Guide to J.P. Morgan 2026 Investment Report: Five Key Signals to Spot the AI Bubble, Agentic Succeeds LLM

J.P. Morgan Releases 2026 Investment Outlook Report: AI, Global Fragmentation, and Inflation as Key Themes

J.P. Morgan has released its investment outlook report for 2026, focusing on three key themes running throughout the report: Artificial Intelligence (AI), Global Fragmentation, and Inflation. The firm notes that the era of low inflation and seamless globalization has clearly come to an end.

In its place, three powerful and interrelated forces are shaping a new market landscape. Together, they pose a fundamental challenge: In an environment where the productivity growth driven by AI is at odds with more persistent, volatile inflation and a fragmented world order, how should we invest? This article focuses on summarizing the sections related to the AI bubble.

2026 Investment Keywords: AI, Global Fragmentation, and Inflation

J.P. Morgan believes the most significant change brought by AI is the potential to drive the cost of professional skills to zero, describing this shift as comparable to the rise of computing technology. This new technology may improve productivity and enhance corporate profit margins, but it also introduces new challenges. The labor market will face significant disruption, and market bubbles may emerge. The key in the future will be how to seize the opportunities brought by this transformation while avoiding the risks of technological obsolescence and irrational exuberance.

On the geopolitical front, the global order is breaking down, forming competing blocs, fiercely contested supply chains, and fragile alliances. Securing natural resources and energy has now become a strategic imperative. As these dynamics reshape trade and capital flows, they also create interesting opportunities for investment returns (and losses). As the pursuit of efficiency is replaced by the need for resilience and security, which regions and industries are poised to stand out?

J.P. Morgan warns that inflation volatility will be higher than in the pre-pandemic era, reflecting persistent fiscal deficits and rising household wealth. However, they also expect that in the coming year, solid economic fundamentals will provide tailwinds for investors. The Fed’s rate-cutting cycle and reduced economic policy uncertainty should help global economic growth return to expected levels. Lower short-term U.S. interest rates may boost global equities and risk assets like credit.

J.P. Morgan 2026 AI Keyword: Agentic

J.P. Morgan notes that since OpenAI launched ChatGPT at the end of 2022, AI’s potential has drawn significant investor attention. Three years later, the AI boom continues to intensify. The core driver is the rapid improvement in generative AI capabilities and a dramatic reduction in costs. Today’s models hallucinate less, can handle much longer context windows, and exhibit stronger reasoning abilities. LLMs are expected to reach human-level performance by spring 2026.

Although the performance growth curve of large language models is slowing, J.P. Morgan believes Agentic AI may be the next breakthrough. Agentic AI refers to AI systems that can proactively take action, plan multi-step tasks, and autonomously operate tools. The core concept: AI no longer merely answers but acts as an assistant, proactively completing tasks. For example, if I input: “Fetch today’s ETH price and update the Google Sheet,” Agentic AI can automatically fetch data from the API and update the Google Sheet.

(Google Officially Launches Gemini 3: The Most Powerful AI Agentic and Vibe Coding Large Language Model to Date)

Editor’s Note: Agentic AI has become the hottest keyword today. Not only are large language model platforms competing, but many retailers are also rapidly iterating their products to capture shopping opportunities from Thanksgiving through the Christmas season.

(Walmart, Amazon, and Google Upgrade AI Shopping Agent Functions, Driving Strong Year-End Shopping Season Performance)

Learning from History: Five Key Insights on the AI Bubble

On the hotly debated issue of the AI bubble in the investment market, J.P. Morgan shares its observations, first pointing out that AI-related investment has contributed more to U.S. GDP growth than consumer spending. In terms of market cap, nearly 40% of the S&P 500 is related to AI.

There is a set pattern to the formation of market and economic bubbles. Most bubbles are born from investors’ belief that the world is changing, prompting them to accumulate resources to meet future demand. Part of bubble formation is due to the widespread availability of credit. Looser lending standards and higher leverage lead to a disconnect between economic fundamentals and market valuations. More investors pour into the bubble until fundamentals finally prevail and the bubble bursts. Here are J.P. Morgan’s observations:

Bubbles Stem from Paradigm Shifts—No Signs of Overcapacity in AI Yet

Bubbles often arise from the belief that a new technology, demographic trend, or policy shift will profoundly change the world. Famous historical examples include the 19th-century railway mania and the late-1990s internet bubble. Both did change the world, but timing was critical. In the former, Britain’s railway network doubled in a decade, but revenue per mile stagnated or declined. In the latter, telecom companies laid millions of miles of fiber optic cable, but only 10% was used, and each cable used only 10% of available wavelengths.

