Alex Sacerdote — How to Invest Through Technology Cycles

In one sentence

Successful investing requires identifying the specific moment when barriers to adoption are removed, triggering a "tornado" of demand along an S-curve for companies with structural moats.

Why read this (and who it's for)

This material solves the problem of being a "reporter" of news rather than a predictor of technological trajectories. It is for investors who want to move beyond surface-level metrics to identify high-conviction plays in complex cycles like AI, EVs, and cloud computing. After reading this, you will be able to map S-curves, identify hardware decommoditization, and spot the "Code" unlock before the broader market reacts.

The big idea

The central thesis is a Growth Formula:

S-curve + Competitive Advantage + Underappreciated Earnings Power = Exponential Unit Growth.

Most investors view growth as linear or erratic; Sacerdote argues that technology cycles follow a predictable S-curve. When barriers to adoption—such as price, technical capability, or supply chain constraints—are removed, demand becomes vertical (a "tornado"). If you can identify the height of that curve by determining the Total Addressable Market (TAM) and find a company with a durable moat (like platform status or critical IP), you can capture high-quality companies at extremely low PEs because their earnings grow from $1 to $10 far more frequently than the market expects.

The logic hinges on identifying "inflection points." By recognizing when a technology moves from a niche tool to an industry standard, you can position yourself before the "sell side" catches up. This requires moving from "left brain" analysis (pure numbers) to "right brain" investing—using intuition and anecdotal evidence to connect dots and predict future cycles.

Key ideas

S-Curve Dynamics & The "Tornado" Effect

What it is — Technology adoption follows an S-curve where growth remains flat until barriers to adoption are removed, at which point demand becomes vertical. Why it matters — Identifying the removal of these barriers allows you to predict a "tornado" of demand before it shows up in standard financial reports. In the book — The author cites EVs: Tesla hit a vertical inflection in 2011 after hitting specific targets for price ($40k), range (300 miles), and supply chain scale.

The AI Stack Hierarchy

What it is — Value flows through a hierarchy: Power $\rightarrow$ Chips $\rightarrow$ Clouds $\rightarrow$ Foundational Models $\rightarrow$ Applications. Why it matters — Infrastructure layers are currently high-conviction plays because winners are clearer and they capture demand first, even though penetration is still low (~10%). In the book — The author notes that while applications are often where market cap sits, the "moat" for apps remains unclear as the ecosystem forms, making infrastructure a more reliable conviction play.

Hardware Decommoditization

What it is — For 40 years, data center hardware followed Moore’s Law (commodity growth). AI workloads grow 10x/year, pushing physical limits and causing "decommoditization." Why it matters — This shifts the investment thesis from "cheap components" to "critical supply chain winners" who provide liquid cooling, high-layer PCBs, or specialized power. In the book — Examples include Celestitas (sole supplier of Google TPU server), Elite Materials (copper clad laminate for 40-layer PCBs), and Corning (high share of fiber).

The "Code" Unlock

What it is — A shift from simple grammar/bug fixing to agentic coding tools, targeting a ~$0.5 trillion market of ~20 million coders. Why it matters — This represents a massive usability unlock where high usage ($100/day per user) and superior PDF ingestion create a "panoply" of software lock-in. In the book — Anthropic is highlighted as a "dark horse" focusing on enterprise, building an ecosystem (SDK, Cloud for Cowork) similar to how AWS was initially seen as commodity servers in 2013.

Alpha Identification in Large Cap

What it is — There is significant alpha in large-cap stocks because it takes only one person to identify a winner (e.g., Google), whereas it takes 100 diversified PMs to realize a company is not a loser. Why it matters — Many endowments are underweight on the world's largest tech companies due to a false belief that there is no alpha in large cap. In the book — Sacerdote argues that maintaining conviction during multi-year growth (like NVIDIA) is difficult because "scary" charts trigger bubble fears, but success requires a holistic view of the foundational model layer.

Rate of Change vs. Absolute Metrics

What it is — The rate of change in market share (e.g., moving from 10% to 30%) is more critical than the absolute percentage itself. Why it matters — As market share moves into that 10-to-30% window, growth rates and margins accelerate significantly because the "sell side" hasn't caught up yet. In the book — The author notes that while the world doesn't think exponentially, identifying these shifts in rate of change is the key to capturing the S-curve.

Frameworks & tools

The Growth Formula

Use this to screen for "Exponential Unit Growth":

1. Identify S-Curve: Find a technology where barriers (price/tech) are falling.
2. Assess Competitive Advantage: Does it have Network Effects, Industry Standards, Scale, Platform Status, Critical IP, or Brand?
3. Evaluate Earnings Power: Is the earnings power underappreciated? Can it grow from $1 to $10 faster than expected?

The Tripod (Conviction Building)

Align three perspectives to build high-conviction:

1. The Investor's View: Your own thesis and "right brain" connections.
2. The Analyst's View: The data, numbers, and "left brain" metrics.
3. The Respected Peer's View: What do the smartest people in the space think?

Scuttlebutt Approach (Philip Fisher)

Identify leading companies via direct engagement:

1. Talk to suppliers.
2. Talk to customers.
3. Talk to competitors.
4. Identify key characteristics of winners through these face-to-face interactions.

Adoption Models

Use these to predict the "Slope" of an S-curve:

• Radio Model: Fast adoption (reached ~100% penetration in 7 years).
• Dishwasher Model: Slower adoption because it must be "plugged into the back end" (B2B systems, security concerns).

The New Rule of 40 (AI)

Measure software/AI health by:

% of sales from AI + % market share in that category.

Memorable examples & stories

AWS as a Hidden Line Item: Initially, AWS was a hidden line item in retail; WhaleRock identified it as addressing $600 billion of IT systems. Lesson: Look for "hidden" infrastructure lines.

The EV Inflection (2019): Tesla hit a vertical inflection point only after hitting three specific targets: price ($40k), range (300 miles), and supply chain scale. Lesson: S-curves require specific barrier removals to go vertical.

Audion Case Study: The firm underwrote based on 25–30 bps take rate vs. Stripe’s 40–50 bps. Actual results were higher than assumptions, allowing for a $100 million block upsize from the seller. Lesson: Deep due diligence on customer feedback reveals actual earnings power.

The "Standing Room Only" Signal: Identifying demand by looking for physical cues, like "standing room only" at major symposiums (e.g., Splunk, VMware 30 years ago). Lesson: Physical demand signals often precede financial data.

Key quotes

"When the gun went off with OpenAI, ChatGPT in November 2022..." Why it matters: Marks the beginning of a new S-curve where "the world doesn't think exponentially."

"Going from 80 to 85 is a huge unlock." (paraphrased) Why it matters: Small improvements in accuracy for foundational models can lead to massive market shifts.

"The old way of software is like using a pen and paper... the new way of software is like a jet engine." (paraphrased) Why it matters: Highlights the shift from incremental tools to agentic, high-power "transporter" software.

"Decommoditization of the hardware industry." Why it matters: Signals the end of Moore's Law as the primary driver for data center growth.

How to apply it

This week

1. Audit your portfolio: Identify one company that is currently in a "slow" part of an S-curve and identify what barrier (price, tech, or supply chain) needs to fall for it to go vertical.
2. Identify a "Radar" signal: Find a symposium or community related to a tech cycle you follow; look for "standing room only" cues.
3. Apply the Tripod: Pick one high-conviction idea and write down your view, an analyst's likely counter-view, and what a respected peer would say.

Ongoing

• Scuttlebutt Habit: Whenever evaluating a private or public company, prioritize direct engagement with customers/suppliers over reading news reports.
• Monitor the "Code" Unlock: Track how many companies are moving from simple wrappers to agentic coding tools.
• Track Rate of Change: Instead of asking "how much market share do they have?", ask "how fast is that market share moving?"

Common pitfalls

• Being a Reporter: Many investors simply report on what happened yesterday (news) rather than predicting the trajectory of the S-curve.
• Ignoring Hardware Decommoditization: Assuming data center hardware will always be a commodity because it was for 40 years is a trap; AI workloads change the physics of the industry.
• Missing the Exit Point: Failing to sell when penetration hits 30–40% allows investors to get caught in the "sell side" catch-up, where exponential growth slows and margins compress.

Connects to

• Philip Fisher (Scuttlebutt): Extends the idea of direct engagement with the value chain to identify moats.
• S-Curve Theory: The foundational map for predicting tech cycles and identifying inflection points.
• Network Effects: One of the primary "moats" used to identify durable competitive advantages in the software layer.

If you remember only three things

• The Growth Formula: S-curve + Competitive Advantage + Underappreciated Earnings Power = Exponential Unit Growth.
• Barrier Removal: Demand becomes a "tornado" only when specific barriers (price, tech, supply chain) fall.
• Rate of Change: The speed at which market share moves from 10% to 30% is the primary driver of accelerated margins and growth.

Internalize it

• Reflection questions

  1. In my current portfolio, which companies are "reporters" of news and which are "predictors" of S-curves?
  2. What specific barrier (price, tech, or supply chain) is currently preventing a technology I believe in from going vertical?
  3. Am I looking at absolute market share, or am I looking at the rate of change of that share?

• Flashcards Q: What is the "Growth Formula" for identifying exponential unit growth? A: S-curve + Competitive Advantage + Underappreciated Earnings Power. Q: When does demand typically become a "tornado"? A: When barriers to adoption (price, technical capability, supply chain) are removed. Q: What is the "Code" unlock in the AI stack? A: A shift from grammar/bug fixing to agentic coding tools ($0.5 trillion market). Q: Why is hardware currently "decommoditizing"? A: Because AI workloads grow 10x/year, pushing physical limits beyond Moore's Law growth rates. Q: What is the "Tripod" method for building conviction? A: Aligning three perspectives: The investor's view, the analyst's view, and a respected peer's view. Q: At what penetration level should one generally consider an exit strategy? A: Generally when penetration hits 30–40% as exponential growth slows.

Fidelity note

Confidence is high. The summary captures the core investment framework (S-curves, moats, and earnings power), the specific AI stack hierarchy, and the hardware decommoditization thesis. Note that "Rule of 40" specifics are distinguished between software and AI as per the notes provided.