Skip to main content
Episode 04

Kevin Jiang: From Founding SoftBank's $100B Vision Fund to Launching Mangusta Capital at 32

Kevin Jiang·1:35:01·2025
About this episode

Kevin Jiang spent nearly a decade at SoftBank helping build the Vision Fund, the largest venture growth equity fund in the world, where he led roughly 12 deals and worked side by side with Masayoshi Son. He recounts how Masa would push to own as much as 20% of companies, why bold high-conviction core positions are what generate alpha, and how his path from Harvard to Goldman Sachs to Apollo to SoftBank was driven more by curiosity and "increasing the surface area of luck" than by a rigid plan.

Now co-founder and managing partner of Mangusta Capital, Kevin is betting on a vertical AI application thesis, backing companies like TaxGPT and Markvision that build defensible, monetizable products on top of foundational models rather than competing on hardware. He explains the firm's origin with co-founders Leonardo Del Vecchio and Tommaso Chiabra, anchored by the Luxottica family office, his "four M's" framework for evaluating deals, and why management and authentic relationships matter most at the early stage.

Masa was like, Kevin, I need you to close this deal. I want to own 20% of this company and you need to make it happen.

Kevin Jiang · 00:21:57

You've also got to make these high conviction, large, impactful investments when you see something and know something that other people don't.

Kevin Jiang · 00:27:12

You never know where an opportunity comes out and you never know who you might be working with.

Kevin Jiang · 00:58:13

Key takeaways

  • 00:05:50Kevin spent nearly 10 years at SoftBank helping build the Vision Fund and working directly with Masa, flying founders to Tokyo to negotiate deals.
  • 00:21:21He led around 12 deals at SoftBank, including a 2020 robotics fulfillment deal where Masa demanded a 20% ownership stake and told Kevin to make it happen.
  • 00:27:12A key lesson from Masa was to make high-conviction, large core investments when you know something others don't, since core positions are what create alpha.
  • 00:29:54Mangusta's thesis is vertical AI applications built on top of foundational models, targeting defensible, monetizable products for specific industries.
  • 00:36:32Kevin's career advice is to triangulate three things: what you enjoy, what you're good at, and where the world is going, aiming for at least two of three.
  • 01:07:03His Mangusta co-founders are Leonardo Del Vecchio of the Luxottica family and serial entrepreneur Tommaso Chiabra, met through a former SoftBank colleague.
  • 01:13:01Mangusta closed $50 million from its anchor investor, the Luxottica (Delvecchio) family office, and is raising a total of $75 to $100 million.
  • 01:18:31Kevin evaluates deals using a "four M's" framework: market, model, management, and money, with management mattering most at the early stage.

Chapters

  • 00:00:00Cold open
  • 00:03:54Why the name Mangusta (mongoose)
  • 00:05:50A decade building SoftBank's Vision Fund
  • 00:07:52Building authentic founder relationships
  • 00:11:49Track record: Flexport, Wiz, xAI
  • 00:20:26Working with Masayoshi Son
  • 00:21:21The 2020 robotics deal: Masa wants 20%
  • 00:27:12High-conviction bets create alpha
  • 00:29:19The vertical AI investment thesis
  • 00:30:36TaxGPT and Markvision in the portfolio
  • 00:36:32Career advice: triangulate three things
  • 00:40:03From Harvard to Goldman to Apollo
  • 00:50:30Increasing your surface area of luck
  • 01:07:03Meeting his Mangusta co-founders
  • 01:13:01Fundraising: $50M closed, raising $75-100M
  • 01:18:31The four M's framework for deals