Deutsche Bank & NVIDIA AI collaboration is a promising leap forward for Financial Services
This week`s announcement: Banking on AI: Deutsche Bank, NVIDIA to Accelerate Adoption of AI for Financial Services
The 3rd phase of digital transformation in financial services includes incumbents leveraging AI beyond the conventional, siloed ways that they are already deploying.
We live in a world that is being shaped by collaborative technologies, be it AI, blockchain, or IoT. Legacy technology systems are not suitable for this world that insists, requires, and demands the seamless collaboration of technologists and business people.
When I worked at Bankers Trust in Manhattan at 130 Liberty Trust Plaza next to the World Trade Center, IT and Sales & Trading were so separate, so siloed that I never met a person from IT in the 4 years that I worked there. I was involved in re-building models for Latin American fixed-income structured products just after the 1994 Mexican currency crisis, calibrating models for risk management of the restructured sovereign debt of LATAM countries, and later running a local currency-denominated loan book. Deutsche Bank (DB) acquired Bankers Trust in 1999, a year after I left.
In 2021, nearly 50% of DB’s revenues (c. €12.5B) were from investment banking & asset management. If the traditional silos between DB’s business divisions and its business divisions and IT persist, then its risk management processes will become a dangerously archaic workflow that nobody can extract value from.
We live in a world where a Trillion parameter model for sensitivity analysis in financial services is just around the corner. Gil Perez, the chief innovation officer of DB, told me that DB is already running models for sensitivity analysis with Billions of parameters. According to NVIDIA, in total in 2019 NVIDIA`s models had 1 Billion parameters and in 2022 the models were training 175 Billion parameters. Undoubtedly, High Performance Computing (HPC) and AI (Machine Learning specifically) are converging. Accelerated compute, meaning GPUs and HPC software development kits plus refined models optimized for performance and efficiency by NVIDIA and the likes, are needed to run such large models. The expectation is 1 Trillion parameters by 2025!
This week’s announcement of an innovation partnership between DB and NVIDIA on AI is a major move in the journey of digitalization in financial services. An incumbent of the size and breadth of DB is more than ready to embark on a marathon that will change the business from inside out. Accelerated by the partnership, DB will advance their AI Center of Excellence to support the experimentation and development of AI and ML services and professional skills development. They have already tested concepts with NVIDIA over the past months and are planning to strengthen the collaboration further in 2023 and beyond as part of the multi-year agreement.
DB’s AI experts work with colleagues across the bank to drive the adoption of AI across the organization and they believe the Center of Excellence model provides both the central support and the extensive business representation to support this cultural change in the bank.
Needless to say that this journey is happening in parallel with DB’s shift to the public cloud that started in 2019–2020 and continues with the ultimate goal being “Real-time compliant access to data”.
The NVIDIA & DB collaboration will hugely benefit the entire financial services sector. Manuvir Das, vice president of Enterprise Computing at NVIDIA, emphasized that the combined improvements in hardware and software we can expect going forward are still huge. He said 1 million times improvement is expected. To me, this opens up huge opportunities for financial services providers that are serious about transforming into data businesses to leverage internal data and external data from their value chain and from alternative data providers.
The first North Star of the DB NVIDIA collaboration is to unlock customer value hidden in their data, beginning with next-generation risk management. This means real-time risk management, real-time valuation, increased alpha signal discovery, improved price discovery, better cyber security, better ESG analytics, and so on.
This week’s announcement of the DB NVIDIA collaboration was combined with the announcement of NVIDIA AI Enterprise 3.0 software suite. This includes improvements in their software in the form of SDKs in all the areas of their focus, vision AI, speech AI, cyber AI, document recognition AI and more. NVIDIA is also adding more workflows for specific use cases.
The second area of DB’s focus is deriving insights out of unstructured data. Starting from the increasing amount of financial texts (regulations, policies, reports, news, etc.) that have been challenging for existing large language models. Deutsche Bank and NVIDIA have started testing a collection of new large language models called Financial Transformers (Finformers). These AI and ML models are looking to identify early warning signs of counterparties, to detect data-quality issues, and to retrieve faster data.
The third area of the DB & NVIDIA collaboration is centered around NVIDIA Omniverse. DB strongly believes that avatars (be it digital twins of a human or not) will be part of the future immersive 3D world. They have already started developing an early concept of a 3D virtual avatar aimed at helping employees to navigate internal systems and respond to HR-related questions. This way DB will learn a lot about the use of this technology without worrying about compliance and regulators interfering and the learnings will be used when it is time to create immersive experiences with banking clients. NLP models trained for intelligent Avatar applications can be directly transferred to chatbot applications.
Avatars, Finformers, GPUs for new machine learning models, workflows and SDKs for specific financial use cases, are highlights from this landmark collaboration. NVIDIA has traditionally worked with clients to co-create their hardware and software for specific use cases. Now is the time for financial services to take a leap. And the excitement just starts as there will be more leaps to come.
I am excited about the DB and NVIDIA collaboration and the forthcoming benefits for the entire financial industry from this collaboration.
I am an NVIDIA ambassador and a Global Fintech Thought leader. This article expresses my own opinions.
📌 Twitter: https://twitter.com/efipm
📌 Subscribe to my YouTube Channel with my insights and industry leader interviews. New video every Wednesday: https://www.youtube.com/EfiPylarinou
📌 Spotify Podcasts. Follow here: https://open.spotify.com/show/5bRkZEYHSwPiGx7vTqylw6?si=Mg3hN5PDQ86K10GjeK52jw
📌 Linkedin: https://www.linkedin.com/in/efipylarinou/
📌 TikTok: https://www.tiktok.com/@efiglobal
📌 Web: https://efipylarinou.com/