DeepSeek a game-changer for financial services AI

The arrival of DeepSeek’s R1 open-source model has rocked the AI industry – which has previously contended that large amounts of power and money would be needed to achieve results the challenger is offering at a fraction of the price. According to Chris Probert, global head of data & generative AI at Capco, the platform’s advent will also have major implications for financial services firms.

With the economies of the West stagnating, investment into AI had long been treated as a bubble that could provide a solution. States and capital pouring billions into infrastructure and technology to make the most of AI’s apparent potential was seen as a golden opportunity to stimulate growth across Europe and North America.

In particular, the idea hinged on the assertion that to create a powerful AI that could quickly analyse data to generate results, there would always be a need for bigger models, trained and run on bigger and even larger GPUs, based ever-larger and more data-hungry data centres. This would be incredibly expensive, meaning any AI boom would necessitate huge amounts of spending on every front.

But the launch of a Chinese artificial intelligence company called DeepSeek shattered those illusions in a chaotic week – leading to a market panic which wiped $590 billion off the value of chip-manufacturer NVIDIA. As trade barriers meant DeepSeek could only be developed on less powerful chips, the fact that it is reportedly as effective as ChatGPT, while being open source and 30 times cheaper to run, means many investors are suddenly worried about how much of a return they may ever see on their investments.

Beyond this chaos, however, Capco expert Chris Probert believes that there is a real opportunity for businesses to avail themselves of. According to the Capco partner, the launch of DeepSeek R1 both underlines how AI innovation is still accelerating, but also shows “that smaller language models would be a compelling option” for addressing an organisation’s problem statements – especially in the lucrative financial services sector.

 The Capco head of data and generative AI expanded, “For example, smaller models give firms the opportunity to leverage and curate their own training datasets, due to the lower data requirements needed to train smaller language models.”

The open source nature of the technology, and its ability to be run on relatively modest in-house hardware also means organisations could use their own training data – rather than relying on “hyperscaler datasets”. This could enable several key benefits: helping financial services firms to develop more fine-tuned and relevant models; reducing concerns about data security and privacy, where organisations no longer need to leverage hyperscaler models that operate in the cloud and can control where data is stored and how it is used; driving greater opportunities for competitive advantage and differentiation, and increasing “AI transparency and explainability”, giving firms greater visibility of how a model generates a specific output.

“This will be critical in a highly regulated industry such as financial services,” Probert explained. “There has already been plenty of discussion around the benefits of building AI capability in an agnostic way – that is, avoiding vendor lock-in to ensure firms have sufficient flexibility to adapt to market changes and benefit from ongoing AI innovation. The R1 model underlines the importance of this agnostic perspective.”

Looking ahead, he contended that firms should focus on “scalable enterprise solutions that allow easy model swaps, providing flexibility while also minimising transition costs”. Rather than hefty partnerships with particular firms, they might instead now embed “modularity and interoperability into the solution architecture early on”, and “future-proof their AI investments against rapid advancements in the field”.

He concluded, “To build for adaptability and high pace of change you can't be locked into a single vendor, and a ‘model of models’ approach will be the optimal path forward. Robust model benchmarking will be crucial, allowing financial services organisations to evaluate which AI models best align with their specific use cases, maximise performance, and deliver the highest return on investment. R1 has been described as AI’s ‘Sputnik moment’—and just as Sputnik triggered a massive acceleration in change, we will now see the same in AI. The main challenge for the financial services industry will be keeping pace.”