AI is not just the latest tech buzzword.
The ai supercycle is a once-in-a-generation long-term trend that is changing how companies make products, how economies grow, and how investors think about long-term compounding. Every ten to twenty years, the market hits a major inflection point in technology. The personal computer boom, the internet age, and the smartphone revolution created trillions of dollars of wealth for investors who could spot the long-term structural shifts. This time around, the ai supercycle dwarfs all of them combined.
In this post, I want to take you through what’s happening under the hood and how to think through positioning. I plan to keep this grounded, free of buzzwords, and as useful as possible for long-term investors who care about understanding value creation, not chasing the loudest headlines.
Understanding the AI Supercycle
The ai supercycle is a multi-year period of technological and economic expansion driven by advancements in artificial intelligence. The difference between this cycle and the hype cycles of the past is that this is being driven not just by expectations and narratives but by real dollars of revenue, hundreds of billions of dollars of capex, and a wave of enterprise adoption.
Some of the key structural trends and forces driving this expansion include:
Exponential compute improvements and GPU supply.
AI accelerator companies like Nvidia, AMD, plus a host of smaller and deeper-in-the-stack companies are shipping hardware with compute capabilities doubling every year that unlocks new classes of software applications.
Enterprise capex surge.
Companies are not just dabbling or testing out new AI tools. We are seeing real workflow integration, AI optimizations of supply chains, customer support, marketing, sales, logistics, and new products.
Productivity improvements that compound.
Every industry, one way or another, will automate the manual, repetitive aspects of knowledge work. This will compound over time and expand the total addressable market for ai tools.
Infrastructure buildout.
Companies like AWS, Google Cloud, and Microsoft Azure are spending tens of billions of dollars a year on ai-powered data centers. These are long-duration capex bets.
New business models.
Ai agents, copilots, autonomous robotics, foundational models, edge applications — these all together create an economic moat around products and services that have almost-unlimited demand.
This is not a short-term boom. It is not speculative or frothy. It is closer to what the early 2000s looked like in cloud computing, except it is happening faster and in more industries.
How early is the AI Supercycle?
Despite many ai companies rallying hard this year, we are still in the infrastructure building phase of the ai supercycle, which is similar to when cloud companies were still building out data centers before SaaS took off.
It helps to think of the AI wave as four interrelated layers:
Layer 1: Infrastructure Layer
GPU hardware, data centers, networking and edge infrastructure, cooling systems, power supply, semiconductors, and hyperscaler capex. Revenue and growth rates are highest in this layer because everything above it depends on it.
Layer 2: Model Layer
Model companies like OpenAI, Anthropic, Meta, plus open-source communities, are training more and more powerful ai models with regular inflection points. In 2026 and beyond, we are likely to see cheaper, more powerful, and more specialized models that open up new types of users and uses cases.
Layer 3: Application Layer
This is where most of the end-user value and revenue growth will be created. Productivity tools and ai agents, vertical SaaS applications, design software, robotic orchestration, and literally thousands of new workflow types and products will be built on top of the models. We are still very much in the early innings.
Layer 4: Automation Layer
Ai-native software companies will begin to replace entire stacks of human work. This includes self-driving fleets, warehouse robots, autonomous agents, and even new industries that do not even exist yet.
It’s no surprise that most of the investment dollars and public attention is focused on layer one. But historically, the largest long-term compounding and re-rating happens at layers three and four.
How to benefit from the ai supercycle
There is no single way to invest in the ai boom. The problem is so large and touches so many parts of the economy that you need to think about it strategically from a number of different angles. Each strategy has different risk, volatility, and potential returns.
1. Semiconductors and Hardware: Picks and Shovels
In any technology gold rush, the companies that sell the picks and shovels always benefit the most. Think of this as the raw building blocks and physical hardware that power the entire ai ecosystem.
The obvious leader is Nvidia with incredible software ecosystem lock-in. AMD, Broadcom, ASML, TSMC, and even memory companies like Micron and SK Hynix should all continue to see tailwinds for at least several more years.
Key Point: Every new model will require more compute to run. Hyperscalers are not slowing down on capex. The buildout here is so massive that we are more likely to see supply constraints, not demand constraints, in the coming years.
Risk to Watch: Eventually, supply will catch up. As an investor, you need to revisit valuations periodically and look for signs of normalization or normalization of expectations.
2. Hyperscalers
Microsoft, Amazon, and Google are not only the backbone of ai deployment. They are innovating in this space by increasing GPU utilization, signing multi-year ai contracts with customers, and building vertical ai platforms for developers.
Key Point: These companies have enormous distribution power, embedded enterprise relationships, and long-term incentives to monetize ai in many different ways — whether it is higher cloud workloads, more SaaS subscriptions, or new ai-native tools.
Risk to Watch: Competition is cutthroat. Margins may fluctuate as these companies scale their ai infrastructure.
3. AI-Enabled Software
These are not companies that just sprinkle ai on top of an existing product. These are companies where ai fundamentally changes the unit economics or the product’s capabilities.
