Stableton

$16.8B. $90B. $30B. Private tech put up big numbers again this week: • Anthropic told investors its third-quarter revenue rose to $16.8B, and that it was profitable on an adjusted basis for a second quarter in a row. • Anduril opened a tender offer that lets employees sell shares at a valuation near $90B. In May, it was $61B. • Applied Intuition is reported to have received offers at $30B or more, double its last official valuation of $15B. The full stories are in this week’s newsletter. 👇 Find this helpful? [ 𝗥𝗲𝗽𝗼𝘀𝘁 ] Anything to add about this subject? [ 𝗖𝗼𝗺𝗺𝗲𝗻𝘁 ] Nice newsletter, Andreas. Next! [ 𝗟𝗶𝗸𝗲 ] Want to stay close to private blue-chip tech? [ 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝘁𝗼 𝗺𝘆 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 𝗯𝗲𝗹𝗼𝘄 ]

Private blue-chip tech companies are funding growth in three different ways this week. ➡️ Ramp is said to have raised about $1.85B from investors at a $60B pre-money valuation, up from $44B in June. ➡️ Anduril is spending $3.7B of its own capital on a Maryland shipyard, backed by a Navy contract worth up to $2.9B. ➡️ Figure AI is said to be in talks with Nvidia, an existing investor, about another $1B. The full stories are in this week's newsletter. 👇 Find this helpful? [ 𝗥𝗲𝗽𝗼𝘀𝘁 ] Anything to add about this subject? [ 𝗖𝗼𝗺𝗺𝗲𝗻𝘁 ] Nice newsletter, Andreas. Next! [ 𝗟𝗶𝗸𝗲 ] Want to stay close to private blue-chip tech? [ 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝘁𝗼 𝗺𝘆 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 𝗯𝗲𝗹𝗼𝘄 ]

A kid asked Jensen Huang whether AI would take everyone's jobs. Huang did not reassure him. Here’s what he told him instead. Huang said to get into the technology or get left behind, and to hold on to 2 traits: curiosity and generosity. I run a 40+ person asset manager, and that pairing describes how the firm is built better than anything I could have written myself. Curiosity is the hiring filter. Nobody has done what Stableton does before, so I cannot hire for experience in it. I hire people who want to find out how things work and who pick up a new tool the day it appears. In 2026, that is a hard skill. The people on my team who use AI all day are doing work that used to need a team. Generosity is the harder one, and running a company taught me a version of it I did not expect. It means giving people full ownership of their area and staying out of it, including when they do it differently than I would. Many of the people I work with are better at their jobs than I could ever be, and a flat organization only works if I act as though that is true. The 2 traits depend on each other. Curious people with no ownership leave. Ownership without curiosity turns into a silo. Together, they are how a small team runs $1bn without the hierarchy most firms our size would have built by now. Huang was talking to a kid. The advice is for anyone running a team this decade.

Broadcom supplies Anthropic's computing hardware. Now it has agreed to lend Anthropic up to $42B to help pay for it. The money would come as convertible notes that could turn into Anthropic shares, helping finance a $125.2B, five-year commitment to lease TPU computing capacity. Anthropic is expected to become Broadcom's largest compute customer in 2027. This week's other notable news: • Anduril won an Army contract worth up to $1.8B over five years to scale its command and control system. • Stripe agreed to buy Parafin, its second acquisition in two months. The full stories are in this week's newsletter. 👇 Find this helpful? [ 𝗥𝗲𝗽𝗼𝘀𝘁 ] Anything to add about this subject? [ 𝗖𝗼𝗺𝗺𝗲𝗻𝘁 ] Nice newsletter, Andreas. Next! [ 𝗟𝗶𝗸𝗲 ] Want to stay close to private blue-chip tech? [ 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝘁𝗼 𝗺𝘆 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 𝗯𝗲𝗹𝗼𝘄 ]

In 2005, YouTube had about 40 videos and a discouraged founder. 18 months later, Google paid $1.65 billion for it. The discouraged founder was co-founder Steve Chen. There's a clip of him from that year, admitting on camera that he wasn't sure anyone would ever bother watching what was on the site. By July 2006, YouTube was serving 100 million video views a day with 65,000 new uploads. Three months later, Google called. I watched a smaller version of this at Stableton. Our first $500 million in assets took 7 years. The next $500 million took 10 months. The early years went into the process of refining the framework, overcoming all the operational hurdles, and getting the initial scaling right. You can't skip that stage, even when progress feels incremental while you're inside it. I'm proud of the team for staying disciplined since day one. Every steep curve starts with a flat stretch where the only thing to trust is the work. If you're in the flat stretch right now, keep building. The compounding phase happens because the groundwork was done right. 🔔 If you enjoyed this one, follow me for more like it.

It's World Teachers' Day. Sam Altman had Paul Graham. Melanie Perkins had Bill Tai. I had… my dad. He started as an apprentice at a big German retailer and ended up building his own shirt business out of the Port of Hamburg. Then one day a storm flooded his warehouse and destroyed every shirt he owned. The business was gone overnight and he had to start from scratch again and went on to build 3 more. He taught me without ever meaning to. I just watched him work. In hindsight, it was the best business education I ever got. I see the same pattern behind the biggest names in private tech, including the ones in our Stableton Morningstar PitchBook Unicorn 20 strategy - an index portfolio of the 20 most valuable private tech companies in the world. Instead of betting on single names, institutional investors, banks, wealth managers, and family offices invest in the whole portfolio, and more than 150 financial institutions work with us as partners. Take 3 of the companies in it: - Ali Ghodsi arrived at Berkeley's AMPLab in 2009 and found his closest collaborator in Ion Stoica, one of the professors running the lab. They co-founded Databricks 4 years later. - Melanie Perkins met investor Bill Tai in 2010, and he did more than write a check. He pulled her into his network of investors and engineers, and Canva's advisor Lars Rasmussen later introduced her to Cameron Adams, now her third co-founder. - Sam Altman went through Y Combinator's first batch in 2005. Paul Graham spotted him early, mentored him for close to a decade, and in 2014 handed him the keys to YC itself, serving as President until 2019. Today he runs OpenAI. Lessons like that stick because you learn them next to someone who's already been through it, not from a textbook. So here's to the teachers, and to every mentor who takes the time to pass on what they know. I owe a lot of where I am today to the people who showed me how things work early on.