In November 2025 we did a webinar called “2026: IFRS 9 Challenges”. Here is a link to a youtube recording for reference:
Some of the key takeaways were:
- The forecasts for 2026 were relatively close to 2025.
- We expected significantly higher risk than the forecasts indicated. As I said: “The risks that we see now are in our view much higher than the risks that were there one or two years ago”.
- Critical areas of risk included: global conflict, global debt, AI.
It seems that our projections were correct and 2026 until now is a much more volatile, especially with regard to global conflicts.
With regard to global debt, it is good to note a small decrease in Debt/GDP over the EU and Eurozone area as indicated by the latest figures (as of the end of 2025) from Eurostat:
https://ec.europa.eu/eurostat/web/products-euro-indicators/w/2-22042026-bp
Nevertheless the debt levels remain high, especially in some EU countries (Greece 146%, Italy 137% and France 116%), which creates uncertainty because Italy and France are significant part of the Eurozone.
UK’s debt/GDP has also decreased slightly as of March 2026 to 94%.
The trends in the US continue to be negative with Debt/GDP reaching over 122% as of the end of 2025.
https://fred.stlouisfed.org/series/GFDEGDQ188S
It remains to be seen what effect the current war between the US and Iran would bring, but the supply shock may lead to notable increase in CPI/PPI, which in turn would affect negatively the economy. Given the very large debt/GDP ratio to some economies, there could be issues with both monetary and fiscal measures, in that regard we continue to see this as persistent threat globally.
On the last risk we mentioned, AI, we continue to view this as a very significant risk. There seems to be a division of people in two camps. One group of people (mostly AI outsiders) claim this is a bubble and it will burst possibly sooner than later. The other (mostly AI insiders) claim this is nothing like the .com bubble as there are basically no GPU/TPUs left unused, in essence how can this be a bubble when we are running into severe supply limitations. At this moment we feel this could go either way. One thing is clear, though. In my view frontier models need radical improvement in model performance and utilization to make the financials work. We do not believe small improvements would get trillion dollar investments in the black.
