The global markets are currently riding a wave of euphoria that has seen Morningstar’s US and global indices have risen by more than 20% over the past year. Against a backdrop of enthusiasm for artificial intelligence, investors are pouring unprecedented capital into a sector that is rapidly redefining the economic landscape.

Yet, beneath the surface of these soaring indices lies a stark divergence: while semiconductor and hardware stocks have surged by an average of 80%, software companies have seen their valuations retreat by roughly 20%, said Michael Field, chief equity strategist at Morningstar, in a webcast on 16 July 2026.

As the industry grapples with whether this is a sustainable revolution or a repeat of the 1999 dotcom bubble, the complexity of the AI value chain suggests that not all companies will emerge as beneficiaries.

The great divide in technology stocks

The current market environment is defined by a massive shift in capital towards infrastructure. According to Field, an estimated $750bn has been earmarked this year alone for AI-related capital expenditure, a figure that dwarfs the combined spending of the rest of the S&P 500.

This “picks and shovels” phase of the cycle has disproportionately benefited hardware providers. For instance, Nvidia recently reported revenue growth of 85% year-over-year, an extraordinary feat for a $5trn market cap company, stated Field.

These firms are generating gross margins near 75%, illustrating why investors have been willing to drive stock prices up by as much as 1,000% over a three-year period, noted Field. However, this concentration of wealth in hardware raises questions about the long-term sustainability of such spending if software revenues do not eventually follow.

Field noted that OpenAI has 1 million business customers and 900 million weekly active ChatGPT users, but revenue still trails its heavy spending, resulting in cash burn. He argued that if it converts more businesses into higher-paying customers, its products could become highly profitable given the scalability of the business.

Mapping the five stages of the AI economy

To understand where the money is going, Field identified five distinct nodes in the AI economy. The first and currently most dominant node is hardware and computing, which accounts for approximately 75% of all AI spending from companies like Nvidia and  Amazon Web Services. This includes the chips, energy networks, and factories required to power the system.

This leads into the second node, data management operated by companies such as Western Digital, followed by the third: the models themselves, such as those developed by OpenAI and Anthropic. The fourth node consists of the software applications that integrate these models, and the final stage involves the integrators who bring the technology to the end user.

The current bottleneck—and the source of the most intense investment—remains firmly at the beginning of this value chain, as several city councils are starting to push back against the buildup of data centres. “We are seeing lobbyist groups visiting the local area and sponsoring football teams,” Field said.

Are we reliving the 1999 dotcom bubble?

The scale of investment has sparked a fierce debate over whether we are witnessing a financial bubble. In an informal audience poll, 20% of the participants believed the market is firmly in a bubble, while 50% suggested that certain areas are showing “bubbly” characteristics.

Concerns are amplified by the "circularity" of the current market, where major AI players rent spare capacity to one another to maintain their positions in the race for first-mover advantage. "If they don't move quickly, now someone else will, and they're going to lose out,” said Field. He admitted that if these massive investments do not translate into significant revenues, the situation could become "very concerning".

However, unlike the 1999 era, today’s leaders are generating massive cash flows and high margins, providing a fundamental cushion that was largely absent during the dotcom crash. Comparing Morningstar index constituent prices with their estimated fair values shows the market trades at roughly an 8% discount (see Chart 1). “This suggests that we're not in a bubble, but the discount is moderate,” said Field.

Why software moats are starting to shrink

The most significant impact of AI uncertainty has been felt in the software sector, where Morningstar recently downgraded the "economic moats"—a concept of competitive advantage promoted by Warren Buffett—of several companies it covers. Morningstar's analysts reviewed 132 firms and reduced the moat ratings for many, shifting them from wide to narrow or narrow to no moat.

The primary driver is a lack of visibility; where firms like Salesforce or Adobe – with wide moats – were once expected to generate excess returns for 20 years, that window has been shortened to 10 years (see Chart 2). Investors are worried that AI could allow businesses to "do less with more," potentially creating their own internal software and reducing their reliance on established vendors.

Until software companies can prove that AI enhances rather than cannibalises their business models, Field argued that their share prices are likely to remain under pressure.

Finding value at the core of the AI chain

Despite the high prices of hardware giants, the most “Core AI stocks”—those with the highest revenue exposure to the theme—actually appeared lightly undervalued or materially undervalued in recent research (see Chart 3). “We found it a bit strange but also pleasantly surprising,” said Field.

This paradox exists because investors have already "played out" many of the peripheral themes, such as industrial companies like Siemens or Caterpillar that provide infrastructure.

Furthermore, popular mega-cap stocks like Apple are often held in AI thematic portfolios (see Chart 4) despite having low direct revenue exposure to the technology, noted Kenneth Lamont, principal of manager research at Morningstar. Yet their equity team gave Apple a score of zero for AI exposure, stressing that it remains primarily an “iPhone company for the foreseeable future when we're talking about revenues,” he said.  He added, "Why are you paying extra for thematic funds for something that gives you broad technology exposure?”