A look at SoftBank's controversial process of participating in multiple rounds to pump up a startup's valuation and how the WeWork debacle is changing that
In early 2018, the founders of Chinese artificial intelligence startup SenseTime Group Ltd. flew to Tokyo to see billionaire investor Masayoshi Son. Tweets: @mvolpe , @lisaabramowicz1 , and @mdudas . Thanks: @pelstrom Tweets: Mike Volpe / @mvolpe : “They pump up valuations to get higher returns to look good to investors... That kind of fundraising apparatus is essentially unicorn porn.” https://twitter.com/... Lisa Abramowicz / @lisaabramowicz1 : “WeWork is not just a mistake, it is a signal of weakness in the whole model. If you screwed up that valuation so badly, what about all of the other companies in your portfolio?” https://www.bloombergquint.com/ ... Mike Dudas / @mdudas : The @SoftBank Vision Fund blow up shows major gaps in accounting standards which allow a company to realize paper gains by setting their own ever-higher valuations on investments https://www.bloomberg.com/... Thanks: @pelstrom
Context & Ripple Effects
The piece revisits the machine behind the Vision Fund's markups: founders like SenseTime's flew to Tokyo for direct access to Masayoshi Son, whose fast, committee-bypassing deal style meant one man's conviction could set a startup's price. By repeatedly participating in successive rounds of the same companies, SoftBank effectively bid against itself to lift valuations — the practice Mike Volpe calls 'unicorn porn' and Mike Dudas ties to accounting rules that let firms book paper gains by marking their own positions higher.
The reckoning was already visible before this story ran: portfolio names like Wag and Fair were struggling, and sources said Oyo's founder had invested $2B in his own company via a loan backed by Son, per coverage of the Vision Fund's struggling portfolio. WeWork turned those warning signs into a verdict — Lisa Abramowicz's point being that if that valuation was wrong, every other markup is in question.
First-order effects
- SoftBank is abandoning or curbing its participation in multiple rounds to boost valuations, directly changing how its largest portfolio companies raise money and how their marks are set.
- Vision Fund's limited partners now face scrutiny over returns built on self-referential pricing — Dudas's argument that higher valuations let firms realize paper gains puts reported performance itself in doubt.
Second-order effects
- Rival mega-backers such as Tiger Global, which out-funded SoftBank as the top startup funder by mid-2021, compete for deals under a new norm where follow-on bids no longer auto-inflate marks — pricing discipline becomes a selling point to founders and LPs alike.
- Auditors, secondary buyers, and future-round investors inherit the problem: any startup priced across consecutive SoftBank rounds carries a markup that outside capital must independently verify rather than take on faith.
Third-order effects
- If the correction sticks, the industry moves toward external validation of private valuations — third-party-led rounds and stricter accounting treatment of unrealized gains — weakening the fund-size-as-moat model Son pioneered.
- The pattern's endpoint is already legible in later coverage: by 2022 Son himself blamed a bubble for a $23B quarterly loss at SoftBank — a bubble the same reporting says his own fund helped create — suggesting the multi-round markup strategy was structurally self-defeating rather than an isolated WeWork error.
The trend: Private-market returns are being repriced from self-set paper marks toward externally verified values, with SoftBank's post-WeWork retreat from round-pumping marking the turn.