Return on Luck

NVIDIA got lucky that deep learning happened. But they got lucky because they had spent 15 years building the platform that deep learning needed. Jim Collins's return-on-luck: the luck is the same for everyone, what differs is the preparation.

TSMC got 'lucky' that smartphones happened. But smartphones needed advanced chip manufacturing at massive scale — exactly what TSMC had spent 20 years building. The luck was universal; the preparation was unique.

Meta's 2023-2024 AI renaissance (LLaMA, AI-powered recommendations) was 'lucky' timing. But it worked only because Meta had spent a decade building the GPU infrastructure, data pipelines, and ML research teams that AI required.

Simons's NSA training in code-breaking gave him skills that turned out to be perfect for finding market patterns. Lucky that his skills transferred — but only because he spent decades developing the computational infrastructure to exploit them.

The GLP-1 weight loss effect was a lucky side effect noticed during diabetes trials. But only Novo Nordisk had 100 years of insulin manufacturing expertise, regulatory relationships, and patient data to turn that observation into a global product.

Intel was offered the iPhone chip contract and turned it down. The luck (mobile revolution) was available to Intel. The negative return came from a decade of underinvestment in mobile-appropriate chip design.

Sanchez: 10,000 baby boomers retire daily, and 70% of small businesses have no succession plan. This demographic luck is available to everyone — the return depends on who develops the skills to acquire and operate these businesses.

Bloom's Luck Surface Area formula: Doing interesting things x Telling people about them = Luck Surface Area. Each variable is within your control. The formula systematizes preparation for unpredictable lucky breaks.

Sam: everyone has access to the same AI models. The luck (AI exists) is universal. The return depends on who has deep domain expertise to apply AI to specific problems. A plumber who understands AI will make more from it than an AI researcher who doesn't understand plumbing.

Sam: Substack and Beehiiv make starting a newsletter free and easy — the 'luck' of accessible tools is universal. The return depends on who can consistently write, grow an audience, and monetize. Same luck, wildly different returns.

Sam: the 'boring business' trend of 2023-2024 was luck for small business operators who happened to have content skills. The return on that luck went to operators who had both the business expertise AND the ability to tell their story publicly.

Sam: AI makes SaaS development faster and cheaper — that luck is available to everyone. The return depends on who has deep domain knowledge of the specific industry's problems. A dentist who can code (or prompt) will build better dental software than a coder who has never been to a dentist.

Winters: Pinterest got 'lucky' that Google's algorithm valued user-generated content pages. But Pinterest prepared for that luck by building SEO-friendly page templates, adding proper metadata, and optimizing page load speeds. Same Google, different return.

Collison: Stripe launched when e-commerce was exploding but payment integration was still terrible. Lucky timing — but the Collisons had spent years as teenage entrepreneurs understanding exactly why payments were broken.

Bloom: you can't create luck, but you can increase the surface area for luck to land on. Write publicly, meet people, start projects, share your work. Each action increases the probability that lucky opportunities will find you.

Rajaram: he was lucky to join the right companies at the right times. But the return on that luck came from preparation — deep product skills, leadership capability, and the ability to operate in ambiguity that he had built over years.

Balfour: HubSpot got lucky that Google's algorithm rewarded content marketing in 2010-2015. But the return on that luck came from HubSpot's massive investment in content infrastructure, SEO expertise, and the inbound marketing framework.

Buffett frequently credits his birth in America as his greatest stroke of luck. But billions of Americans had the same luck. Buffett's return on that luck — reading every financial document he could find from age 7 — was what differentiated him.

Knight's Stanford thesis on importing Japanese running shoes was a school assignment. His trip to Japan was impulsive. His meeting with Onitsuka was unplanned. Lucky — but Knight then spent 10 years building the distribution and brand infrastructure that turned that luck into Nike.

Dyson saw industrial cyclone separators at a sawmill and wondered if the same principle could work in a vacuum cleaner. Lucky observation — but the return on that luck required 5,127 prototypes and 15 years of development.

Senra: NVIDIA's AI dominance was 'lucky' — but the luck only landed because Jensen had spent $10B+ and 15 years building CUDA. The preparation was so expensive and long-term that no competitor can replicate it.

The hosts argue that companies like Microsoft and Amazon that locked in long-term power agreements in 2022-2023, before the AI electricity crunch became obvious, will benefit enormously from luck-amplified-by-preparation.

The hosts frame the soft landing as a return-on-luck event: the Fed's policy was sound, but the favorable supply environment (healing supply chains, falling energy prices) amplified the outcome. The Fed was prepared AND got lucky.

The hosts argue that dollar dominance is a return-on-luck phenomenon: the US was lucky to be the largest economy after WWII (luck), but then designed institutions (Bretton Woods, Treasury markets, Fed system) that amplified that lucky position into durable dominance (preparation).

The hosts frame the immigration-growth link as return-on-luck: the US didn't plan for 3+ million annual immigrants, but the surge arrived at exactly the moment the economy needed workers to avoid either inflation or recession. The lucky timing amplified the policy outcome.

The team traces NVIDIA's CUDA investment back to 2007, when Jensen Huang committed resources to GPU computing software with no clear commercial application. When AI exploded in 2022, NVIDIA was the only company with 15 years of software ecosystem development.

