This conversation features Stripe co-founder John Collison on the internet economy. He believes it's still early, only 5%-6% of the total economy, with huge growth ahead. He's optimistic about globalization, where developers from developing countries can more easily participate. Key holdings: Stripe (itself, valued at $36B, helping global entrepreneurs register US companies via Atlas), Shopify (bullish, as internet infrastructure), and Constellation Software (bullish, a rare tech conglomerate with ~500 business units).
Stripe co-founder John Collison discussed the growth of the internet economy in an interview, stating that his company’s mission is "to increase the GDP of the internet," with its latest valuation reaching $36B. Core views include: the internet economy is still in its early stages, and more infrastr
John Collison (Stripe co-founder) in conversation with Patrick O'Shaughnessy. Stripe's mission is to "increase the GDP of the internet," with a latest valuation of $36B. The core theme: the internet economy is still in its early stages (only 5%-6% of total economic activity), and the infrastructure opportunity is far from saturated. Collison also explores whether the industrial conglomerate model can be applied to the technology sector, how to measure the value of intangible assets, and the boundaries of the "no-code movement."
The most impactful judgment of the episode: John Collison argues that current Generally Accepted Accounting Principles (GAAP) systematically understate the true profitability of technology companies—because R&D spending is treated as an expense rather than capital, and there is simply no accepted method for measuring the "useful life of software." He states bluntly: "I have no patience for people who constantly complain about non-GAAP metrics—accounting standards are human inventions, and we should be able to discuss better ways."
John Collison believes that the internet economy accounts for only 5%-6% of the overall economy. Most people mistakenly think it is mature, but it is actually far from the inflection point.
Deduction and Falsification Signals: If, after the pandemic recedes, the share of the internet economy does not increase significantly (e.g., stays below 10%), then the "early stage" thesis is falsified. If the second-order effects—such as the globalization of "long-tail business"—fail to continue emerging, then Collison's optimistic assumption is invalid.
Collison argues that the tech industry lacks single-industry conglomerates like Danaher or LVMH because the VC system provides an alternative “shared platform,” and high tech valuations make it difficult to execute a “buy low” strategy.
Uncertainty: Collison says, “Stripe will become more acquisition-oriented in the future” — when the company matures and platform breadth becomes a core need. But he does not provide a timeline.
Collison argues that current GAAP accounting standards are "nearly useless" for tech companies, as they treat R&D spending as an expense rather than capital, and completely fail to measure the useful life of intangible assets like software.
Falsifiable Condition: If future accounting standards are reformed (e.g., capitalizing SaaS subscription customer acquisition costs), the "profitability" of certain tech companies could shrink significantly — this would validate Collison's view.
Collison argues that payments is fundamentally a scale business (extremely high fixed costs), but when technology paradigms shift, speed becomes the true moat.
Implication: If a paradigm-level shift occurs in the future of payments infrastructure – such as central bank digital currencies or new blockchain protocols – whether Stripe can adapt quickly will be the key test of whether its "speed culture" is genuine.
Collison outlines several key principles in Stripe's internal management practices.
Uncertainty: Collison acknowledges, "We still haven't found a good way to measure software engineering output." – Currently, Stripe extensively uses engineer surveys as a substitute, but "it's still in the early stages."
