Aleph Alpha

Aleph Alpha is a German artificial intelligence (AI) startup company founded by professionals with experience as employees of Apple, SAP and Deloitte, based in Heidelberg. Aleph Alpha attempts to provide a full-stack sovereign technology stack for generative AI, independent from US companies and comply with European data protection regulations and the AI Act. It built one of the most powerful AI clusters inside its own data centre and develops large language models (LLM), which try to provide transparency of its sources used for the results generated and are intended for enterprises and governmental agencies only. Training of its model has been done in five European languages.

History
Aleph Alpha has been founded 2019 by the serial entrepreneur Jonas Andrulis and Samuel Weinbach. Jonas Andrulis received his academic economics engineering degree from Karlsruhe Institute of Technology with a thesis in AI. He worked in consulting at Deloitte and started several AI software companies. Before starting Aleph Alpha he worked as AI R&D engineering manager position at Apple's Special Projects Group and with Siri AI R&D working on classified research. Samuel Weinbach received a degree in business administration and was with Deloitte consulting from 2010 building the Deloitte Analytic Institute that was focused on spearheading corporate AI initiatives.

Funding

 * After receiving 5 million in seed money in 2020, a second round of financing by several European venture-capital companies totaled 23 million Euro in 2021.
 * In a financing round in November 2023, German companies Schwarz Gruppe and Dieter Schwarz Foundation with the Innovation Park Artificial Intelligence (IPAI) participated as the co-lead investors, along with Bosch together with SAP, Hubert Burda Media and Christ&Company Consulting as well as US-Hewlett Packard Enterprise. The total amount of this third round was more than 500 million US dollars.

Luminous
Aleph Alpha developed its own AI language model, named Luminous, based on its own research and codebase with the architecture of generative pre-trained transformers (GPT) with self-supervised learning. Next to the standard functionality all GPT models share Aleph Alpha contributed some proprietary innovation:


 * They were the first team to offer multimodality in 2021 - the ability to prompt their models with any combination of text and images. The innovation behind this capability was published at the leading AI conference EMNLP 2022 and open-sourced under the name MAGMA.
 * In 2022 they were the first team to develope the ability to create images based on multimodal input. This method has been published at NeurIPS 2023 under the name Multifusion.
 * With an innovation called AtMan published at NeurIPS 2023 Aleph Alpha developed a method to make the patterns learned by GPT models to create results visible and controllable. This methods addressed the problem that generative AI had been a black box before. It created interest from enterprises and governments to address complex and critical problems where chatbots are sometimes insufficient, making AI generated results more trustworthy.

As a tool to build and train its foundation models, the HPE Machine Learning Development System is used. Using the GPT-type concept allows adaptation and fine-tuning of the foundation model to various applications.

Luminous is already used for the citizen information system Lumi of the city of Heidelberg.

Partnerships

 * For R&D, Aleph Alpha is working with the University of Duisburg-Essen and the Technical University of Darmstadt. It is also participating in open-source organizations such as EleutherAI, a collective of researchers working to increase accessibility to AI research.
 * Both Hewlett Packard Enterprise (HPE) and SAP have entered non-exclusive partnerships with Aleph Alpha. In the case of HPE, it is the first partnership of this kind. HPE will offer Luminous on its Greenlake platform.