How Altimeter Is Investing $100M+ Checks | Anduril, SpaceX, K2 Space

Summary

• Erik Kriessmann’s single biggest position at Altimeter is Anduril, which he pushed to invest in almost immediately after joining in fall 2023. His case: a “modern defense prime” spanning air, land, sea, and space, stitched together by the Lattice software platform, where “the government is just getting a lot more value per dollar of spend” — and where strategic acquisitions produced Dive and Fury, products he says “never would’ve found their way into the hands of the government or the warfighter without Anduril.”
• The K2 Space math is the episode’s sharpest thesis: SpaceX bent the launch-cost curve so far that satellites should get big again. Viasat-3, one of the most powerful satellites ever launched, ran ~26 kilowatts and cost $500M+, launched one at a time; K2’s buses are 20 kW each, stack 10 in a Falcon 9, at roughly a 10X price reduction — “bigger, more power equals more dollar signs.” With Starlink/Starshield already owning LEO, “MEO and GEO are up for grabs,” backed by a Space Force mandate to proliferate across orbits.
• Altimeter’s model is extreme concentration: fewer than 70 total private investments across seven venture funds, roughly one or two per partner per year. “Diversification is the enemy of outlier returns” — Series A checks typically run $15–20M, sometimes $25M; growth checks run $50–250M+, and the firm is “happy to put a significant percentage of a fund into one company that’s clearly hitting escape velocity.”
• The crossover structure is a sourcing engine, not just a label: publics and privates explicitly feed each other. Private investment work in 2022 informed the public NVIDIA position, which informed the private CoreWeave bet; the pre-revenue Snowflake relationship informed Simuha, which Snowflake later acquired, as well as Tabular, Sigma, and ClickHouse.
• Kriessmann’s core investing heuristic comes from his human-capital career: “talent is the leading indicator of success of any organization.” In diligence, “if there’s 10 reasons we invest in a company, maybe the first eight are team” — and the signal he thinks many investors miss is tracking who a company is hiring, which helped him build conviction in Base’s “density of talent.”
• On exits: “max tech uncertainty, max political uncertainty, max economic uncertainty” — not appealing for IPOs, but great companies now have more optionality than ever. SpaceX-style biannual tenders, Stripe, and Databricks show value accruing to private-market holders; the risk is bifurcation into “haves and have-nots,” where overcapitalized companies with no secondary interest get stuck.
• Sector calls: missiles and munitions are a major open opportunity given drained stockpiles from Ukraine and Israel; autonomy across air, land, and sea is “inevitable.” On drones he’s selectively bullish — “I don’t think there are gonna be 100 massive drone companies. I think there will be a couple” — and Altimeter hasn’t made a standalone drone bet yet. Rapid-fire: ChatGPT is his clear favorite, with o3 “incredible”; Grok is useful inside X, while he isn’t using Gemini; SpaceX above 160 launches this year; and he doesn’t think the U.S. will disclose that aliens exist this year.

Deep dive

1. From “Who is Erik?” to a seat at the table

• Kriessmann’s decade in venture started in human capital at Khosla, where Vinod gave him a blunt success metric: “I’m just gonna ask the entrepreneurs… ‘What do you think of Erik?’ And if they tell me, ‘Hey, he’s great,’ I know you’re doing your job.” At Index he rebuilt the function from a blank slate, treating founders, future founders, and ecosystem talent as distinct “relationship journeys, almost like you’d think of a customer journey.”
• His operating method with portfolio CEOs — Alex Wang, Dylan Field, Asaf at Wiz — was to ask: “what’s the number one thing any investor could do for you this year that would be transformational?” Almost always people-related; then “you just deliver. You just do the thing.” The Scale AI CTO search landed Brad Porter (later founder of Collaborative Robotics), and knowing Zach Abrams and Sean before they had decided to start a company led Index to source and co-lead the seed for Bridge, later acquired by Stripe.
• The Altimeter route is the punchline: through co-investments like Confluent and ClickHouse, founders kept crediting him in board meetings he wasn’t in — “Who is this guy that’s not here that’s getting called out in the board meetings?” The firm’s pitch: “you’ve done the hard part of showing that you can build trust with and be valuable to exceptional entrepreneurs… perhaps you could be great at that here.”

2. Talent density is the deep-tech thesis

• The through-line of the episode: “talent is the leading indicator of success of any organization… the team you build is the company you build” — a maxim he says was ingrained at Index, and a central thesis behind his focus on aerospace, defense, energy, autonomy, mining, and robotics, beyond the fact that “one, it’s just cool.”
• The mechanism: these are “truly just huge slices of the economy” that are “too daunting for most. Most won’t even try. Even fewer are capable of pulling it off.” Because there aren’t 50 companies attacking each problem, the Palantir/Tesla/SpaceX-trained founder class can consolidate a dense pool of exceptional people into one business — “companies where you can stack the deck in your favor.”
• His relationship philosophy, offered with a disarming origin story: “maybe ‘cause I’m an only child… you gotta take an interest in other people.” The rule: “it’s not what everybody can do for you. It’s about what you can do for everybody else to earn the right to have a relationship.”