| 标的 | 嘉宾态度 | 关键数据 |
|---|---|---|
| Stripe | Self-assessment (not explicitly stated, but implied confidence in its own model) | Latest valuation not disclosed; $36B from the previous round; business covers global markets; engineering teams in Dublin, Singapore, Mexico City; Atlas helps global entrepreneurs register US companies |
| Visa | Neutral (business partner; mentions Stripe is a direct integrator) | No specific data disclosed |
| MasterCard | Neutral (mentioned alongside Visa) | No specific data disclosed |
| Amazon | Bullish (as a model of a "capital allocation technology company") | Cites: Alexa team has "thousands or even tens of thousands" of people |
| Netflix | Neutral (compared to Disney, believes the competition between the two is "more evenly matched") | No specific data disclosed |
| Disney | Neutral (compared to Netflix) | No specific data disclosed |
| Warby Parker | Bullish (as a representative of "digital-native disruptors") | No specific data disclosed |
| Shopify | Bullish (as part of "internet infrastructure") | No specific data disclosed |
| GitHub | Bullish (as "developer tool infrastructure") | No specific data disclosed |
| Salesforce | Neutral (as a case of "acquisition-driven technology company") | Mentions acquisition of desk.com turned into Service Cloud; no specific data disclosed |
| Oracle | Neutral (as a historical B2B case; references the book Softwar) | No specific data disclosed |
| Bullish on execution speed (as a positive example of "large tech companies maintaining speed") | No specific data disclosed; mentions its rapid execution of Live video strategy | |
| Neutral (as a case of "good technology execution") | No specific data disclosed; mentions OKR system originated from Intel | |
| Danaher | Neutral (as a positive example of an industrial conglomerate) | Mentions "lean manufacturing" methodology |
| LVMH | Neutral (as a case of a single-industry conglomerate) | Mentions Louis Vuitton and champagne brands |
| Berkshire Hathaway | Neutral (as a classic multi-industry conglomerate case) | No specific data disclosed |
| Constellation Software | Bullish (as the only similar industrial conglomerate case in the tech sector) | Mentions it owns approximately 500 independent business units |
| WorldCom | Risk warning (as "one of the largest companies during the internet bubble; after the bubble burst, excess fiber capacity gave rise to subsequent infrastructure") | Mentions the then-common claim that "the internet doubles every 4 months" |
| Liberty Media / John Malone | Neutral (as one of "the most successful serial acquirers") | References the book Cable Cowboy |
| Lambda School | Bullish (a remote coding education company funded by Stripe) | Mentions a 9-month course helps students transition into programming careers |
| Pioneer | Bullish (a company funded by Stripe; Collison believes it "helps people go down the rabbit hole") | No specific data disclosed |
1. "The internet economy is only 5-6% of the overall economy, not the 'mature' state most people assume." (John Collison) — Supporting evidence: The vast majority of economic activity has not yet been internetized; the S-curve is still in its early stages. Falsification signal: If this ratio does not materially increase over the next five years, the judgment is wrong.
2. "GAAP accounting standards are almost useless for technology companies — R&D spending is capital, not an expense, but accounting standards treat it as an expense." (John Collison) — Supporting evidence: The useful life of software cannot be measured; internally at Stripe, "investment-type spending" and "operational-type spending" are distinguished at the "system level." Direct corollary: If accounting standards are reformed, the reported earnings of certain technology companies could change dramatically.
3. "Board members should be like Pokémon — each with a different set of combat skills and weaknesses." (John Collison) — Supporting evidence: The composition of Stripe's board — Mike Moritz (Sequoia, early-stage strategy and recruiting), Jonathan Chadwick (former VMware/Skype CFO, responsible for CFO recruiting). Core principle: Board members should be clear about what they bring, and the company should be clear about what it needs.
4. "Economies of scale vs. speed — payments is a classic economies-of-scale business, but when technology paradigms shift, speed is the true moat." (John Collison) — Supporting evidence: Stripe is in a business with extremely high fixed costs (a global team of engineers connecting with local banking systems in each country), but when faced with technological change, being large doesn't guarantee success. Facebook's fast execution on Live video is a positive example.
5. "From the first line of code to the first customer, Stripe took only three months. A common mistake big companies make is: 300 people working for three years without ever getting market feedback." (John Collison) — Supporting evidence: For all of Stripe's new product launches, the core team was ≤10 people, typically ≤5. Methodology: Validate the market before committing resources, not the other way around.
6. "The deepest insight from Warren Buffett's concept of 'owner earnings' in his 1986 letter is the distinction between 'CapEx to maintain the status quo' and 'CapEx for growth' — technology companies, especially, must be honest with themselves: which spending is the cost of 'standing still' and which is true expansion." (John Collison) — Supporting evidence: Collison cites the early Berkshire Hathaway textile mill case — a business that perpetually needs capital just to maintain the status quo is the worst kind of business. The same applies to technology companies.
7. "Excel is one of the most underrated programming environments in the world — sequential execution mode, interleaved code and data space, single-cell scale naturally limiting complexity." (John Collison) — Supporting evidence: Collison believes that any designer of a "no-code" tool should study Excel, because Excel has already enabled countless people to master "lightweight programming." However, he is skeptical about whether "no-code can fully replace coding" — "No-code is still fundamentally doing engineering — you're just reasoning about object relationships and data flows, rather than writing code."
8. "The second-order effect of the internet economy — globalization — allows a child in India, Venezuela, or the Gaza Strip to also participate in the internet economy. This is what excites Stripe the most." (John Collison) — Supporting evidence: Stripe Atlas enables entrepreneurs worldwide to incorporate a US company; Collison himself grew up in rural Ireland and is a beneficiary of the internet.