SAP Chief Quantum Officer

Qubits are set to be the next ERP revolution

The German-speaking Handelsblatt reported on SAP's plans for quantum computing. Apart from the general euphoria, any serious consideration should assume that quantum computing is to classical computer science as a frozen pizza is to a three-Michelin-star restaurant.

The super-rich IT companies are currently experimenting with quantum superiority, which means that there is a tiny hope that quantum algorithms could one day be superior to classic abap algorithms. Why SAP has taken Formula 1 as a model, where every part is developed from scratch, remains a mystery even after reading the Handelsblatt text.

In many successful motorsport scenarios, you start with production vehicles that are modified accordingly. Why didn't SAP buy a few quantum computing simulators from Atos or Fujitsu and try to gain relevant qubit experience with simple sorting algorithms (second semester computer science)? In the past, SAP boss Christian Klein refused to develop a generic LLM (Large Language Model) for ERP end-to-end processes, i.e. a business ERP LLM. Now it is to be the much more expensive and sophisticated quantum computing. Failure is immanent.

Traveling Salesman Problem

Transforming algorithms and data structures from the ERP world to a quantum computer is not only a Herculean task, but also completely pointless due to the technical possibilities. The German-language Handelsblatt reports favorably on SAP's plans, which leads to confusing statements. The text gives the impression that it is only a matter of time before existing SAP customers can solve their traveling salesman problem (TSP) with the help of a quantum ERP.

The traveling salesman problem is a classic NP optimization task that involves finding the shortest route for a traveling salesman. The traveler should visit the locations exactly once and then return to the starting point. Due to its computational complexity, it is solved approximately in classic ERP practice for logistics and production with the help of approximation algorithms, heuristic methods or dynamic programming. NP problems (non-deterministic polynomial time) are

Computational problems whose solutions are very difficult to find, but can be checked quickly (in polynomial time).

Going it alone or partnerships

SAP has been conducting research in the field of quantum computing for several years, including in cooperation with: IBM (IBM Quantum), Google (Quantum AI) and various universities. It is emphasized that the goal is not to replace classic ERP systems, but to integrate quantum algorithms as optimization modules into existing SAP landscapes. SAP would continue to run on traditional servers (Intel processors and Linux) - but individual complex optimization problems could be outsourced to a quantum processor (cloud-based).

ERP systems such as S/4 serve production planning, supply chain optimization, inventory management, financial and risk modelling and transport and route optimization. Many of these challenges belong to the class of combinatorial optimization problems - precisely the types of problems for which quantum algorithms could theoretically provide advantages. These problems grow exponentially with complexity. Quantum algorithms such as QAOA could theoretically find good solutions more quickly.

The SAP Quantum Abap Framework

Where SAP could provide added value would be in the development of a framework for quantum-based business algorithms. The goal should be to develop a quantum abap. Even if quantum hardware is moderately successful in the coming years, the development of quantum algorithms will remain the far greater task for many years to come.

Just as the SAP founders provided a valid basis for the development of standard business software with the Abap programming language, future ERP quantum success can probably only be achieved with a suitable framework for the creation of ERP algorithms for different quantum hardware. It remains to be seen whether SAP will manage to set a similar standard in the quantum area as it did with Abap in classical IT.

Cobbler, stick to your last! Even if the idea of a quantum computer for business processes may be tempting, SAP managers should keep their feet on the ground: SAP did not succeed in developing cloud computing independently and autonomously (S/4 Hana was originally planned as an on-prem system). The failure with AI is obvious everywhere: the collaboration with Aleph Alpha was a flop and the ERP-LLM RPT-1 is equipped with far too few resources, see also Jonas Andrulis and Roland Berger. But now SAP wants to convince with quantum computing - a technology that is thousands of times more complex and expensive than cloud and AI. Failure is immanent!

SAP was unable to significantly develop cloud computing, failed to build an LLM (Large Language Model) for end-to-end business processes, falls far short of the possibilities of universal agentic AI with Joule, and now wants to reclaim a lost unique selling point with quantum computing.