Both the railway and internet booms saw massive overcapacity, misaligned with consumer demand or unit economics at the time. So far, there are no signs of overcapacity in AI—data center vacancy rates are at a historic low of 1.6%, and three-quarters of computing power in under-construction data centers is already booked. Across the value chain—compute, energy, and data centers—demand far outstrips supply. Recent earnings seasons confirm that AI applications are driving revenue growth for large enterprises.

AI-Related Debt Markets Set for Continued Growth

The key factor in the tulip bubble was Amsterdam’s deep credit markets; the 1980s Japanese asset bubble relied on bank loans collateralized by artificially inflated corporate equity; the pre-GFC real estate bubble was fueled by subprime mortgages. In the 2010s, with policy rates at zero, oil producers accessed cheap financing, creating an energy stock bubble.

Oracle’s recent entry into the bond market signals that the next phase of the AI infrastructure cycle will rely more on credit. Public markets are willing to finance large tech firms, whose spreads are below the overall investment-grade bond index. As the Fed’s rate-cutting cycle progresses, AI-related credit is likely to keep growing.

2025 AI Still Built on Cash Flow

Bubbles often expand rapidly because financial structures amplify gains while masking underlying risks. The South Sea Bubble was driven by massive debt-to-equity swaps; before the 1929 crash, the U.S. market was hooked on even higher-leverage margin lending. The recent SPAC boom scaled quickly thanks to put options at redemption and free warrants, stacking huge financial leverage.

Recently, several companies have directly collateralized AI infrastructure: Lambda and CoreWeave have issued GPU-backed bonds, and Alibaba has announced zero-coupon convertible bonds for data center expansion. Data shows that tech company bond issuance and asset-backed securities related to data centers and commercial mortgage-backed securities are back to levels seen during the 2020-2021 boom.

But these are just the surface of capital markets. If hyperscale data center operators were to leverage up to investment-grade company norms—net debt at 2.8x EBITDA—the market could theoretically unlock another $1 trillion in funding.

What matters now: Will underwriting standards start to loosen? Whether it’s PPAs (power purchase agreements), private equity, or venture capital, if standards begin to relax, that’s a clear sign of rising risk. So far, large tech firms’ operating cash flow still exceeds capex and dividends, meaning the current wave of AI investment, while large, is still mainly funded by internally generated cash. Leverage may gradually increase in the future, but for now, this wave of AI spending remains on a relatively healthy footing.

Listed AI Companies Corrected by the Market, Bubble Already Appearing in Private Markets

In bubble eras, valuations often disconnect from fundamentals. During the dot-com bubble, some companies went public with zero revenue. Cisco’s stock price increased 40-fold from 1995 to 2000, while profits rose only 8-fold. J.P. Morgan notes that the private market is already showing signs of a bubble, as AI startup valuations are consistently outpacing those of non-AI firms at every funding round.

However, in public markets, AI company returns are entirely driven by profit growth. Over the past three years, the forward P/E of listed AI stocks has declined, while expected EPS has more than doubled. Over the past five years, NVIDIA’s stock price is up 14x, while profits have grown 20x.

J.P. Morgan: Focus on Who Captures Value, Not Just the AI Bubble

Every bubble is the same: as prices rise, more people believe. Dutch craftsmen once spent several times their annual income on tulip bulbs; in 2005, Las Vegas bartenders rushed to flip homes. Recent IPO trends also hint at similar bubble behavior. Market sentiment is indeed heating up, but only when the frenzy reaches another level does it become truly alarming.

Overall, most of the conditions for bubble formation are already in place. However, J.P. Morgan believes the current risk is not that the bubble has peaked, but that the probability of a bubble forming is increasing. The more important question than whether AI is entering bubble territory is: who will capture the majority of value from this technological shift? Unfortunately, history offers no consistent, replicable model.

In some industries, early movers fared poorly—UK railways, fiber optics, and telecoms all saw pioneers take the risk and invest capital, only for later entrants to profit after asset prices collapsed. Conversely, in IT, early movers like IBM, Microsoft, Cisco, and Amazon maintained their ecosystem dominance and market share. U.S. utilities defended their market share, but regulatory constraints limited investor returns, so early movers didn’t capture all the profit.

This article, A Guide to J.P. Morgan’s 2026 Investment Report: Five Key Signals for the AI Bubble, Agentic Succeeds LLM, first appeared on Chain News ABMedia.