A prominent warning signal is quietly flashing in the financial markets, echoing the structural imbalances that preceded the collapse of the dot-com bubble. Analysts at JPMorgan Chase & Co. have flagged a growing divergence within the artificial intelligence sector, where the stock prices of hardware winners—including semiconductor designers, memory makers, and equipment manufacturers—have separated dramatically from the largest technology companies funding this capital expenditure boom. This widening performance gap represents an unsustainable long-term trend, raising the risk of a major market correction if the massive spending does not soon translate into clear commercial revenues.
According to a research note by JPMorgan strategist Jason Hunter, the growing contrast and negative trend in the stock prices of hyperscalers are highly reminiscent of the market behavior of 1999 and early 2000. During the final months of the internet boom, companies building the physical infrastructure of the web experienced spectacular, near-vertical growth, while the telecom firms spending billions of dollars to buy that equipment saw their stock prices peak and plummet. Since the dot-com bubble burst less than a year after this divergence was first observed, the current split has forced Wall Street banks to keep a close eye on the individual charts of tech giants to see if they can find a secure foothold this summer.
The Mechanics of the 1999 Dot-Com Analogy
To understand why the current market split is causing anxiety among institutional investors, one must examine the specific mechanics of the 1999 telecom boom. During the late 1990s, the undisputed darlings of the stock market were the telecommunication and network equipment suppliers, such as Cisco Systems, Nortel Networks, and Lucent Technologies. These companies manufactured the routers, switches, and fiber-optic cables required to build out the physical infrastructure of the emerging internet.
Backed by seemingly unlimited investor optimism, these equipment makers experienced parabolic growth, with their market capitalizations reaching historic heights. At the same time, the major telecommunication companies—the “spenders” who were borrowing heavily to purchase this equipment—struggled to monetize their massive new networks. As investors grew increasingly concerned about the low return on investment (ROI) and the slow pace of consumer adoption, they began selling off the telecom stocks. When the spenders were forced to freeze their capital budgets, the demand for networking equipment collapsed overnight, popping the dot-com bubble and triggering a multi-year market downturn.
The parallel to 2026 is striking. Once again, investors are pouring hundreds of billions of dollars into the companies building the physical infrastructure of a technological revolution—this time, artificial intelligence. However, the software giants who are actually buying the chips and constructing the data centers are facing growing skepticism over their monetization timelines. This has created a highly fragmented market where the providers of the “shovels” are enjoying record-breaking profits, while the “miners” are seeing their stock prices suffer, creating a classic structural imbalance.
The Stark Divergence in 2026 AI Stock Classes
The physical data from the first half of the year clearly demonstrates this deep split in the market. Since January, the companies that manufacture semiconductors, high-bandwidth memory, and advanced cooling equipment have posted some of the strongest returns in history. The Philadelphia Semiconductor Index (SOX) has soared approximately 87% in 2026, completing its best-ever quarter, while specialized hardware plays have seen even more explosive growth. For example, the Roundhill Memory ETF has surged 141% since its initial launch in April, reflecting the critical supply shortages and rising prices of advanced memory components.
Conversely, the massive technology conglomerates that are actually financing this capital supercycle have experienced a significant pullback:
- The Spenders’ Slump: The Roundhill Magnificent Seven ETF, which tracks the largest U.S. technology giants, has fallen approximately 7% from its peak earlier this year.
- Microsoft’s Market Capitalization: Despite investing tens of billions of dollars into its Partnership with OpenAI and its native Azure AI services, Microsoft’s market capitalization has fallen by more than 18% this year, marking its worst monthly stock slump since the dot-com era.
- Meta Platforms: The parent company of Facebook and Instagram has seen its shares drop 5% from their peak, as investors grow increasingly nervous about the company’s massive $145 billion capital expenditure guidance.
- The Valuation Gap: While hardware suppliers are enjoying massive upward earnings revisions, the actual buyers of these chips are being punished by investors who are airing growing concerns over whether AI will pay off from a monetization perspective.
This stark divergence means that the broad market indexes are resting on a highly concentrated, vulnerable foundation, where a pullback in just a few high-flying semiconductor names could drag down the entire financial system.
Two Scenarios for Market Re-convergence
In a separate note to clients, JPMorgan analyst Nikolaos Panigirtzoglou outlined two potential scenarios for how this unsustainable performance gap between chip winners and hyperscalers might ultimately resolve itself. The first is an optimistic “catch up” scenario, while the second is a highly disruptive “catch down” correction.
The Optimistic “Catch Up” Scenario
Under the positive scenario, the massive capital investments made by the hyperscalers are validated by a rapid acceleration in commercial revenues. Enterprises, small businesses, and retail consumers begin to adopt generative AI tools at a scale that generates robust, high-margin subscription and transaction revenues.
As these software revenues materialize, the profit margins and stock prices of the hyperscalers rise to “catch up” with the hardware winners, allowing them to capture a larger share of the overall AI value-added pie. This outcome would confirm that the current capex boom is structurally sound, supporting a sustainable, multi-year expansion for the entire technology sector. JPMorgan noted that its official house view continues to favor this more optimistic outcome, pointing out that corporate demand for advanced computing power remains historically high.
The Pessimistic “Catch Down” Scenario
Under the negative scenario, the return on investment for generative AI tools remains disappointingly slow. Enterprises find that the high costs of running and integrating advanced models do not deliver sufficient productivity gains to justify their expensive software subscriptions, leading to a cooling of market demand.
