The Magic S-Curve Trick That Let Smart Investors Buy Wonderful Companies For a Bargain – And How You Can Spot the Next Big Opportunity

The Magic S-Curve Trick That Let Smart Investors Buy Wonderful Companies For a Bargain – And How You Can Spot the Next Big Opportunity

In a recent wide-ranging conversation on "Invest Like The Best" podcast, Alex Sacerdote of Whale Rock Capital laid out a clear framework for spotting technology winners. His firm’s approach rests on three pillars: the S-curve of adoption, durable competitive advantages, and earnings power that the market has not yet priced in. He called the true enterprise application layer of AI an “L-curve” – nearly vertical from a base of well under 1% penetration. His highest-conviction position sits with Anthropic, the company behind Claude and its coding tools that have shown early explosive traction in agentic workflows.

Most investors do not think this way. They extrapolate the next quarter or the next year in a straight line. That linear habit is precisely why S-curves matter for anyone who wants to improve their results in technology investing.

What an S-Curve Actually Is

Technology adoption rarely follows a straight line. It starts slow while a small group of tinkerers and early users experiment. Barriers such as high cost, poor usability, missing infrastructure, or lack of complementary products keep the majority away. Once those barriers fall, adoption accelerates sharply. Growth compounds as more users attract more developers, more data improves the product, and network effects or switching costs lock in customers. Eventually the curve flattens as the market saturates or a newer wave arrives.

Classic examples include the smartphone. Early devices were clunky and expensive with weak networks. Apple removed the main frictions – price, touch interface, app ecosystem, and carrier support – and the curve went vertical. Cloud computing followed a similar path inside Amazon; AWS began as a hidden line item and only later revealed itself as a massive new platform. Electric vehicles showed early promise but have faced real resistance around 10-15% penetration in several key markets due to price, charging infrastructure, and consumer habits.

The pattern repeats because human organizations and supply chains change slowly at first, then all at once once the economics and usability cross a threshold.

Why Serious Investors Pay Attention to the Shape of the Curve

When you catch the right segment of the S-curve, unit volumes can grow exponentially rather than linearly. If the business model is strong – recurring revenue, high incremental margins, pricing power – those units translate into earnings that compound faster than most models assume. Sacerdote has pointed to concrete cases: his team effectively bought Nvidia in 2023 at roughly 4 times projected earnings power a few years forward, Tesla in 2019 at about 5 times for the vehicle curve, and Amazon’s AWS exposure early on at what felt like almost no cost relative to the eventual cash flow.

The market rarely prices these inflections correctly in real time. Wall Street focuses on this year’s numbers and next year’s consensus. Very few analysts build detailed models three or four years out that incorporate falling barriers and accelerating adoption. That gap creates opportunity for those willing to do the work.

How to Use the Framework in Practice

Start by identifying genuine platform shifts, not incremental improvements. In the current cycle the stack breaks into layers: power and chips at the base, cloud infrastructure, foundational models, and applications on top. Demand for compute is broad and relatively certain regardless of which model ultimately leads. That is why infrastructure names often move first and why shortages in advanced GPUs and power have already appeared, pushing spot prices higher.

Next, assess competitive advantage. In foundational models the market is consolidating toward a small number of leaders rather than becoming pure commodity. Differentiation exists in specific domains – coding and agentic workflows for one player, document understanding for another. These skills are hard to replicate quickly because they rest on proprietary training methods, data feedback loops, and scale.

Then ask whether earnings power is underappreciated. Look for businesses that can fund their own growth from cash flow, maintain or expand margins as they scale, and operate with low maintenance capital needs relative to cash generation. Track the key operating metrics the industry actually watches: for AI these include token consumption, coding tool adoption rates, enterprise movement from pilot to production, and compute utilization.

Private access adds another layer. Sacerdote’s team built conviction in Anthropic through deep primary research, including a 90-page deck assembled with the help of the company’s own tools, and by developing direct relationships. Retail investors rarely get equivalent allocations at early valuations; recent rounds have taken Anthropic to post-money levels of $380 billion then $965 billion in rapid succession. That reality forces most individuals to use public proxies or to wait for clearer entry points.

Current AI Context: Real Progress, Real Friction

The excitement around coding agents and agentic AI has clear early roots. Anthropic’s Claude tools, particularly in coding and agentic workflows, have demonstrated rapid internal adoption and strong early external traction in finance and software development. OpenAI continues to lead in consumer-facing applications while pushing enterprise versions, and Google has invested heavily in Gemini across search, cloud, and productivity tools. Nvidia has captured the majority of the GPU supply for both training and inference, with demand driving higher spot prices and clear revenue growth in its data center segment.

Yet adoption data shows a more measured picture. While worker-level generative AI use has grown, many enterprises remain in pilot or experimentation stages. Integration challenges, skills gaps, data quality issues, and questions about measurable returns persist for a large share of organizations. Full labor replacement at scale faces higher hurdles than headline demos suggest. Power and cooling constraints on Earth are real, and downstream companies building massive clusters face heavy capital expenditure and depreciation schedules that will pressure returns if utilization or pricing does not keep pace.

Beyond AI: Other Potential S-Curves Worth Watching

AI itself contains sub-curves. The move from general chat interfaces to specialized agentic systems that execute multi-step workflows is one visible acceleration, particularly in software development and certain enterprise tasks. Vertical applications in specific industries may follow once integration costs drop and reliability improves. Hardware layers beneath the models – new chip architectures, advanced packaging, and power delivery – are also shifting as demand for efficiency grows. Public companies such as Nvidia, AMD, and Broadcom are directly exposed to these dynamics, though competition and cyclical capex patterns introduce volatility.

