J.P. Morgan Mid-Year Outlook: AI Narrative Has Not Peaked; Reduce Cash, Increase Allocation to Real Assets
- Core View: J.P. Morgan believes the market is overly pessimistic in pricing the three major global risks (fragmentation, inflation, AI disruption). The current volatility presents an entry window. The firm recommends continuing to bet on the AI super-cycle and US equities, hedging against inflation with real assets, reducing cash holdings, and watching for opportunities in emerging markets.
- Key Elements:
- AI Capex Unabated: Capital expenditure expectations for the top five cloud giants by 2026 have surpassed $650 billion, an increase of $130 billion from the previous quarter. The market is pricing in an "AI peak," but the data does not support this.
- Changing Financial Profile of Tech Giants: The free cash flow of these five giants is projected to drop from $240 billion in 2024 to $73 billion in 2026, marking a shift from an asset-light to a capital-intensive, high-investment model.
- Pressure on the SaaS Industry: Nearly half of the components in the software index have fallen over 50% from their highs. AI is disrupting the subscription-based software business model, with approximately 21% of direct lending funds flowing to software companies.
- Inflation More Sticky Than Pre-Pandemic: Core PCE was already sticky at 3% before the energy shock. Prices have risen 25% cumulatively, while core fixed income has only returned 6%, meaning cash and bonds face real losses.
- Historical Data Supports Buying the Dip: After a ~10% correction in US stocks, buying when the VIX breaks above 30 yields a positive return 70%-83% of the time within six months, with an average return of 12.4%.
- Emerging Market Potential Emerging: EM corporate earnings are expected to grow by 46%, with a P/E ratio of just 11.8x. The discount of Chinese equities relative to Asia is at its deepest in 20 years, suggesting a potential structural revaluation if policy shifts occur.
Original Author: David
Key Takeaways:
On June 1st, J.P. Morgan Wealth Management released its Mid-Year Outlook 2026 report, essentially offering guidance to its high-net-worth clients on investment strategies for the second half of the year, marking the halfway point.
Against a backdrop of rising oil prices due to the Strait of Hormuz blockade, resurging inflation, and the AI narrative shifting from euphoria to skepticism, the report's overall tone is cautiously optimistic, albeit with specific adjustments to investment allocations.
J.P. Morgan believes that the market has overly pessimistically priced the three major global risks (fragmentation, inflation, AI disruption), and the current volatility presents a window of opportunity for entry.
The overall assessment is:
Continue betting on the AI supercycle and US equities, hedge against inflation with real assets and alternative strategies, reduce cash holdings, and focus on emerging markets.
If you hold US tech stocks, or are considering increasing or decreasing your position in the second half of the year, the framework and data in this report are worth reviewing. We have condensed and interpreted the original report, rearranging priorities based on investment relevance.
Six Key Conclusions:
① The AI Supercycle Is Not Over; The Market Is Overly Pessimistic.
The combined 2026 capital expenditure expectations for the five Hyperscalers (Microsoft, Meta, Oracle, Google, Amazon) exceed $650 billion, an increase of $130 billion from the previous earnings season. AI-related investments contributed 25 basis points to US real GDP growth in 2025. Taiwan's GDP growth rate exceeded 7%, the fastest since 2010, driven primarily by semiconductor exports. JPM believes the market is pricing in "peak AI," but the data does not support this narrative.
② But the Financial Profile of Hyperscalers Is Changing.
Free cash flow is projected to decline from $240 billion in 2024 to $73 billion by the end of 2026. Microsoft's forward P/E ratio has fallen from a high of 35x during the AI era to 22.5x. These companies are transitioning from "asset-light, high-return" to "asset-heavy, high-investment" models, a shift the market is still digesting.
③ SaaS Is Experiencing an Underwater Bloodbath.
Approximately half of the components in the S&P Software Index (IGV) have declined over 50% from their all-time highs. JPM's basket of "AI vulnerable targets" is down nearly 20% this year. In the private credit market, 21% of exposure is to software companies, rising to 40% when including technology and business services. The impact of AI on the subscription software business model is already underway.
④ The Inflation Floor Is Higher Than Pre-Pandemic; Cash Is Bleeding Out.
