In a significant development for cryptographic research, a new academic paper has introduced an improved algorithm designed to solve the superspecial isogeny problem in dimension 2. Authored by Maria Corte-Real Santos, affiliated with the École Normale Supérieure de Lyon and the French National Centre for Scientific Research (CNRS), the research outlines a sophisticated approach that achieves substantial performance enhancements over previous methodologies. The findings were published in the Cryptology ePrint Archive as Paper 2026/2808, drawing immediate attention from researchers and cryptographers studying post-quantum security frameworks and abelian surfaces.
The core contribution of the research centers on the efficiency of solving instances of the superspecial isogeny problem in dimension 2. According to the study, the newly developed algorithm requires a one-time precomputation phase that depends exclusively on the prime $p$. This initial phase succeeds with overwhelming probability, establishing a robust foundation for subsequent computations. Following this precomputation, each individual instance can be solved in an expected time and memory complexity of $p^3/4+o(1)$. Crucially, the authors note that this primary result is entirely unconditional, meaning it relies on no heuristics to achieve its performance guarantees.
When factoring in the cost of the initial precomputation, a single instance of the problem is solved in a time complexity of $p^19/23+o(1)$. This represents a notable improvement over the established Costello-Smith algorithm, which operates at a time complexity of $p^1+o(1)$. Furthermore, the author emphasizes that this performance enhancement holds true even when operating under the exact same memory constraints as the rigorous version of the Costello-Smith algorithm, demonstrating a pure algorithmic efficiency gain rather than merely a hardware-based trade-off.
The foundational engine behind this performance leap involves a specialized algorithm designed to compute the explicit equations of the Humbert surfaces, specifically denoted as $mathcalH_N^2$ modulo $p$, for all integer values of $N$ up to a dynamic bound $D$ that grows proportionally with the prime $p$. These computed equations provide a highly efficient membership test for optimally $(N,N)$-split surfaces. By evaluating these surfaces in batches through the application of fast multipoint evaluation techniques, the research successfully accelerates what has historically been the primary computational bottleneck in the Costello-Smith algorithm.
This specific bottleneck involves the complex task of finding an isogeny that maps a given superspecial principally polarised abelian surface directly to a product of supersingular elliptic curves. By streamlining this transformation via fast multipoint evaluation and advanced geometric insights into Humbert surfaces, the new algorithm bypasses the traditional computational walls that previously limited performance scaling in higher-dimensional isogeny-based cryptography.
The implications of this research extend across theoretical mathematics and advanced cryptographic protocol design. Isogeny-based cryptography has emerged as a major pillar of post-quantum cryptography, relying on the hardness of finding maps between elliptic curves or higher-dimensional abelian varieties. Enhancements in solving isogeny problems in dimension 2—such as the abelian surfaces explored in this work—provide deeper analytical tools for evaluating the security parameters of cryptographic schemes that depend on supersingular or superspecial abelian varieties.
The publication of Paper 2026/2808 adds to the ongoing global discourse on optimizing cryptographic primitives for future computing landscapes. Researchers in the cryptographic community are expected to analyze the unconditional nature of the new algorithm and its reliance on the precomputed Humbert surface equations to explore further optimizations in higher-dimensional settings. As cryptographic standards continue to evolve in anticipation of quantum computing developments, theoretical breakthroughs of this nature offer critical insights into the mathematical boundaries governing isogeny graphs and their structural properties.