Looking further out, several other areas show early characteristics of potential S-curves, though most sit firmly in the slow initial phase with substantial technical, economic, and regulatory barriers still ahead. The same disciplined questions apply: What concrete problem does this solve at meaningful scale? Are the primary barriers falling in measurable ways? Is there evidence of defensible positioning or cash-flow potential forming, or mostly technical promise and narrative?

Humanoid robotics stands out as one of the more plausible near-to-medium term candidates. Tesla’s Optimus program targets factory deployment first, with ambitions for broader use later. Other players such as Figure (backed by major tech investors) and Boston Dynamics (under Hyundai) are advancing dexterity and AI integration. Cost curves are improving, and labor shortages provide a tailwind in manufacturing and services. China is investing aggressively at scale. Real-world reliability, safety certification, and integration into existing workflows remain significant hurdles. Public investors can gain exposure through Tesla, though the robotics business is still small relative to the company’s other segments and timelines for meaningful revenue contribution are uncertain.

Quantum computing is advancing from pure research toward selective early commercial use. Publicly traded companies such as IonQ and Rigetti are among those targeting hybrid quantum-classical systems and early industrial pilots in optimization and simulation for manufacturing, logistics, and pharmaceuticals. Some roadmaps point to first application-specific commercial advantage in the late 2020s, with better error correction and hybrid approaches as near-term milestones. It will not replace classical computing broadly anytime soon. The realistic path is narrow, high-value niches rather than general-purpose disruption. Demand concrete evidence of paying customers achieving measurable advantages before assuming rapid scaling.

Space infrastructure has clearer momentum in parts. SpaceX’s Starship developments are lowering launch costs and increasing cadence, which benefits satellite constellations and potential future orbital activities. Concepts for orbital data centers are being explored by various players, including ideas tied to SpaceX and Blue Origin ecosystems, as well as specialized startups testing storage and compute off-Earth (such as Lonestar Data Holdings with missions to the ISS and lunar surface). Advantages include abundant solar power and radiative cooling for AI workloads. Major hurdles remain in launch economics at scale, on-orbit power and thermal management, data latency, orbital debris, and overall cost competitiveness versus terrestrial alternatives. Most serious assessments see any meaningful complementary capacity arriving in the 2030s or later. Public proxies are limited; broader space economy exposure can come through suppliers or related communications companies, but pure orbital compute plays are mostly private and highly speculative at this stage.

Other areas with longer or more uncertain timelines include breakthroughs in energy (fusion efforts at companies such as Commonwealth Fusion Systems continue to progress on milestones, while small modular reactors see steadier deployment interest) and advanced energy storage. These often act as enablers for other curves rather than standalone explosive platforms in the near term.

Across all of them, history shows that technical feasibility rarely equals rapid economic adoption. Many promising curves plateau early or take far longer than initial projections. The framework helps by forcing attention to measurable barrier reduction and real usage data instead of excitement alone.

The Benefits Are Real – But So Are the Drawbacks

Done well, the S-curve lens gives you a multi-year roadmap instead of quarter-to-quarter noise. It helps you stay calm when short-term results disappoint and gives you a framework for adding on weakness. It also highlights when a story has already priced in too much optimism.

The drawbacks are significant and worth stating plainly. Many candidate curves never reach the steep part. Timing errors hurt in both directions: buying too early ties up capital while the business burns cash or fights for survival; buying after the inflection has already been widely recognized often means paying a full or premium price. Hype cycles inflate valuations ahead of actual cash flow, creating drawdowns when reality catches up. Private valuations in hot areas can reach extraordinary levels quickly, concentrating risk for those who gain access.

Where to Begin If You Want to Apply This Yourself

Pick one narrow area you already understand or are willing to study deeply. Technology investors often focus on semiconductors, software platforms, or specific vertical applications. Read the actual 10-K and 10-Q filings rather than summaries. Listen to earnings calls and note what management emphasizes versus what sell-side models assume. Compare operating metrics across direct competitors instead of relying on headline growth rates.

Build a simple habit of tracking the few numbers that truly move the needle for that industry. For AI-related businesses these might include forward-looking commentary on token or compute demand, customer expansion from pilot to scaled deployment, and cash generation relative to growth spending. Cross-check narrative claims against primary data from government statistical agencies or reputable industry surveys.

Stay skeptical of both extreme optimism and reflexive dismissal. The best operators in this space combine long-term conviction with continuous updating as new information arrives. They also recognize that not every exciting technology becomes a durable, high-return business.

A Final Realistic Note

S-curve thinking helps separate durable platform shifts from temporary excitement. It rewards those who do the hard work of understanding adoption dynamics, competitive position, and realistic earnings trajectories several years forward. It does not remove risk, eliminate volatility, or guarantee outcomes. Markets can stay irrational longer than expected, and even well-researched theses can be derailed by execution missteps, regulation, or faster-than-anticipated competition.

Technology investing rewards patience and realism more than excitement. The S-curve is simply one clear lens for bringing both qualities to the table. Use it carefully, update it constantly, and never confuse a compelling story with a guaranteed result.

Resources & References

Note for readers: Always cross-reference the latest company filings (10-K, 10-Q), earnings transcripts, and official regulatory sources before making any investment decisions. This article is for educational purposes only and is not financial advice.