3. Anduril: the biggest bet, and why the vibe shift is real

• Investing in Anduril was one of the first things he pushed for after joining — “God, this is just obvious to me… destined to be a wild success.” A mutual friend, Christian, reconnected them to the Anduril team, and the fall 2023 pitch was pure Altimeter: “we are highly concentrated investors… we wanna concentrate and build a really large position and help you go from the private to public markets.”
• The business overview: a modern defense prime built on software, AI, and design-for-mass-manufacturing — sentry towers to Dive, Copperhead, counter-UAS, Roadrunner, Barracuda, Fury (the Group 5 autonomous drone for the CCA program), plus the Klaus acquisition in ruggedized edge compute — all stitched together with Lattice. Acquisitions are framed as technology that “probably never would’ve found its way into the hands of the government or the warfighter without Anduril.”
• On the cultural reset, he flags a paradox worth keeping: “there’s almost unwavering support for the men and women who serve… yet so much hate and negativity for the companies that provide them with the capabilities and tools.” Palmer and Brian’s answer was mission plus authenticity — anime, merch (“good luck getting their merch”), and a recruiting pitch that ends: “oh by the way, the wealth generation upside for you personally is massive.”
• Molly’s pushback from Brian’s Hill & Valley clip — bottlenecked by regulation, not manufacturing — draws a hedge, not a confirmation: certification and range testing create “hoops to jump through, in some cases for good reason… I’m sure they could be moving even faster in certain areas.”

4. K2 Space: when launch gets cheap, bigger is better

• Founders Karan and Neil (brothers), with a core team from SpaceX’s Dragon, Falcon 9, Starship, Starlink, and Starshield programs, had a specific insider insight: SpaceX will own launch and proliferated LEO, but MEO and GEO require purpose-designed satellites “SpaceX had no desire or interest” to build.
• The comparison that carries the argument: Viasat-3 at ~26 kW cost $500M+ and launches one at a time; K2’s buses are 20 kW each, ten to a Falcon 9 — 200 kW of capacity per launch at “10X more capability at 10X cheaper cost.” Vertical integration is the how: an integrator buys a Honeywell reaction wheel at $2M apiece; K2 builds its own at “a 10 or 100X reduction in price… across the entire bus.”
• The historical logic chain: launch cost per kilogram was so high that everyone shrank satellites despite massive capability trade-offs; SpaceX bent that curve, so “it was just inevitable that things are gonna get bigger again.” More power maps directly to bandwidth and throughput, positioning, navigation, and timing are in MEO, and the Space Force has a mandate to proliferate across all orbits — hence multi-mission, multi-orbit buses for government and commercial customers.

5. Crossover structure plus radical concentration

• Brad founded Altimeter as “an OG crossover” — earn the right to invest in exceptional companies “regardless of whether they’re public or private” — and the two funds explicitly compound: 2022 private investment work informed the public NVIDIA position, NVIDIA informed private CoreWeave; the pre-revenue Snowflake relationship informed Simuha, which Snowflake later acquired, as well as Tabular, which Databricks essentially acquired, Sigma, and ClickHouse.
• The discipline: fewer than 70 total private investments across seven funds (his own hedge: “don’t quote me on this”), no toe-holds, and Brad’s standing counsel — “wait for your pitch… one or two investments per partner per year,” not as a hard rule. Checks: typically $15–20M, sometimes $25M, leading a Series A; $50–250M or more at growth; and very large checks into quasi-public rounds like OpenAI.
• Diligence weighting: “if there’s 10 reasons we invest, maybe the first eight are team, then product and market.” Software gets ACVs, NDR, churn, gross margins; hard tech early on is technical progress and milestones plus customer pull — partnerships and early program wins — before bookings, pipeline coverage, and the path to free cash flow.

6. Exits: maturation of capital markets, and a haves/have-nots split

• His read on the current exit environment: “extreme volatility and uncertainty… max tech uncertainty, max political uncertainty, max economic uncertainty” — not appealing for IPOs, though “there’s always gonna be appetite from investors to invest in great businesses.”
• The structural shift: a lot of late-stage private capital gives founders “more optionality than ever.” Most companies still aim to become standalone, enduring public companies, but that is not the only path. SpaceX’s biannual tenders provide liquidity for early investors and employees, and similar approaches are appearing with Stripe and Databricks, while Anduril remains private. “Some of those companies may never go public. Some of them absolutely will.”
• To Molly’s question about getting “stuck with a class of secondaries”: he doesn’t think stuck is the word — staying private lets companies manage dilution and, if very profitable, even buy back stock, and “capital markets are gonna form themselves around these special assets.” The warning is for the have-nots: overcapitalized, high-valuation companies with no secondary interest, no IPO path, and no acquirer — “that’s another situation entirely.”

7. The sector map, Base, and rapid-fire calls

• Beyond Anduril and K2, the big defense opportunities he names: missiles and munitions — lack of supply of solid rocket motors and munitions, with stockpiles drained by Ukraine and Israel, makes rebuilding them “super important for deterrence” — plus inevitable autonomy across air, land, and sea. On drones: “I don’t think there are gonna be 100 massive drone companies. I think there will be a couple,” no standalone Altimeter bet yet, but he likes use cases such as Flock Safety’s acquisition of Aerodome for public-safety drones.
• Base (Justin and Zach) is his case study in the hiring signal: “a big, bold, hairy market opportunity, but I really think they’re the only ones who have a shot” — and one thing he thinks many investors miss is tracking who Base and similar companies are hiring, “that’s really how I build conviction.”
• The Calshi rapid-fire: best AI is “ChatGPT, not even close” — “o3’s incredible”; Grok is useful inside X, while he isn’t using Gemini because he isn’t coding or using coding agents; Perplexity is no longer his main tool beyond basic search. Wealthiest three: “Elon, Elon, and Elon” (then Jeff and Mark). SpaceX launches this year: above 160, “for sure above 150,” maybe 170. On whether the U.S. will say aliens exist this year, he doesn’t think it will; at 6% odds, “the odds are in my favor.”