US Core PCE was already sticky at 3% before the energy shock. Since the 2020s, consumer prices have risen by a cumulative 25%, while core fixed income has only returned 6%. JPM clients hold nearly 20% of their assets in cash and short-term bonds. The report's message is clear: while you think you are de-risking, you are actually losing money.
⑤ The Strait of Hormuz Blockade is the Biggest Oil Supply Shock Since WWII, but JPM Advises Buying the Dip.
Oil prices nearly doubled, US stocks experienced roughly a 10% correction, and the S&P 500's P/E ratio briefly dipped below 20x. JPM's historical data shows that buying after the VIX breaks above 30 yields a 70% to 83% probability of positive returns within 6 months, with an average return of 12.4%.
⑥ Emerging Markets Could Be an Opportunity in the Second Half.
EM corporate earnings are expected to grow by 46%, with a P/E ratio of only 11.8x. Taiwan and Korea are key nodes in the AI hardware supply chain. Latin America holds over 40% of the world's copper and nearly 60% of its lithium reserves. The discount of Chinese stocks relative to other Asian markets is at its deepest in 20 years, and JP's stance is "cautiously warming up."
On AI: Market Pricing 'Peaked,' JP Morgan Says It's Premature
JPM opens by stating that Wall Street's narrative on the AI supercycle "has become too pessimistic."
Key data supporting this assessment:
The combined 2026 capital expenditure expectations for the five cloud giants (Microsoft, Meta, Oracle, Google, Amazon) exceed $650 billion. Cloud rental prices for GPUs (the core chips for training AI models) have risen 40% since last October, with supply still unable to meet demand. Nvidia's stock is trading at a 40% discount to its average P/E ratio over the past decade. The market is pricing in 'peak chip sales,' but cloud business revenue is still accelerating.
At the same time, the financial profiles of these five companies are changing. Free cash flow is projected to decline from $240 billion in 2024 to $73 billion by the end of 2026. Microsoft's P/E ratio has fallen from a high of 35x during the AI era to 22.5x. The asset-light model that attracted investors over the past decade is being rewritten by heavy capital investment. JPM believes that for now, focus should be on revenue growth rather than cash flow, but this also means that if demand slows down, these investments could become a drag.
Other judgments on AI serve as partial risk warnings under the major trend:
Traditional software companies are among AI's first true victims. Approximately half of the components in the US software sector index have declined over 50% from their highs, with a median operating profit margin of only 4%. The logic of the impact is simple: SaaS charges per user, and AI reduces the number of users. This has already transmitted to the lending market. About 21% of loans in the US direct lending market are to software companies, and publicly traded tech loan fund prices are near cycle lows. JPM's stress tests show potential losses including leverage could reach 4% in extreme scenarios, but it does not currently pose a systemic risk.
SpaceX, Anthropic, and OpenAI might go public this year in a cluster, which historically is not a good omen. Following the 25 largest IPOs in history, the median new stock underperformed the broader market by 30 percentage points in its first year; 12 out of 18 fell in their first year. In years with mega-IPOs, the median annual return of the broader market was only 3%, far below the long-term average of 10%. JPM did not definitively state a peak, but explicitly treats the market reaction to SpaceX's IPO as a cyclical thermometer.
On Inflation: Inflation Won't Return to 2%, Your Cash and Bonds Are Losing Money
The key point of this inflation section is not that the Strait of Hormuz pushed up oil prices, but that before oil prices were pushed up, US inflation had not returned to normal levels.
In January 2026, Core PCE was at 3.1% year-over-year, with particularly stubborn increases in local service categories like dining and personal care. Then oil prices doubled. The Fed's model shows that for every $10 increase in oil per barrel, inflation rises by about 0.3 percentage points; this time it rose by $40.
JPM believes a full repeat of the 1970s is unlikely. There is no wage-price spiral in the labor market; the voluntary quit rate is declining; housing inflation has fallen from 5% at the end of 2024 to just over 3%; and China's overcapacity is depressing global goods prices. However, the inflation floor is significantly higher than before the pandemic, likely hovering around 3%.
JPM's recommended response is to increase allocation to real assets.
Since 2020, US consumer prices have accumulated a 25% increase, while bonds have only returned 6%, and cash even less. You might think your money is sitting idle, but it's actually shrinking every year. Among JPM's own clients, nearly 20% of assets are still in cash and short-term bonds.
