JPM Tracking the AI buil

Sales & Trading FICC Market Structure & Liquidity Strategy JPMORGAN 13 August 2026

Key developments from July 2026 1. Does AI investment create financial risk? 2. Trading infrastructure turns to tokenization 3. T+1 settlement deadlines enter the orbit 4. Dubai launches sweeping review of its funds regime 5. Another go at IGB electronic trading 6. The EU Emissions Trading System gets a revamp 7. What’s needed to activate overnight KRW liquidity? 8. Back to the drawing board on best ex? 9. Self-certification of event contracts refined 10. Hong Kong’s fixed income and FX ambitions take shape

Explore our content on J.P Morgan Markets and contact our team for more information on this report, or if you would like to be added to the J.P. Morgan FICC Market Structure Liquidity Strategy distribution list.

1. Does AI investment create financial risk?

As AI build-out becomes a defining feature of the global investment landscape, standard setters are increasingly asking whether the same forces driving scale and competition could also introduce new sources of fragility. While this may not yet be the case, certain policy makers are moving pre-emptively rather than waiting for a stress event as two channels that could create risk (leveraged, circular AI financing and AI’s correlated role within markets) accelerate faster than oversight.

In a research paper published on July 14, the Bank for International Settlements (BIS) compares the current environment with prior investment booms, arguing that the AI race can generate over-investment when firms prioritize speed and scale over more measured capacity expansion.[1]

The BIS characterizes the build-out as a winner‑take‑most contest in which private incentives to secure early scale and lock in dominant positions can exceed the socially efficient level of investment. In that framework, the pace and financing structures for capital deployment can become a channel through which competition translates into systemic vulnerability.

The paper estimates that hyperscaler capital expenditure is on track to exceed $700 billion in 2026, with spend potentially running into the trillions over coming years. The authors consider whether the rising share of build-out activity funded through debt and circular financing structures increases interconnectedness and contagion risks.

Figure 1: Visual summary of AI financing structures

Sales & Trading FICC Market Structure & Liquidity Strategy JPMORGAN 13 August 2026

A complementary FEDS Notes[2] published on July 17 presents a framework for using publicly available data to track the generative AI build-out and its potential impact on the economy, organizing indicators into three categories: capabilities and costs, firm investment and adoption, and productivity and labor.

Summarizing recent trends across each indicator, the Fed demonstrates that the economic effects of AI have been concentrated within certain highly exposed sectors such as financial markets, creating divergence in productivity gains. The authors conclude that as the broad-based adoption of AI takes shape over time, productivity and labor market effects will become more evident and evenly distributed.

Separately, a post from the International Monetary Fund (IMF) published on July 23 flags AI adoption within the financial services sector as a potential source of financial stability risks. The author focuses on how AI has become increasingly embedded in the decision-making architecture of finance, reshaping how firms price risk, allocate credit and respond to stress.[3] The IMF suggests three priorities for central banks and supervisors:

1. Stronger oversight of AI-driven trading and lending 2. Better visibility into AI use and correlation risks that can arise when many models respond to the same signals 3. Deeper cross-border coordination on operational resilience and cyber defense Noting the potential for AI to improve execution, liquidity and risk assessment in normal conditions, the IMF cautions that speed, opacity and synchronized reactions could amplify shocks during stress and expose system-wide vulnerabilities.

With AI policy initiatives proliferating across global jurisdictions, the common principles and…