification status would publish contracts as a separate optional document — effectively shelving them. Cancelling C++26 entirely would fold all four flagship features into C++29, which is years away.
Meanwhile, the language continues its competition with memory-safe alternatives that do not require navigating ISO national body politics to ship a feature.
Where This Leaves C++ Developers
The practical answer: start experimenting with reflection now. GCC 16.1 already supports most of it experimentally. MSVC does not yet have a public implementation, so Windows-primary developers will wait regardless of what ISO decides. The compile-time introspection model is genuinely transformative, and the standard — with or without contracts — will catch up to the implementations.
The harder answer: the ISO process has a documented vulnerability. A determined minority, working through national body channels rather than the technical committee, can override years of work and near-unanimous expert consensus. C++26’s delay is likely measured in months, not years, if WG21 acts quickly. However, the process failure is real — and it is the kind of structural weakness that accelerates conversations about whether C++ governance can keep pace with modern software development.
Contracts will resurface. They were ejected from C++20 and returned. If stripped from C++26, they will come back again — redesigned, reargued, re-voted. The question is whether the committee’s legitimacy can survive another cycle of the same fight.