Therefore, its advice is to move some money into inflation-linked assets:
- Commodities, infrastructure, real estate – things that rise with prices – a combined allocation of around 5% of the portfolio is suggested.
- Gold individually: 3% to 6% is recommended.
- Also, hedge funds: in 2022, when stocks and bonds fell together, macro strategy hedge funds returned 9%. However, JPM acknowledges that 94% of its own private bank clients have never bought a hedge fund, and 86% have never bought infrastructure products.
To summarize this section in one sentence:
Inflation may not spiral out of control, but it will not return to 2%. If your portfolio is still the classic 60/40 stock/bond mix plus a pile of cash, JPM believes you are preparing for a world that no longer exists.
On Geopolitics: Chinese Stocks May See a Structural Revaluation
This section covers the most diverse topics, from the Middle East conflict to US-China rivalry to Europe's predicament. We will only highlight points directly relevant to investment decisions.
1. The Strait of Hormuz blockade was the biggest market shock in the first half of this year. Approximately 20 million barrels of oil pass through this chokepoint daily, accounting for one-fifth of global oil consumption. After the US-Israel joint strikes on Iran, oil prices nearly doubled in days, and European natural gas prices surged nearly 100% in two days. QatarEnergy's CEO stated that 15% of LNG capacity could be offline for up to five years. Qatar also supplies about 30% of the world's helium, which is essential for chip manufacturing; South Korea has already warned of potential chip factory shutdowns.
JPM believes the conflict is moving towards de-escalation, but the damage to physical infrastructure and the risk premium on energy will not disappear quickly.
Therefore, their advice to investors is: Buy US stocks on the dip.
US stocks fell about 10% in the first half of the year, with the S&P 500's P/E ratio briefly dipping below 20x. Historically, buying after the VIX breaks above 30 yields a 70% to 83% probability of positive returns within 6 months, with an average gain of 12.4%.
2. The US and China are building their own ecosystems, and markets may accelerate into two camps. The US is restricting chip exports to China, pressuring the Netherlands and Japan to restrict semiconductor equipment. China is expanding exports to non-US markets; Belt and Road investment reached a record high in 2025, investing $53 billion in Brazil in one year, and its total trade volume with Latin America has surpassed that of the US. JPM's assessment is that future investment returns may increasingly depend on which camp your assets belong to, not just the growth of the company itself.
But fragmentation also creates opportunities, especially in emerging markets.
JPM lists several directions:
- Latin America holds over 40% of the world's copper and nearly 60% of its lithium, also rich in nickel, rare earths, and agricultural resources. Foreign direct investment has doubled in the past two decades, central banks are better at controlling inflation than developed countries, and politics are shifting towards more pragmatic, pro-business governments.
- Gulf States in the Middle East are using oil revenues to build AI data centers. Saudi Arabia partnered with BlackRock on a $3 billion data center project, with costs 30% lower than the US.
- East Asia (Taiwan, South Korea) controls key nodes in the AI hardware supply chain. If AI capital expenditure continues to accelerate, these economies' exports and pricing power will continue to strengthen.
- Chinese stocks trade at their deepest discount to other Asian markets in 20 years. 80% of Chinese consumers are excited about AI products (compared to 38% in the US), and electricity costs are roughly half of the US level. JPM's stance is "cautiously warming up." If clearer pro-business signals emerge from the policy side, Chinese stocks could see a structural revaluation.
In contrast, Europe is the market where JPM is most conservative. Electricity prices are two to four times those in the US, R&D spending is only 2.2% of GDP (US 3.6%, South Korea 5.2%), and the venture capital scale is one-tenth of the US.
The energy shock also pressures the European Central Bank into potentially raising rates again. In Europe, JPM only recommends buying defense and infrastructure-related assets, avoiding automobiles and consumer goods.
What JPM is Betting On, and What It Isn't
Condensing the 60-page report into one sentence: Volatility is an entry opportunity, but the entry strategy needs to change.
What You Should Bet On:
- AI Infrastructure Supply Chain (chips, optical modules, power), Emerging Market Stocks and Bonds, Real Assets (commodities, infrastructure, gold), Defense-related Targets, Chinese AI Concepts (cautious addition).
What You Should Not Bet On:
- Cash, Traditional Subscription Software Companies, European Automobiles and Consumer Goods, and a pure 60/40 stock/bond investment model for the second half of the year.
Attached original report link:
