# Alex Sacerdote - How to Invest Through Technology Cycles - [Invest Like the Best, EP.477]

Podcast: Invest Like the Best with Patrick O'Shaughnessy
Published: Jun 17, 2026
Reading time: 17 min
Canonical: https://podbrew.app/briefs/invest-like-the-best-with-patrick-o-shaughnessy-alex-sacerdote-how-to-invest-thr

This segment features Alex Sacerdote, founder of Whale Rock Capital Management. His firm, managing over $17 billion, has been one of the top-performing hedge funds in recent years, consistently delivering strong returns. Alex invests with a distinct framework honed over two decades.

The discussion unpacks how Sacerdote applies his investment methodology, focusing on technology S-curves, robust competitive advantages, and identifying underappreciated earnings potential. A significant portion of the talk delves into the entire AI stack, from foundational chips and models to various applications, using his highest conviction position, Anthropic, as a starting point.

Understanding Sacerdote's approach offers valuable insights into navigating the dynamic world of technology-focused investing. His perspective on identifying and capitalizing on major technology cycles, particularly the current AI revolution, provides a crucial roadmap for investors seeking to build durable portfolios in an ever-changing market.

## Key takeaways

- WhaleRock Capital's initial AI investment strategy prioritized chips and infrastructure, anticipating universal demand regardless of which foundational models ultimately became dominant.

- The foundational model market is consolidating into an oligopoly of 3-4 key players, including Anthropic, OpenAI, and Google, mirroring the competitive structure seen in the cloud computing industry.

- Agentic coding capabilities are identified as the "true unlock" for AI, creating a half-trillion-dollar market by enabling models to largely replace manual coding.

- The private 'unicorn' market holds substantial economic significance, often exceeding the market capitalization of entire public stock exchanges, necessitating specialized investment strategies.

- Exponential growth in leading companies like Nvidia, Tesla, and Apple can be anticipated by understanding the S-curve pattern, enabling investments at low P/E ratios well before mainstream recognition.

- Strategic inflection points occur when critical adoption barriers, such as price, ease of use, or infrastructure, are removed, igniting a rapid "tornado of demand" for the technology.

- Evaluating the total addressable market (TAM) or "height" of an S-curve is crucial for determining investment longevity, while early signs of exploding demand can be spotted through intuition, anecdotal evidence, and market observations.

- Digital competitive advantages like network effects, industry standards, rapid scale, critical intellectual property, and strong brand can be more powerful and enduring than traditional moats.

- AI foundational model companies like Anthropic and OpenAI are establishing significant moats through critical IP, strong enterprise branding, rapid achievement of scale, and recursive improvement loops that accelerate their innovation.

- Traditional enterprise software companies are struggling to monetize AI products, leading to reduced investment and market share shifts as CIOs prioritize AI APIs and tokens.

- The 'AI Rule of Forty' evaluates a company's strength in the AI landscape by combining its percentage of sales from AI with its market share in the AI category.

- While many software incumbents face significant headwinds, platforms with strong network effects like Slack could become more important as central hubs for AI agents, but others risk being relegated to 'headless' databases.

- AI workloads, growing tenfold annually, are pushing data center hardware to its physical limits, triggering a "de-commoditization" and renaissance in an industry previously characterized by slow innovation and commoditization.

- Significant opportunities are emerging across the hardware supply chain, including specialized high-bandwidth memory, advanced liquid cooling for AI servers, high-speed networking components with rapid upgrade cycles, and multi-layer printed circuit boards and their raw materials.

- Companies in the hardware sector are experiencing a transformation from low-margin, commodity suppliers to high-growth businesses with 35-50% annual revenue growth, rising average selling prices, and improved margins due to the specialized and innovative demands of AI infrastructure.

- Investors frequently miss tech opportunities due to discomfort with perceived high valuations, a lack of holistic market understanding, and insufficient conviction, especially when specialized views obscure broader foundational shifts.

- The rate of change in market share and adoption metrics is more critical than absolute percentages, as it signals accelerating growth and margin expansion that indicates significant underlying market momentum.

- Key risks to the AI market include potential negative regulation driven by public and governmental apprehension, a slowdown in AI model improvements that could enable open-source models to create a "race to the bottom," and the risk of major players faltering, leading to excess compute capacity.

- The AI application layer is currently considered risky due to an unclear ecosystem, making it challenging for applications to establish sustainable business models and competitive moats.

- AI functions as an augmentation tool for investment research, accelerating information gathering and improving report quality, but it does not replace the human analyst's role in providing unique insights or stock-picking judgment.

## 02:00 - 14:03 WhaleRock Capital's AI investment thesis centers on Anthropic and the foundational model layer's evolution into an oligopoly.

Alex Sacerdote's WhaleRock Capital, a technology-focused investment firm, conducted a deep dive into the AI stack following ChatGPT's release in November 2022. Recognizing a new compute paradigm, the firm initially focused on chips and infrastructure, viewing them as essential components regardless of which foundational models would ultimately succeed.

Over time, the firm gained clarity on how the foundational model layer would evolve. While sixty different companies initially competed, the market consolidated. Many startups failed, and efforts from large companies like Amazon and Meta faltered. Anthropic emerged as a dark horse, focusing on the enterprise, alongside OpenAI and Google's Gemini, forming an oligopoly akin to the cloud market's three-player structure.

A significant unlock for AI, according to Sacerdote, has been the advent of agentic coding tools. Early models were not robust, but with advancements like Anthropic's Claude code, which can generate code from English prompts, the market exploded. Some users were spending $100 a day on tokens, translating to a potential half-trillion-dollar market from coding alone, even with technology that was only months old.

Contrary to expectations of commoditization, foundational models exhibit significant differentiation and sticky ecosystems. Anthropic excels in finance, while Google is strong with PDF ingestion. These companies are building comprehensive product ecosystems around their APIs, similar to AWS's strategy, creating lock-in and competitive advantages. The enterprise AI application market is still less than one percent penetrated, indicating massive S-curve growth potential in the coming years.

> The big kicker was code, and this is the true unlock of AI.

## 14:03 - 18:04 WhaleRock's Investment Strategy in Private Tech Companies

WhaleRock initially hesitated to invest in Anthropic due to its $60 billion valuation and negative gross margins. However, after extensive engagement with CEO Dario and his team, observing their dedication and the quality of their code, WhaleRock conducted deep due diligence. This involved creating a 90-page PowerPoint deck, utilizing Claude code to scour the internet for market feedback, and building a strong relationship with Anthropic's CFO, ultimately leading to a successful investment.

The private 'unicorn' market has expanded significantly, now surpassing the size of many European stock markets, including Germany and the UK. To effectively navigate this landscape, WhaleRock conducts between 2,000 and 3,000 face-to-face meetings with management teams annually. A substantial 10-15% of these meetings are dedicated to private companies, underscoring the importance of understanding these increasingly impactful businesses.

WhaleRock's first private investment was in Stripe, a decision that stemmed from extensive due diligence on public payments company Adyen. By surveying 200 Adyen customers, they discovered Stripe's critical role, realizing the two companies were like "Coke and Pepsi." After meeting the Collison brothers, WhaleRock acquired a $100 million block of Stripe shares during COVID-19 at a $35 billion valuation. Their rigorous underwriting, based on metrics like total processing volume (TPV) and take rates, revealed Stripe was significantly undervalued, with TPV closer to $1 trillion than the disclosed $550 billion.

> We realized this is Coke and Pepsi.

## 20:04 - 30:05 Identifying exponential S-curve growth and strategic inflection points in technology adoption

Technologies often follow an S-curve pattern, experiencing a long period of slow, hidden development before reaching a visible inflection point and rapid exponential growth. Examples include smartphones, which existed for a decade before the iPhone ignited mass adoption, and electric vehicles, where Tesla spent 15 years developing before its significant growth in 2019. Understanding this pattern allows investors to predict long-term earnings and acquire companies like Nvidia, Tesla, and Apple at remarkably low price-to-earnings ratios during their early growth phases.

Exponential growth is triggered when key barriers to adoption are eliminated. For the iPhone, this meant reducing the price to $200, integrating a 3G network, and creating an intuitive touchscreen interface with an ecosystem that made it simple for anyone to use. Similarly, Tesla spurred EV adoption by lowering prices to $40,000, extending range to 300 miles, and establishing a robust supply chain. These barrier removals create a "tornado of demand" where the market recognizes an immediate need.

Beyond just identifying the takeoff, investors must also assess the "height" of the S-curve, or the total addressable market (TAM), to determine how long to hold an investment. For instance, Amazon Web Services (AWS) initially appeared as a small line item, but its TAM was re-evaluated to be significantly larger than traditional enterprise IT markets like routers and storage, especially once it was realized the service wasn't deflationary. Mega S-curves like the internet, mobile, cloud, and AI build upon one another, offering vast potential, but it's crucial to adjust expectations if a curve, like electric vehicles, hits an unexpected wall.

Identifying these strategic inflection points often relies on intuition and anecdotal evidence rather than just data. Observing market shifts, like seeing a 12-year-old in China playing advanced mobile games or witnessing packed ballrooms at IT symposiums for companies like Splunk, VMware, or AWS, can signal exploding demand before it's reflected in official metrics. It's also acceptable to miss the initial phase of growth, as a large S-curve can offer significant returns over a prolonged period.

> When you have strategic inflection points, you can't- Trust the data, and strategic inflection points are about intuition, anecdotal evidence.

## 30:05 - 38:08 Identifying Competitive Advantages and Moats in the Tech Industry

In the rapidly evolving tech landscape, understanding competitive advantages, or 'moats,' is crucial for long-term success. While many worry about constant disruption, digital-era moats can be as, or even more, powerful than traditional ones. These include network effects, where a product gains value with more users (e.g., LinkedIn, Facebook), and becoming an industry standard (e.g., Oracle, Bloomberg) which creates a chokehold through ingrained software and user bases.

Other significant competitive advantages involve achieving massive scale quickly, as seen with Amazon Web Services reaching Walmart-sized scale in just five years, or newer players like Anthropic rapidly growing. Critical intellectual property, like Qualcomm's indispensable mobile patents or ASML's lithography technology, also creates strong barriers to entry. Furthermore, a powerful brand, exemplified by Google, Amazon, and even Elon Musk's ventures, allows companies to grow without extensive advertising, reducing customer acquisition costs.

Leaders in S-curve markets often combine multiple moats. For instance, Amazon built a seven-year lead, becoming an ecosystem and platform, then achieved immense scale, making it impossible for competitors to match their R&D. The foundational model layer of AI demonstrates this, with many early contenders falling away as only two or three, like Anthropic and OpenAI, have emerged at the top by developing their own distinct competitive advantages.

For companies like Anthropic, current moats include critical intellectual property, a strong enterprise brand (where CIOs often name 'Claude' first), and achieving escape velocity through scale and fundraising. They also benefit from a recursive improvement loop, where their leading code development feeds back into their models, accelerating innovation. OpenAI similarly focuses on enterprise and coding tools, leveraging its position to foster accelerating growth, recognizing the high value customers place on solutions that replace human effort.

> What we found over the years is some of the competitive advantages within the digital world are more powerful, if not equally or more powerful than in the offline world.

## 38:07 - 44:08 AI's Disruptive Impact on Enterprise Software and the 'AI Rule of Forty'

WhaleRock has significantly reduced its software holdings, moving from 40-50% of the portfolio five years ago to being net short entering this year. Initially, there was optimism that large software companies with their sales forces, data, and APIs could integrate AI effectively. However, their AI products were not good, did not move the needle, and could not be charged for, leading to a quick divestment.

Traditional enterprise software faces disruption on multiple fronts. CIOs are shifting budgets to faster ROI AI APIs and tokens, pushing legacy software down their priority lists. This creates budget pressure and makes it difficult for software companies to raise prices annually. Furthermore, AI's impact on jobs could reduce demand for 'seats' and self-built apps, while new AI-native companies pose a threat by potentially obviating incumbents' data advantages and allowing for easier replacement.

The 'AI Rule of Forty' is introduced as a new metric to evaluate companies in the AI era. While the old rule combined growth rate and operating margin, the AI version focuses on the percentage of sales derived from AI plus the company's market share within that AI category. For many traditional software companies, AI currently accounts for only one or two percent of their sales, indicating a significant journey ahead to achieve strong AI exposure and market position.

Despite these challenges, there's a developing idea that AI could make some existing software platforms more critical. Platforms with strong network effects, like Slack, might become key repositories where AI agents operate, embedding them as permanent fixtures within organizations. However, there's also the risk that CRM systems, for example, could become 'headless,' relegated to being mere databases accessed directly by AI, with the human interface becoming less relevant.

> The old way of software is like using pen and paper or it's like a horse and buggy. The new way of software is like a jet engine or frankly the transpo- Order from Star Trek, it's so revolutionary changing that it feels like it has to be disruptive.

## 44:08 - 52:10 AI is de-commoditizing data center hardware and driving a renaissance in the industry.

For the past four decades, data center hardware, including Intel x86 chips, memory, and networking components, saw little innovation and became largely commoditized. Compute workloads grew by 25-40% annually, but Moore's Law kept pace, meaning there was minimal growth or requirement for significant hardware advancements across servers, printed circuit boards, and memory.

The emergence of AI workloads, which are growing 10x annually, is now pushing every aspect of data center hardware to its physical limits. This unprecedented demand is leading to a "de-commoditization" of the hardware industry, creating a renaissance where significant innovation is required and unit growth is tremendous across the entire supply chain.

Specific components benefiting from this shift include high-bandwidth memory, which is stacked and requires significant lead time from manufacturers like Samsung. AI servers, costing hundreds of thousands of dollars each, necessitate advanced liquid cooling solutions and robust contract manufacturing. Networking, once a seven-year upgrade cycle, now sees annual advancements with critical software layers like open-source Sonic, making it incredibly network intensive.

Even seemingly mundane components like printed circuit boards are evolving; AI servers require 40 layers compared to the typical 10. This is driving huge growth, rising ASPs, and improved gross profits for specialized suppliers, transforming them from low-margin, slow-growth businesses into high-growth entities with long-term design roadmaps and 35-50% annual revenue growth. Fiber optics for connecting data centers and eventually GPUs within racks are also experiencing huge demand for specialized, high-margin products, as are power supplies for the increased power demands of AI chips.

> Not only are you creating tremendous unit growth, but we call it the de-commoditization of the hardware industry.

## 52:10 - 58:10 Identifying Investor Blind Spots and Addressing AI's Major Risks

Many investors overlook significant tech opportunities, especially in hardware and chips, because they are uncomfortable with what appear to be high valuations and lack the conviction needed to stay invested. Specialized analysts often miss the broader picture by focusing too narrowly, failing to connect foundational model layers with wider market trends. A holistic view, experience across multiple S-curves, and understanding the rate of change in metrics like market share are crucial for identifying true growth opportunities.

A significant factor often missed is the importance of the rate of change over absolute percentages. An acceleration from 10% to 30% market share, for example, signals rapid growth and margin expansion, which is a key indicator of underlying momentum that many financial analyses might not fully capture.

Several risks threaten the AI bull case. Widespread public and governmental negativity, as evidenced by actions like Maine banning data centers, could lead to restrictive regulations. Another major concern is a potential slowdown in AI model improvements. While existing models have broad adoption potential, a plateau in innovation could allow open-source alternatives to catch up, potentially creating a "race to the bottom" for proprietary AI companies, though chip manufacturers might still benefit.

Additionally, the concentration of power among a few key AI players presents a risk. If one or two leading companies falter and lose their competitive edge, it could result in a significant surplus of computing resources. While historically, other players tend to absorb such capacity, a scenario where major entities like Meta choose to withdraw from intense AI development could pose a challenge to the overall market.

> If good enough is good enough, I won't have a business.

## 58:10 - 1:04:11 WhaleRock's Cautious Stance on AI Applications and Emphasis on Human-Centric Research

WhaleRock maintains a cautious approach to the AI application layer, noting that the ecosystem remains unclear and risky. Similar to the early days of the iPhone, it takes time for applications to develop sustainable business models and build moats. The firm has observed this struggle within the enterprise world, making it challenging to identify clear winners despite the emergence of some promising startups.

Despite advancements, AI primarily serves to augment WhaleRock's research process rather than replace human analysts. AI helps the team quickly get up to speed on complex technical areas and enhances the quality of research notes by handling reporting tasks. However, it cannot generate the crucial insights, wisdom, or conviction needed for effective stock picking.

The core of WhaleRock's research strategy is the human-driven "scuttlebutt" approach, inspired by Philip Fisher. This involves extensive engagement: meeting numerous companies, developing deep relationships with management teams, and conversing with competitors, suppliers, and customers. This hands-on method allows analysts to uncover key characteristics of leading companies and build strong investment conviction, as exemplified by their detailed research on AppLovin.

Cultivating a strong network of respected investors is another vital human element of WhaleRock's process. Sharing ideas with 10-15 like-minded individuals, as Philip Fisher suggested, helps validate investment theses. The firm calls this the "tripod" method, where conviction is strengthened when an analyst, a respected outside investor, and the firm's leader all align on a particular opportunity.

> The AI can be a great reporter, it can't pick into the future.

## 1:04:11 - 1:08:12 WhaleRock's Evolution to a Multi-Product Firm and its Compounding Research Machine

WhaleRock expanded its offerings beyond a long/short fund to include long-only and private investment products, formalizing private options around 2015 and acting on them in 2020. This evolution was driven by the observation that large-cap tech companies are significantly underweight in many institutional portfolios, which often leads to missed opportunities despite these being major drivers of market performance.

Institutional investors, such as endowments, frequently underallocate to the world's largest tech companies like Apple, Amazon, and Tesla. This often stems from a belief that large-cap stocks offer no alpha, leading them to focus on small and mid-cap managers or hold extensive private investments. WhaleRock, however, identifies 'tremendous alpha' in large-cap tech, arguing that recognizing the winners among these companies before the generalist fund managers presents a significant competitive advantage.

The firm attributes its success to its 'WhaleRock learning machine,' a research engine comprising ten highly experienced individuals. This team prioritizes compounding knowledge over two decades, conducting 2,500-3,000 face-to-face meetings with management teams annually. This consistent, in-depth research process enables WhaleRock to support its multi-product strategy, applying its insights across both public and private investments.

> We call it the WhaleRock learning machine, and it's a group of ten highly experienced individuals. In tech, you gotta go out and talk to people, so we do twenty-five hundred, three thousand face-to-face meetings with management teams.

## 1:08:11 - 1:10:12 Alex Sacerdote Details His Father's Legacy of Mentorship and Humility at Whale Rock

Alex Sacerdote recounts the kindest act done for him: his father joining Whale Rock. His father had a distinguished 41-year career at Goldman Sachs, where he ran corporate finance in the 1980s and chaired private equity in the 1990s. Despite his success, he was known for his remarkable humility and gentlemanly demeanor.

When Alex launched Whale Rock, his father offered to join as the firm's chairman, providing oversight and the voice of experience. They collaborated for six years until his father's passing in 2011. Alex values their professional relationship, noting they never raised their voices, and his father was a significant mentor to many people.

Alex's father was recognized as modest, whip smart, and wise, in addition to being a great investor. He was instrumental in keeping Goldman Sachs out of difficult situations through his role on their commitments committee. People found him approachable, and he handled both personal and work-related issues with grace, a soft manner, and a strong sense of humor.

> If I could be half the person that he is, I'd be completely winning.

---

Get podcast briefs for shows you follow: https://podbrew.app/