Categories of companies to watch:
- Ai copilots for coding, finance, legal work, customer support, and design.
- Vertical SaaS companies building domain-specific ai models or workflows
- Workflow automation companies building on ai agents
- Creative and design tools powered by generative ai models
- Cybersecurity companies with ai-powered threat detection
Key Point: This layer has the highest long-term revenue growth once infrastructure normalizes and becomes more stable.
Risk to Watch: There will be many companies that overpromise. You want to look for ones with durable distribution, network effects, and real monetization.
4. Robotics and Automation
Robotics, especially those using vision models and autonomous ai agents, stand to gain from the next wave of “intelligence” being put into hardware. Manufacturing, warehousing, agriculture, and logistics will see rapid adoption. This is particularly interesting when combined with the internet of things (IoT).
Key Point: The quality of ai models is finally good enough to make autonomous systems practical and scalable.
Risk to Watch: Hardware cycles are longer and more capital intensive.
5. AI Energy and Power Grid Beneficiaries
Ai data centers require enormous amounts of electricity to run. Utilities, clean energy companies, and grid infrastructure companies may quietly become the next big winners as power demand explodes.
Key Point: The ai supercycle is just as dependent on power availability as it is on GPU availability.
Risk to Watch: Regulatory timelines and shifts in energy policy can impact growth.
Building a Balanced AI Portfolio
You do not need to chase hype or try to pick the next Nvidia as the single best company. A more thoughtful portfolio has a mix of exposures.
Strategy 1: Core and Satellite
Core: Diversified hyperscalers and semiconductors. Think of these as core, long-term stable compounding plays.
Satellite: Smaller software and robotics plays with high upside but higher volatility.
Strategy 2: Dollar-Cost Averaging
The ai supercycle will play out over several years. Dca into your highest conviction ideas. Minimizes timing risk.
Strategy 3: “Optionality Basket”
Set aside a small portion of your portfolio for early-stage ai companies that have the potential to 5x or 10x if they succeed. These could include:
- Ai agent platforms
- GenAi creative content tools
- Automation startups
- Robotics and edge ai
- Most will fail, but a few could generate outsized returns.
Strategy 4: Common Investor Mistakes to Avoid
Investors make a number of predictable mistakes when navigating the ai boom:
- Hype first, fundamentals second.
- Focus on revenue, not just demos.
- Don’t ignore compute cost.
- Some software companies will never have good margins because their model inference costs do not scale.
- Short-termism.
- Market pullbacks are natural. Long-term compounding requires patience.
- Biannual re-thinks.
- Ai moves fast. A company that was a “winner” today may be a loser tomorrow.
The ai supercycle for the next decade
The biggest mistake investors make is thinking about how long structural technology cycles can last. Cloud computing had over ten years of hyper-growth. Smartphones had close to two decades of secular growth. The ai supercycle could last even longer for a number of reasons:
- Touches every industry.
- Fundamentally changes productivity.
- Enables new industries that do not even exist today.
- Keeps improving year after year with compounding improvements in models.
- The demand trend underlying the ai supercycle is extraordinarily strong even if valuations fluctuate.
AI Beyond Investing
The ai era is not just for investors. There are practical ways for individuals to improve their earning power and career durability.
Learn ai tools
Whether you are in marketing, finance, design, programming, or operations, there are ai tools that will massively increase your output. Being “ai fluent” will become a baseline skillset within a few years. Today is the time to get started.
Build small ai projects
One of the best ways to learn about ai is to build with the smaller tools that are being developed. This could include:
- Ai-powered spreadsheets
- Chatbots
- Simple agents or automations
- Data analysis workflows
- Ai content systems
Even small projects build intuition and experience that is incredibly valuable.
Leverage ai to start a side business
- Ai makes it easier to start small side businesses in categories like:
- Micro SaaS
- Niche content sites
- Design and creative studios
- Data consulting
- E-commerce stores with ai-powered workflows
The total cost of starting businesses has never been lower.
Focus on skills that ai amplifies
Instead of competing directly with ai, focus on skills that are hard for ai to replicate but are amplified by ai:
- Strategy
- Product thinking
- Writing
- Leadership
- Decision making
- Relationship building
Ai does not replace any of these things. It amplifies them.
Conclusion: The AI Supercycle Will Be This Decade’s Defining Investment Theme
Zoom out far enough and you will realize that the entire economy is reorganizing around intelligence as a new form of capital. Companies that embrace ai will grow faster, run leaner, and build products that were impossible a few years ago. The companies that supply the hardware, models, and cloud infrastructure behind this will enjoy a decade of unprecedented demand and economic expansion.
Investors that treat this as a long-term secular trend instead of a speculative bubble stand to gain the most. Build a thoughtful portfolio, diversify across different parts of the ai stack, and keep your time horizon long. The ai supercycle is just getting started, and the opportunities will only get larger in the years ahead.
Leave a Comment