In short: The second drawdown he sat through: "Cloudflare 83%." Same mechanism as Shopify — the Quality Score and the Selling Rules are what make holding possible, and the stories rather than the price are named as the thing that actually forces people out: "when a stock is down 50% or more, everything you hear and read about it is extremely negative, and those stories influence your thinking." Disclosed entry in Part I: $39.
Cloudflare runs a global network that sits in front of websites to make them faster and to absorb attacks. Its shares fell 83% in the same 2022 sell-off, and it was held on exactly the same basis as Shopify: the quarterly quality score and the pre-written sell rules stayed intact, so the position stayed on, and was added to.
The general rule he draws from both: "when it hurts the most, it's often the best time to buy" — while admitting the honest part, that at the time it feels like throwing good money after bad. The supporting statistic is Chris Mayer's study of 365 stocks that rose a hundredfold: not one of them avoided a 50% fall, and most fell 75% or more, more than once.
In short: A Potential Multibaggers pick and disclosed holding: bought at $39, $310 today. Named again in the dollar-cost-averaging passage as one of the three positions he has "probably bought more than 50 times over all those years" — the entry price is therefore an average built over years, not a single trade. No business analysis is offered in Part I.
Cloudflare runs a global network that sits in front of websites and applications, making them faster and shielding them from attacks. Customers pay a subscription, and the same network serves every one of them, so each additional customer costs very little to add.
The guest bought at $39 against roughly $310 today. No business case is given in this half of the interview — what is disclosed is how the position was built: it is one of three names he has purchased more than fifty separate times over the years through a fixed fortnightly buying programme. The entry price quoted is therefore an average accumulated over a long period, not the price of one decision.
In short: Agentic acceleration. Q2 revenue +36% Y/Y to $696M ($30M beat), accelerating from 34% in Q1, with adjusted EPS $0.29 ($0.02 beat) and free cash flow +69% to $56M; shares jumped nearly 16% after hours "as the 20% workforce reduction showed no visible impact on growth." Large customers (>$100K ARR) grew 27% to 4,698 and now generate 73% of revenue, DBNR improved to 120% from 118%, current RPO +35%, and the Workers platform added nearly 2 million developers in Q2 alone — twice the Q1 pace and more than all of 2025. The positioning is infrastructure for the "agentic internet": non-human traffic exceeded human traffic on its network for the first time, and Cloudflare is building the tools agents need to interact and transact, including Wallets and payment rails — explicitly not competing with hyperscalers by buying GPU clusters, but supplying "the network, security, and orchestration layer around AI workloads." Adjusted gross margin rose sequentially to 73%, its first sequential increase in eight quarters (still below the 75–77% target), absorbing ~$151M of restructuring charges. FY26 revenue guidance raised again to $2.864–$2.870B and Q3 to $736–$737M, both above consensus. "The agentic-AI pivot is starting to look less like a restructuring narrative and more like a growth catalyst." A disclosed author holding.
Cloudflare sits between websites and the internet, making them faster and blocking attacks. Revenue grew 36% — faster than last quarter's 34% — and the shares jumped 16%, largely because the company had cut 20% of its staff and growth didn't suffer at all.
The strategic story is what the company calls the "agentic internet." For the first time, more traffic on Cloudflare's network came from software than from humans — AI agents fetching pages, calling services, and increasingly trying to buy things. Cloudflare is building the tools those agents need, including digital wallets and payment rails.
Crucially, it is not trying to be an AI company in the expensive sense. It isn't buying warehouses of graphics chips to compete with Amazon, Microsoft and Google. It sells the network, security and traffic-management layer that wraps around whoever's AI is running — a much less capital-hungry position in the same boom.
The supporting numbers are healthy: large customers (over $100,000 a year) up 27% and now 73% of revenue, existing customers spending 120% of what they did a year ago, and nearly 2 million developers joining its Workers platform in a single quarter — more than in all of 2025. Guidance rose again. The verdict: what looked like a cost-cutting story is turning into a growth one. The author owns it; analysis, not a recommendation.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.