Cascade: AI's Latest Phase
Artificial intelligence (AI) has moved beyond a singular investment theme and into a multi-phase economic force. What began as a breakthrough in model creation has evolved into a capital-intensive infrastructure cycle and is now entering what may prove to be its most consequential stage: broad economic diffusion/adoption.
We have been writing and speaking about AI's evolution through our "3 Cs" framework; and we continue in this update, with a notable adjustment to the final C. Framing the evolution through the lens of the "3 Cs"—Create, Catalyze, Cascade—helps distinguish where we've been, where we are, and where leadership may rotate next.
Within the initial iteration of the 3 Cs, we had Cultivate as the third C, as a way to capture the "beneficiaries" phase; but given how much attention has recently been devoted to AI-related disruption, we think the concept of Cascade encompasses the idea of both winners and losers as AI continued to gain traction.
Source: Charles Schwab.
"Old economy" is defined as sectors typically thought of as those that dominated when the U.S. was a manufacturing-driven economy: oil and gas, industrials, materials, construction, automotive, among others. Hyperscalers are major cloud service providers that operate vast computing, storage and networking resources through a distributed infrastructure of interconnected servers and software. SaaS (Software as a Service) is a cloud-based subscription service where software is accessed online rather than installed locally on machines. For illustrative purposes only.
Create: Innovation Concentrated
The Create phase represented the foundational breakthrough period: the development of large language models (LLMs), hyperscale computing buildouts, and the emergence of generative AI as a commercial product. The launch of tools like ChatGPT (which helped turn weeks' worth of our notes into a more refined report) marked the inflection point from theoretical capability to mass adoption awareness.
Market characteristics of this phase were likely unmistakable:
- Extreme index concentration
- Explosive capital spending (capex) from hyperscalers
- GPU scarcity and pricing power
- Early monetization experiments
- Narrow earnings and stock market leadership
This phase tended to be defined by capital intensity and intellectual property concentration. The data center buildout was (and in some ways, still is) at the center of that capital intensity. As you can see in the chart below, business investment in data centers soared close to an eye-watering 80% year-over-year (y/y) pace only a year after the release of ChatGPT. That led to explosive growth in computer and equipment investment, which has continued to soar.
Diversified? Tax efficient? Cost effective?
The buildout gets real
Source: Charles Schwab, Bloomberg, as of 9/30/2025.
For illustrative purposes only.
As the data center buildout gained steam, a relatively small cohort of mega-cap companies captured the lion's share of earnings revisions and multiple expansion. Stock market breadth narrowed even as headline indices advanced—a dynamic consistent with early-cycle technological revolutions. To frame it in a stock market index context, by mid-2024, the equal-weighted S&P 500 fell to its lowest relative to the cap-weighted S&P 500 since 2008.
In many ways, Create was about the birth of capability: training models, scaling compute, and building the digital architecture upon which everything else now rests.
Catalyze: From Code to Concrete
The Catalyze phase reflects second-order effects—where AI's computational demands spill into the physical economy. If Create was about silicon, Catalyze is about steel, copper, and electrical current.
AI's power intensity is reshaping:
- Data center real estate footprints
- Electricity demand trajectories
- Grid investment cycles and transmission infrastructure
- Water usage and cooling systems
Leadership has broadened beyond pure technology into Industrials, Utilities, Energy, and Materials. This phase marks AI's transition from a micro software story to a macro societal story. The buildout of data centers and power infrastructure, along with the investment in commercial equipment, has become one of the most capital-intensive investment cycles in decades—reminiscent not of past tech upgrades, but of industrial revolutions (at points, even wartime investment cycles).
Importantly, Catalyze has introduced cyclical dynamics. Capex cycles historically overshoot. Infrastructure investment accelerates, margins expand, and supply chains tighten—before eventually normalizing. The key investor question becomes: Where are we in that capex arc?
Cascade: "3 Ds" of Diffusion, Dispersion, Disruption
The Cascade phase we've apparently now entered is the most complex—and potentially the most economically transformative. Here, AI diffuses across industries, altering productivity curves and competitive moats. The winners will not simply be those who build AI, but those who apply it effectively.
Potential beneficiaries include:
- Automation platforms
- Advanced manufacturing
- Healthcare diagnostics
- Cybersecurity
- Defense technologies
- Enterprise productivity tools
At the same time, the Cascade phase introduces disruption:
- Margin compression risk in legacy Saas (software-as-a-service)
- Pricing pressure for commoditized software
- Automation of routine knowledge work
- Compression of consulting and basic content models
This is where dispersion intensifies. Not all "AI exposure" is created equal. Companies that integrate AI to enhance margins and pricing power may outperform, while those whose offerings are easily replicated by AI models may face structural headwinds. We have learned this recently, if not painfully, in market behavior.
The disruption fear has cascaded from industry-to-industry—with Software getting hit first, followed by financial/wealth management, and then trucking/logistics and commercial real estate; all based on AI's disruptive characteristics (perceived or real). In short, investors are no longer just asking, "Who benefits from AI?"—they're now asking, "Who gets destroyed by AI?" They're selling first and asking questions/doing detailed research later.
Getting disrupted
Source: Charles Schwab, Bloomberg, as of 2/13/2026.
Data indexed to 100 (base value = 1/1/2025). An index number is a figure reflecting price or quantity compared with a base value. The base value always has an index number of 100. The index number is then expressed as 100 times the ratio to the base value. S&P 500 Software is a sub-index of the S&P 500 that tracks companies within the Information Technology sector that are classified under the Software industry according to the Global Industry Classification Standard (GICS). S&P 500 Capital Markets is a sub-index of the S&P 500 that tracks companies within the Financial Services sector that are classified under the Capital Markets industry according to GICS. S&P 500 Ground Transportation is a sub-index of the S&P 500 that tracks companies within the Industrials sector that are classified under the Ground Transportation industry according to GICS. S&P 500 Real Estate Management & Development is a sub-index of the S&P 500 that tracks companies within the Real Estate sector that are classified under the Real Estate Management & Development industry according to GICS. Indexes are unmanaged, do not incur management fees, costs and expenses and cannot be invested in directly. Past performance does not guarantee future results.