The team traces Apple Services back to a chain of returns on luck: the iPhone succeeded partly through lucky timing (3G networks matured, capacitive touchscreens became viable), and the resulting 2B-device installed base became the platform for a $96B services business.

The team frames Uber's entire existence as a return-on-luck phenomenon: the zero-interest-rate environment of 2010-2021 made the $32B in losses possible by keeping capital cheap and abundant. In any other era, Uber would have been forced to be profitable far sooner.

The team frames Microsoft's LinkedIn acquisition as return on luck: the $26.2B price seemed high in 2016 (many analysts called it overpaying), but Nadella's preparation (clear integration strategy, letting LinkedIn operate independently, connecting it to Microsoft 365) converted the acquisition into a $100B+ asset.

Wolfe describes Lux's strategy as manufacturing luck: fund many frontier science companies (creating many chances for breakthroughs), then pour resources into the ones that achieve breakthroughs (amplifying the lucky outcome). Preparation (broad portfolio) meets luck (scientific breakthrough) to generate outsized returns.

Collison acknowledges that Stripe's timing (founded 2010, as smartphones and e-commerce were exploding) was partly lucky. But the Collison brothers had been preparing: they'd been programming since childhood, had already sold a company, and were intellectually obsessed with payment infrastructure.

Housel reframes the skill-vs-luck debate: everyone gets some luck (good and bad). The question isn't whether you were lucky, but what you did with the luck you got. Some people squander good luck; others amplify it. The return on luck is what separates great outcomes from average ones.

Marks frames the declining-rate era as a return-on-luck event for an entire generation of investors: the lucky environment (falling rates) amplified mediocre investment skill into impressive-looking returns. With rates now stable or rising, the luck has ended and only skill will produce returns.

Brown frames Snapchat's story as a return-on-luck failure: Spiegel created a brilliant feature (luck favored him), but Snapchat lacked the platform scale, advertiser relationships, and international distribution to sustain the advantage. When Instagram copied Stories, Snapchat's luck couldn't be converted into durable returns.

Brown argues that Microsoft's 'failed' $15B Bing investment was actually preparation that enabled its OpenAI partnership: the data center infrastructure, AI research team, and search experience built for Bing positioned Microsoft to be the partner OpenAI needed.

Brown traces ARM's origin: it was spun out of Acorn Computers (which failed in the PC market against IBM). Had Acorn succeeded, ARM's designs would have been proprietary to Acorn. Acorn's failure was the lucky event that freed ARM to license to everyone, creating the open architecture that now powers every smartphone.

Schmidt argues that Google's early search dominance was partly luck — they were better at the right time. But their return on that luck — investing ad revenue into infrastructure, talent, and adjacent products — made a lucky break into a permanent advantage.

NVIDIA didn't get lucky with AI — they made a massive bet (CUDA) a decade before the AI revolution, absorbed years of losses, and were positioned perfectly when the luck arrived. The return on that patient investment was the entire AI computing market.

Friedberg notes that SpaceX's success wasn't purely entrepreneurial genius — it built on 60 years of publicly funded NASA research, infrastructure, and talent. Musk's 'luck' was inheriting this knowledge base; his 'return' was commercializing it.

US export controls denied China access to NVIDIA's best GPUs. Instead of stopping Chinese AI, this constraint forced DeepSeek to innovate algorithmically — producing more efficient models. The restriction became the advantage.

The team frames Walmart's e-commerce success as return on luck: the 4,700-store footprint was built for in-store retail, not e-commerce. But the same proximity-to-customer advantage that made stores profitable in 1975 makes them ideal fulfillment centers in 2024.

The team frames low-wage workers' gains as return on luck: the pandemic-induced labor shortage (luck) created bargaining power that workers used to demand higher wages, better conditions, and more flexibility (preparation meeting opportunity).

Wolfe describes Lux's strategy as systematic luck creation: fund 40-50 frontier science companies per fund, knowing that breakthroughs are probabilistic and unpredictable. When a breakthrough occurs, pour in follow-on capital to maximize the return on that lucky event.

O'Shaughnessy connects luck surface area to the return-on-luck Canon: expanding your surface area is the preparation that creates opportunities for luck, and executing well when luck arrives generates the return.

Collins applies his Return on Luck research to personal life: lucky and unlucky breaks happen to everyone, but only the disciplined extract value from good luck and survive bad luck.

Three types of luck: what luck, who luck, and zeit luck. What matters isn't getting lucky — it's your return on the luck you get.

Stebbings argues that venture capital returns are more luck-dependent than the industry admits. The best funds weren't smarter — they met the right founders at the right time. But the best firms maximize their return on that luck through rigorous portfolio support.

Galloway notes the ironic return on luck: US export controls on GPU chips were intended to slow Chinese AI. Instead, they forced Chinese labs to innovate algorithmically, potentially producing more efficient AI architectures. The restriction became the advantage.

The FTX recovery wasn't caused by regulatory protection, legal remedies, or institutional safeguards — it was caused by a crypto bull market that happened to occur during the bankruptcy process. The 'luck' was timing; there was no 'return' that anyone earned.

Luck is not something you have or don't have—it's a skill you can develop through intentional practice and mindset shifts.