JPM Market Intel Global Thematic Views July 2026

THEMATIC PULSE – July 2026 JPM GLOBAL MARKET INTELLIGENCE

Below is a summary of the views which we investigate in this note with the ‘Deep Dive’ section including (i) a detailed rationale for the theme; (ii) key charts; and, (iii) color on positioning.

• GLOBAL TECH – After a prolonged period of strength, the Global Tech/AI theme has come under some pressure in July. This was driven by a combination of: • Challenges to fundamental narrative incl. growing concerns around competition (resurgence from China AI); pricing/costs; excess compute and challenges to capex narrative. • De-leveraging (elevated for retail investors & ETFs, and coincided with large end-of-quarter rebalances); • Elevated positioning (Tech remains stretched, with Global Semis at 98th %-tile for both L/S ratio and Net exposure and Tech Hardware L/S ratio at 86th %-tile, net exposure ~99th %-tile); • Extent of rally across the theme, as well as Momentum, with Tech a core component of the long leg; • Seasonality (as of late-June, starting points of gross flows and leverage were at highs, and we’ve seen HF de-grossing every July for the past 5 years); • BULL / BEAR FOR SEMIS & GLOBAL MONETIZATION • BULL CASE – (i) Fundamental Story Remains Intact; (ii) Historical Drawdown Creates Entry Point; (iii) Earnings Tailwind • BEAR CASE – (i) Increasing Competition / Margin Compression Risk; (ii) Vol Shock Potential; (iii) Geopolitical / Rates Headwind • TRACKING FLOWS IN THEMATICS – We track flows into 16x Secular Equity Themes. The AUMs are relatively small (~$400bn) but are rising fast (20% organic inflows in June, 2.5-3x the rate for broader US Equities). Over the past month, the Themes with the strongest inflows have been AI and Robotics & Automation, while those with weakest inflows have been Space and Blockchain.

THEMATIC PULSE – July 2026 JPM GLOBAL MARKET INTELLIGENCE

New narratives are being reflected an evolving landscape for Global Tech & AI, with key themes to track as Tech progression advances. Below we look at (i) Open vs. Closed models; (ii) Pricing Wars; (iii) Chinese LLMs; and, (iv) Capex Challenges

1. Open (Weight & Source) vs Closed Frontier Models (source here)

Open-weight models are those where the model weights are publicly released. In “Open- weight” models do not release training data and code are often — just the weights themselves (vs open-source). Open-weight models have shown similar competitiveness on general tasks, and tend to be much cheaper. Another advantage is around data privacy, as data is kept on your own infrastructure. Closed frontier models are accessed only through an API – you never see the weights. the provider handles training, infrastructure, safety systems, and updates.

• DeepSeek • GLM-4 from Zhipu AI • Meta Llama 3.1 / 3.3 • Mistral • Phi-4 from Microsoft