Cascade represents the move from concentration to fragmentation—a theme increasingly evident across equity markets, with the progression mirroring a classic "3 Is" innovation cycle: Invention > Infrastructure > Integration.
Market evolution across the 3Cs
Disclosures
Source: Charles Schwab. For illustrative purposes only.
Somewhat ironically, the chaos in certain sectors might be the pricing in of arguably the greatest economic gains that are poised to come from AI integration: structurally faster productivity growth. This is among the more controversial points raised by proponents of any AI technology, given the "benefit" is slower growth in labor via less hiring. At the most basic level, productivity is the measure of output per worker, which means it grows when the workforce shrinks or when output exceeds the growth in the workforce.
Given we have been in a backdrop of low hiring for nearly two years—evidenced recently by the fact that only 181,000 jobs were created on a net basis in 2025—the shrinking workforce has been a key driver of better productivity lately. As you can see in the chart below, now that we are beyond the COVID-related disruptions, we can see that productivity growth is averaging close to 2% on a rolling four-quarter basis. That is stronger than what we saw in the cycle leading up to the pandemic, but not yet close to the average in the late 1990s and early 2000s (a period being used as a benchmark given the internet's role in transforming the economy).
Productivity looking healthy
Source: Charles Schwab, Bloomberg, as of 9/30/2025.
Again, it's worth noting the societal angst with stronger productivity growth. While it was a hallmark of the economic cycle from the late 1990s into the early 2000s, it was consistent with a jobless expansion after the 2001 recession ended. On the one hand, the economy continued to grow as it benefited from the integration of the internet. On the other hand, there was a dearth of hiring and the unemployment rate continued to rise well into 2003. The recovery in the labor market was short-lived: nonfarm payrolls peaked in December 2007.
We see similar risks of a jobless expansion today, but there are admittedly different forces at work in the current cycle. Most notably, the demographics of the United States (and much of the developed world) have worsened considerably as our society has continued to age and birth rates have fallen. Combining that with the significant slowdown in immigration makes for a weaker outlook for labor force growth, not to mention the fact that tepid hiring has likely been driven by geopolitical unrest and volatile policymaking. There is some room for improvements in hiring if the effective tariff rate continues to roll over and business confidence stabilizes, but stabilization shouldn't be conflated with a boom when it comes to the labor market.
Where do we stand now?
The Create phase remains influential given index weigh concentration, but ample signs of maturation are increasingly evident. Case in point is the fact that none of the Mag7 cohort of stocks is outperforming the S&P 500 index itself year-to-date. The Catalyze phase appears firmly underway as power demand projections and industrial capex accelerate. The Cascade phase, however, is likely only just beginning—and could define the next phase of market leadership. It's already defining the rotational phase in which markets are currently immersed.
For investors, this implies:
- Monitoring earnings revisions outside mega-cap tech
- Watching capex-to-revenue ratios for signs of peak infrastructure spending
- Focusing on pricing power and margin durability as AI diffuses
AI is no longer just a technology narrative—it is an economic restructuring force. The most significant opportunity—and risk—may lie not in the invention of intelligence, but in its uneven distribution.
In that sense, the 3 Cs are not sequential endpoints, but overlapping phases—and the transition from Catalyze to Cascade may prove to the defining investment rotation of this cycle.
Diversified? Tax efficient? Cost effective?
Explore more topics
This material is intended for general informational and educational purposes only. This should not be considered an individualized recommendation or personalized investment advice. The investment strategies mentioned are not suitable for everyone. Each investor needs to review an investment strategy for his or her own particular situation before making any investment decisions.
All expressions of opinion are subject to change without notice in reaction to shifting market, economic or political conditions. Data contained herein from third party providers is obtained from what are considered reliable sources. However, its accuracy, completeness or reliability cannot be guaranteed.
Past performance is no guarantee of future results.
Investing involves risk, including loss of principal.
The policy analysis provided by the Charles Schwab & Co., Inc., does not constitute and should not be interpreted as an endorsement of any political party.
Performance may be affected by risks associated with non-diversification, including investments in specific countries or sectors. Additional risks may also include, but are not limited to, investments in foreign securities, especially emerging markets, real estate investment trusts (REITs), fixed income, municipal securities including state specific municipal securities, small capitalization securities and commodities. Each individual investor should consider these risks carefully before investing in a particular security or strategy.
All corporate names and market data shown are for illustrative purposes only and are not a recommendation, offer to sell, or a solicitation of an offer to buy any security.
Diversification and asset allocation strategies do not ensure a profit and do not protect against losses in declining markets.
Source: Bloomberg Index Services Limited. BLOOMBERG® is a trademark and service mark of Bloomberg Finance L.P. and its affiliates (collectively "Bloomberg"). Bloomberg or Bloomberg's licensors own all proprietary rights in the Bloomberg Indices. Neither Bloomberg nor Bloomberg's licensors approves or endorses this material or guarantees the accuracy or completeness of any information herein, or makes any warranty, express or implied, as to the results to be obtained therefrom and, to the maximum extent allowed by law, neither shall have any liability or responsibility for injury or damages arising in connection therewith.
Indexes are unmanaged, do not incur management fees, costs, and expenses, and cannot be invested in directly. For additional information, please see schwab.com/indexdefinitions.