Sam Altman loses 2015 bet against a tech bubble, as 1 of the 3 portfolios failed to achieve its proposed valuation by Jan. 1, 2020, will give $100K to a charity
In 2015, Sam Altman wrote a post titled “Bubble Talk” which lamented the “boring reading” of ongoing press coverage of a bubble … Tweets: @vivekraju93 , @zachgins , @triketora , @balajis , @synopsi , @sama , @nongaap , and @sama Tweets: Vivek Raju / @vivekraju93 : Though Sam Altman lost the bet technically, he's not wrong. Tech valuations (2015) weren't a bubble. Maybe right now, few are, driven primarily by SoftBank. But they are already bursting. Wework burst. Wag burst. Technically, they weren't even tech cos http://www.daniellemorrill.com/ ... Zach Ginsburg / @zachgins : @Techmeme @DanielleMorrill The irony is if Uber was still private and Airbnb public, he may have been right on all three... Tracy Chou / @triketora : looks like sama is donating $100k to charity but the predictions weren't bad http://www.daniellemorrill.com/ ... Balaji S. Srinivasan / @balajis : Sam technically lost the bet but it was very close. He stacked the deck against himself but nailed two out of three predictions. Only the first missed, and by 15-20%. http://www.daniellemorrill.com/ ... https://twitter.com/... Rasty Turek / @synopsi : Reflects the market perfectly. VCs overestimate the hot and sexy and underestimate the “ugly duckings”. https://twitter.com/... Sam Altman / @sama : Thank you to @DanielleMorrill for writing up the results here: http://www.daniellemorrill.com/ ... Proposition 1 missed by around 15%, Proposition 2 is up 5x in 5 years, Proposition 3 went from almost nothing to way over the threshold. Mike / @nongaap : Interesting post by @DanielleMorrill re: the results of a 5-yr “tech bubble” bet made in 2015. My quick take: Winning these sort of bets often comes down to power law & one company singlehandedly driving returns. Also, BVP cloud index ~3x over 5 yrs. http://www.daniellemorrill.com/ ... https://twitter.com/... Sam Altman / @sama : I lost the bet here ( https://blog.samaltman.com/... ) from 2015 to Michael de la Maza, a TechStars mentor, and will cheerfully donate to a charity of his choice. Here are some reflections on it:
Context & Ripple Effects
In March 2015, Sam Altman answered bubble chatter not with an essay alone but with money on the line: a public wager that three tech portfolios would hit proposed valuations by Jan. 1, 2020. Andreessen Horowitz had made the same argument without stakes months later, in its case against bubble talk.
The bet has now resolved as a partial loss — one of the three portfolios missed its target, triggering a $100K donation to a charity of Michael de la Maza's choice. The scoreboard matters because it converts a five-year-old rhetorical fight into a dated, checkable outcome.
First-order effects
- Altman owes $100K to Michael de la Maza's chosen charity, and the public record now shows one of his three 2015 portfolios failing to reach its proposed valuation on schedule.
- The loss hands bubble skeptics a concrete data point: commenters like Vivek Raju read the result as 'technically lost, substantively right' — valuations broadly held, while SoftBank-adjacent names like WeWork and Wag burst separately.
Second-order effects
- The no-bubble camp's argument survives on average but not at the tails, forcing future valuation debates to distinguish index-level appreciation from individual blowups rather than arguing about 'the bubble' as one thing.
- SoftBank's role as the marginal price-setter in the late-decade deals gets elevated from footnote to explanation for the failures, pressuring how its portfolio marks are treated in bubble assessments.
Third-order effects
- Public, dated wagers become a recurring accountability device for valuation claims — the pattern suggests more investors will put stakes behind market calls, accepting that partial losses get scored publicly.
- If the broader portfolios did appreciate even as single names collapsed, the structural lesson points toward barbell markets: blockbuster winners coexisting with high-profile failures, which is exactly what the subsequent wave of VC mega-exits like Snowflake, UiPath, and Roblox would later confirm.
The trend: Tech bubble debates are shifting from op-ed exchanges to public, dated bets whose outcomes get scored in the open, splitting 'the bubble' into index-level strength versus individual blowups.