Ex-Goldman Trader Who Decided to Build Data Centers In Space

Philip Johnston is the CEO and Co-founder of Starcloud, a YC-backed startup building data centers in space to solve AI's growing energy crisis. After working at Goldman Sachs and consulting for Middle Eastern space agencies, Philip assembled what he calls "the 10 most kickass engineers in the world" from thousands of candidates. Their mission: create gigabit-scale orbital data centers with 4km solar arrays that could fundamentally transform computing infrastructure.In this episode, Philip reveals his journey from being rejected twice by YC to securing $21M in funding and preparing to launch Starcloud's first satellite this year. He explains why his small, elite team is developing technology to run NVIDIA's H100 chips in space (the first time in human history), their rapid 18-month satellite development cycle, and why conventional terrestrial data centers simply cannot scale to meet future AI energy demands.-----------------------------------------------Chapters:00:00 Introduction02:02 Growing Up with a YC Founder Brother04:28 Why Trading Skills Are Useless for Entrepreneurship12:13 The Starbase "Holy Sh*t" Moment That Started It All16:42 The First Angel Investor Who Changed Everything21:03 Why Redmond (Not LA) Is the Real Space Capital23:47 Y Combinator & The Viral White Paper That Got Them Roasted29:29 Over 200 VCs Reached Out: The $11M Round in Days32:35 Hiring Only 10 People: The World's Most Elite Space Team37:27 First to Fly NVIDIA H100s in Space42:20 Quick Fire Questions: Why Humanity Will Wipe Itself Out47:52 Why Your Crazy Idea Probably Sucks-----------------------------------------------Takeaways:1) Exposure to high achievers eliminates the impossible - "One thing that having an identical twin brother gives you is... it teaches you what is possible. If he can do something, there is no reason I shouldn't be able to do it." 2) Traditional finance experience is worthless for founders - "I would not recommend anybody become a trader if you want to be an entrepreneur... trading is an incredibly niche and non-transferable skill set." 3) Best founder training: startup experience beats everything - "The best path to founding a company is to go work for a startup." Philip's clear hierarchy of valuable pre-founder experience places startups at the top, followed by management consulting, with VC and trading trailing far behind.4) Lead with your most audacious vision, not incremental steps - Philip's biggest fundraising regret was initially pitching short-term vision: satellite edge computing instead of his revolutionary vision for orbital data centers.5) American angels wire money immediately; Europeans wait for others - "American angels are like absolute godsend because they sign and wire on the same day... in Europe they're like, who else is in?" The cultural difference in investor behavior meant Starcloud's first check came from an American who decided independently without needing social proof.6) Move your company to the talent, not where "space companies go" - "Everybody we want to hire lives in Redmond, Washington... for satellites, Redmond's like the global capital." Philip relocated from aerospace hub El Segundo when he discovered 90% of satellites were actually designed and built in Redmond.7) Hire slowly: interview thousands to find your 10 perfect engineers - "We spent an enormous amount of time interviewing and not hiring lots of people... we went through about 10,000 CVs." Starcloud's elite 10-person team resulted from an extraordinarily selective process that rejected thousands of qualified candidates.8) "Build it and see" beats "spend months in CAD" - "We are very much a build it and see culture rather than a spend six months in CAD and see culture." Philip's team prioritizes physical prototyping and testing over extensive theoretical design, accelerating their development cycle dramatically.