How SVB’s collapse still casts a shadow over global banking risk

When Silicon Valley Bank (SVB) collapsed in March 2023, it wasn’t just a US banking story. The sudden implosion of the third-largest bank failure in US history revealed systemic vulnerabilities that still reverberate across the global financial system.

More than two years later, many structural forces remain, including interest rates that are only gradually cooling in the face of persistent and fluctuating inflation, in part due to the impact of tariff policies and geo-political instability. This drives both short and longer term asset-liability mismatches and ineffective management or scenario planning. In summary, such external factors brought down SVB and today they continue to pose significant risks for banks and financial institutions worldwide.

Concurrently, the financial services industry is bracing for another wave of potential shocks. S&P Global forecast around $850 billion (£600 billion) in credit losses across firms by the end of 2025. For regulators, boards and investors, that raises an urgent set of questions: 

• Have banks’ risk models truly evolved since SVB? 

• Are firms resilient enough to withstand stress events in a higher-for-longer rate environment? 

• And what role should new technologies - particularly AI - play in future-proofing risk management?

From reactive forecasting to intelligent resilience

SVB’s downfall was as much about governance and foresight as it was about balance sheet structure. Rising interest rates eroded the value of its long-dated bond holdings, while inadequate scenario planning left the bank exposed when depositors rushed to withdraw funds. Despite regulatory scrutiny across geographies, ineffective forecasting and presumed resilience remain widespread and potentially significant.

The mitigation to these risks lies in moving beyond stress testing based on historical trends, towards intelligent resilience based on forward-looking assumptions using the right technology. In our latest global risk management survey, we revealed that 75% of banks intend to increase investment in risk technology infrastructure (up from 51% in 2021), while 64% plan to augment spending on third-party software (vs. 43% in 2021).

Traditional models have long been built to look to the past. But the past is no longer a suitable sole proxy for the future. Today, however, banks need to anticipate the future for competitive advantage, using more qualitative information such as management judgement, transformed into intelligent and robust analytics. 

By adopting this forward-thinking approach, institutions can unlock benefits across functions and make more strategic decisions for the benefit of their end customers and financial performance.

The role of AI

AI-enabled risk management tools continuously scan scenarios, model exposures and adapt to rapidly changing conditions in an automated way, proposing rather than waiting to be asked. By combining explainable AI with advanced scenario modelling, these tools enable firms to detect vulnerabilities earlier across portfolios, scanning vast datasets in real time to flag emerging risks before they escalate. Both generative and agentic AI take this a step further, turning these predictive insights into intelligent decisions. 

AI also strengthens stress testing by simulating a much wider range of “what if” scenarios than human analysis or traditional computing - from rate shocks to liquidity crunches overlaid with multi-variate permutations of evolving marco-economic factors. This offers firms a more agile view of resilience than traditional models and drives competitive advantage to anticipate new opportunities. Additionally, this automation streamlines reporting, reducing the burden of regulatory compliance and freeing risk teams to focus on higher-value decision-making.

To summarise, AI is not a substitute for sound governance but a “force multiplier” that helps banks manage uncertainty proactively. Institutions that embed AI into their risk frameworks starting with the end customer in mind will be better prepared to withstand shocks, as well as better positioned to seize opportunities in an unpredictable market.

Preparing to avoid the next headline

The 2008 crisis was meant to hardwire resilience into banking, yet SVB’s collapse showed how quickly old risks resurface if governance and control are ineffective. For banks, the priority is clear: move beyond fragmented models and legacy systems to embed continuous, intelligent risk monitoring & decisioning. 

This means integrating data and automating analysis to enable faster, more confident decisions. Resilience is no longer just about avoiding losses; it’s about creating the agility to seize opportunities when others pull back.

It’s clear that AI will play a central role here, by giving firms the ability to anticipate shocks, strengthen compliance and demonstrate future-readiness to regulators and investors. Those that act now will not only protect themselves from becoming the next headline, but also set the pace for the next era of financial stability and prosperity.