Investor anxiety over soaring artificial intelligence valuations has intensified in recent weeks, with market pullbacks triggered by concerns that the AI trade has run too far, too fast. Yet JPMorgan Asset and Wealth Management CEO Mary Callahan Erdoes believes the narrative is being framed incorrectly. Speaking at CNBC’s Delivering Alpha conference, she argued that investors should focus less on fears of an AI bubble and more on the scale of opportunity that remains ahead. Despite periodic turbulence, markets continue to trade near record highs, underscoring strong institutional conviction about long-term AI adoption.

Rethinking the Bubble Narrative

Erdoes challenged the idea that AI valuations—particularly in megacap names like Nvidia and AMD—represent speculative excess. Instead, she said markets are struggling to price in a technological shift whose full economic impact has barely begun.

According to Erdoes, AI adoption today resembles the early stages of past industrial revolutions, where initial investments ran ahead of visible corporate earnings. She referenced the Hemingway quote—“gradually, then suddenly”—to describe how AI-driven productivity gains may emerge. Companies have not yet absorbed the full benefits of integrating AI into operations, she noted, leading to a mismatch between market expectations and reported financial performance.

Her view contrasts with short-term market behavior, where investors periodically rotate out of AI-heavy names during risk-off sessions. Thursday’s sharp sell-off—the worst day in more than a month—illustrated how quickly sentiment can shift when valuations appear stretched. But Erdoes stressed that volatility should not be mistaken for structural fragility.

A Market Still Early in the Adoption Curve

Erdoes’ assessment aligns with broader institutional sentiment. Ares Management CEO Michael Arougheti echoed that current investment levels are modest compared to AI’s long-run economic potential. Demand for compute capacity, infrastructure, and enterprise AI deployments continues to outstrip supply, he said, creating a multi-year runway for growth across hardware, software, and services.

This reinforces a key dynamic of the AI economy: before efficiency gains materialize, corporations must undergo a period of heavy investment in data centers, chips, cloud architecture, and workforce restructuring. For markets, this translates into revenue expansion before cost synergies begin to meaningfully lift profitability.

Erdoes emphasized that U.S. corporations are still in the early stages of integrating AI across “the bottom line”—a process that could transform every major sector. The disconnect between valuations and realized earnings, she argued, reflects timing rather than mispricing.

Macro Conditions Support Risk Appetite

Beyond AI, Erdoes offered an upbeat assessment of the macro environment. While economists have warned repeatedly of a looming recession, she noted that five years of recession predictions have yet to materialize. Credit markets remain healthy, consumer spending resilient, and corporate balance sheets relatively strong.

For investors, she argued, the absence of recession risk should encourage a proactive rather than defensive approach. In her view, this is a favorable backdrop for taking advantage of dislocations—especially in credit—before the broader market recognizes the full economic impact of AI-driven growth.

Looking Ahead

The coming years will reveal whether AI valuations are justified by transformative corporate adoption or if markets have priced in too much too soon. But for institutional investors like Erdoes and Arougheti, the message is clear: the AI revolution is still in its formative phase, with structural growth drivers far outweighing bubble risks. As companies scale investment and the technology begins reshaping productivity metrics, markets may see a shift from speculative enthusiasm to measurable economic returns—suggesting the real inflection point has yet to arrive.

Key Points:
AI valuations may appear stretched, but JPMorgan’s Erdoes argues the technology’s economic impact is still underestimated.
Institutional leaders believe AI investment remains early-stage, with demand far outpacing current infrastructure capacity.
Supportive macro conditions and strong credit markets could amplify AI-driven opportunities in the coming years.