Faced with stagnant revenues and high debt holding costs, the major cloud giants are forced to scale back their capital expenditure plans, cutting their orders for advanced chips and data center components. This sudden reduction in demand would trigger a sharp “catch down” correction, hitting highly valued semiconductor and memory stocks significantly harder than other sectors, as the high earnings expectations currently baked into their stock prices are revised downward. JPMorgan warns that the bottom-up analyst consensus currently points to a sharp deceleration in hyperscalers’ capex growth starting next year, a trend that, if taken at face value, tilts the scale toward this negative scenario.
The Discrepancy in Corporate Earnings Growth
The divergence between the hardware winners and the spenders is also clearly reflected in their corporate earnings trajectories. While the major cloud providers are investing hundreds of billions of dollars into physical infrastructure, the primary financial beneficiaries of this spending today are the companies supplying the hardware.
This earnings mismatch has reached extreme levels:
- Five-Fold Growth Gap: Consensus estimates project that U.S. semiconductor companies’ earnings will grow by a massive 98% in 2026, representing a growth rate more than five times faster than the projected earnings growth of the U.S. hyperscalers.
- Global Upward Revisions: Since January, the 2026 earnings-per-share (EPS) estimates for Asian, U.S., and European semiconductor names have been revised upward by 89%, 22%, and 17%, respectively, reflecting the persistent shortages of high-end processors.
- Valuation Benchmarks: Despite this massive run-up, U.S. semiconductor valuations remain relatively constructive, trading at an average of 22 times 12-month forward earnings—only one point higher than the overall valuation of the S&P 500.
While these low price-to-earnings ratios suggest that the hardware rally is backed by real, current earnings rather than purely speculative optimism, analysts warn that these valuations are highly dependent on the assumption that the hyperscalers will continue to increase their capital spending indefinitely. If the spenders freeze their budgets, these high earnings estimates will quickly collapse, exposing the extreme vulnerability of the hardware sector.
The Wider Macroeconomic and Liquidity Picture
The structural challenges of the technology sector are taking place against a highly complex macroeconomic and liquidity backdrop. While high interest rates have put pressure on corporate borrowing costs, global liquidity remains surprisingly robust, providing a supportive cushion for high-beta risk assets.
Global Money Creation Trends in 2026
According to JPMorgan’s macroeconomic analysis, U.S. money creation is on track to increase from $1.6 trillion in 2025 to $1.8 trillion in 2026. This continuous injection of liquidity into the financial system has helped support asset prices across the board, preventing a more severe market correction even as the Federal Reserve maintains its cautious, high-for-longer interest rate stance.
This abundant liquidity explains why major market indexes like the S&P 500 have managed to cross historic milestones like 7,600, as the massive volume of capital looking for returns continues to flow into high-growth sectors, masking the underlying structural divergences within individual stock classes.
Two-Way Volatility Risk in Crypto Markets
The abundant liquidity has also supported a highly volatile cryptocurrency market, though JPMorgan warns that the digital asset space has begun to face new, structural risks. The bank flagged that concentrated institutional players, such as MicroStrategy, have introduced avoidable “two-way risk” into the crypto markets through their highly leveraged acquisitions of Bitcoin.
Because these companies utilize debt and equity issuance to purchase massive stockpiles of digital tokens, they have built a highly leveraged loop that can magnify both upward and downward price movements. If the price of Bitcoin experiences a sharp decline, the forced selling or margin calls on these highly leveraged corporate balance sheets could trigger a rapid, cascading selloff, inducing more uncertainty and volatility across the entire digital asset ecosystem. This risk is closely tied to the broader tech sector, as a sudden de-risking of growth assets on Wall Street would immediately spill over into the highly correlated cryptocurrency market.
What to Watch for in the Summer of 2026
As the market enters the third quarter of the year, investors must focus closely on the individual stock charts and earnings reports of the major cloud giants to determine whether the AI bubble is heading for a soft landing or a sharp correction. The upcoming summer reporting season will serve as a critical test of whether the massive capital investments are beginning to yield tangible, commercial returns.
If companies like Microsoft and Meta can demonstrate robust revenue growth in their AI and cloud divisions, they will validate their capex budgets, allowing their stock prices to recover and reducing the risk of a “catch down” correction. However, if their reports show that AI monetization remains slow while capital expenditures continue to climb, the pressure on their share prices will likely intensify. As the autumn approaches, this persistent divergence remains the most important risk factor for global markets, proving that even the most successful technology trends must ultimately answer to the basic laws of financial returns.
Conclusion
JPMorgan Chase’s warnings regarding the historic divergence between AI hardware winners and major technology spenders represent a timely, highly analytical wake-up call for global investors. By showing how the parabolic growth of semiconductor and memory stocks stands in stark contrast to the slumping stock prices of the cloud giants funding this capex boom, the bank has exposed a structural imbalance that closely mirrors the final months of the 1999-2000 dot-com bubble. While the short-term earnings of chipmakers remain exceptionally strong, the long-term sustainability of the entire AI sector depends entirely on whether the spenders can successfully monetize their massive investments.
Whether the market achieves a positive “catch up” recovery or suffers a disruptive “catch down” correction will be decided over the coming quarters. As investors monitor corporate earnings, global liquidity conditions, and geopolitical wildcards in the Middle East, the need for extreme cost discipline and rigorous risk management has never been more apparent. In an era increasingly defined by high interest rates and technological transitions, the companies that can deliver real, profitable commercial utility—rather than just building more physical capacity—will be the ones that secure their financial futures, proving that fundamental cash flows remain the ultimate arbiter of value on Wall Street.
