Rex Salisbury

1/ Excited to announce the launch of @Cambrian_HQ 's inaugural $20 million fund to back the next generation of fintech founders! Incredibly grateful for my time at @a16z, and, now, exited to share how I ended up as a solo capitalist by way of community building. 🧵 👇

@thogge 👎 most of those don't even make sense (except "stay out of the way"). they offer brand signaling. that is 90% of the value and it's v. meaningful. the rest is deck chairs.

Every great VC firm has a unique product they offer: -a16z: turn founders into CEO w/ operational help -benchmark: ultimate consigliere/partner -sequoia: we know how to build an enduring co -thrive: ultimate loyalty -FF: stay out of your way -KV: venture assistance agree?

@Area224 shepards pie is great. Agree on Mac N Cheese. it is perfectable editable, but seems like it should be better!

Costco's shepherd's pie (actually cottage pie, but no complaints) is top-notch. Chicken pot pie also solid. Both can easily feed family of five. Meal kits, like the taco kits, are also a strong choice. Note: Sam's Club — "Poor Man's Costco" — actually has superior mac and cheese. Downside of all of them: prep instructions are usually off by a little to a lot. Also Sam's puts them on the underside of the food. But set aside an hour or so to make the big dishes.

Hot take — Costco kicks the crap out of cloud kitchens. Feeing a family of 5 for ~$25 inclusive of delivery. Costco is sitting on a gold mine. They have quite limited selection for the prepared meals, and honestly, they could be better. Both those things are fixable. I’d personally be very happy to buy 10x more from Costco if they did this and I’m sure lots of other folks would to. As in aside, order ordering dinner for five people from DoorDash sucks just from the amount of garbage it generates your trash is full after 2 meals.

Excited to finally announce what I've been working on the last few months: a new exchange to trade the AI takeoff! LFG

Stablecoins are not stable. we need something better. Robert Shiller has the answer, invented by the chilean gov't in the 1960s. First, Why? Stables reference the dollar. but inflation erodes value of the dollar. There is a solution. Instead of a "Unit of Account" for a currency we need a "Unit of Price". It exists. Robert Shiller calls it his favorite underrated financial innovation. 60 years later, crypto independently rediscovered the need for this product — and getting it half right w/ stables. Right b/c the US dollar is relatively stable. Wrong b/c we need "relatively" as a qualifier. The historical context of the first "Unit of Price". In 1960s, Chile was drowning in inflation. Nobody would sign a long-term contract. Mortgages were impossible. Landlords wouldn't commit to rent for more than a month. On Jan 20, 1967, the government created something radical: the Unidad de Fomento (UF) — a non-circulating unit of account. The UF didn't reference a currency, it referenced the price level. the UF was the world's first "Unit of Price". The UF exists today and is pervasive in chili. The UF solved the cold start problem by referencing it in their own development loans -> then requiring adoption by banks. The biggest switch came from the 1981 AFP pension privatization — which created mandatory institutional demand for UF-denominated assets at scale. Once a critical mass of contracts referenced the UF, more contract types followed suit. The UF exchange rate against the peso is constantly adjusted for inflation so that its purchasing power remains nearly constant on a daily basis. it is NOT a currency (it's not in circulation). You can't spend it. It's a purely abstract unit of account — goods quoted in UFs can only be purchased with pesos. The UF has now replaced currency for most long-term contracts in Chile, and for purchases and sales of large items. - Pension payments - alimony, and child support payments are - offices for sale are often quoted in UFs Shiller's a bit frustrated. The UF was adopted 58 years go. It has been copied in Colombia, Ecuador, Mexico, and Uruguay. He tried to design a "Unit of Price" for the US...but it never took off! Ironically, the US's relatively low inflation environment historically means we are more exposed than some high inflation countries. We've never had to invent something that was inflation proof, so we didn't. Instead we choose to trust the dollars value wouldn't get inflated away. Without the pain of extreme inflationary periods, there was no way to build a political coalition to solve a problem we hadn't truly experienced (yes, we've had bouts of inflation, but ppl still had mortgages and long term leases). Now fast-forward to 2025. The stablecoin market is $250B+ and growing. But 99% of that market is pegged to the dollar. USDC. USDT. DAI. They're stable-ish — but silently depreciate with the inflation. A handful of projects are explicitly reinventing the UF on-chain. - Flatcoin designs like Nuon Frax Price Index (FPI) already use Truflation US CPI indices as an underlying source of truth, effectively upgrading "official" inflation to a censorship-resistant oracle — once the decentralized oracles exist, you can finally peg money to purchasing power, not politics (presuming oracles don't get compromised in their own politics or the governments gathering the data don't screw with it or stop *ahem* china or in the US cuts to various federal statistical agencies and economic data collection by POTUS) - Reflexer's RAI tried a different path (since wound down): a non-pegged stablecoin whose exchange rate is determined by supply and demand while the protocol tries to stabilize its price by constantly devaluing or revaluing it Reflexer — essentially a crypto-native unit of account (hard in practice!). But here's what most flatcoin builders are missing, and what Shiller understood deeply: the killer app isn't the token itself. *It's the contracts denominated in it.* Chile's UF worked because it became the standard unit for writing obligations — mortgages, rents, pensions. The token is just a vessel. The durable innovation is the unit of account layer that sits above it, embedded in legal and commercial agreements. So the success of Flatcoins is not their supply (which is likely to be eclipsed by stables), but the extent to which they are referenced. Similar to a LIBOR or SOFR used to credit contracts. That's what took the UF ~20 years to achieve (again big tipping point in 1981 w/ the pension adoption requirement) — but no flatcoin project has cracked yet (and certainly no flatcoin has gotten a government mandate!). The meta-point: Most of crypto's monetary experiments are rediscovering problems that economists solved decades ago — inflation indexation, seigniorage, reserve requirements — but with programmable, permissionless rails and without central banks as the single point of failure. And some of crypto's biggest "innovations" reflect not fundamental "invention" so much as outflanking an sclerotic regulatory state that should be doing these things independently. h/t @RobertJShiller:

TIL; the father of venture capital also created the first currency swap. 1980's were greatest time to be a financial innovator. 1970's we had - no swaps - no options - no electronic trading! - no VCs ...list goes on.

@HaunVentures has exited 2 stablecoin competitors - Aug 2024 -- invested in Bridge (reported in press) - Oct 2024 -- Bridge sold to stripe - Dec 2024 -- invested in BVNK (reported in press) - Mar 2026 -- BVNK sold to mastercard great timing. who's next!?

EXCLUSIVE: interview w/ Co-Founder of @BVNKFinance, Chris Harmse, on the mastercard acquisition. We get into: (00:00) deal rationale (00:34) what this means for customers (00:54)what this means for Mastercard competitor Visa...who previously invested! (01:28) their plan to hit $1 trillion in TPV (up 30x from where they are today). stablecoin wars heating up. major platforms like mastercard don't want to miss the wave, but what is the wave and how does this acquisition position them to win? 👇

Autonomous agents are quietly becoming the internet’s most powerful economic actors. Stripe’s Steve Kaliski explains how machine payments and stablecoins will fundamentally rewrite who captures value in commerce. (0:00) Are Agents Already Economic Actors? (3:48) How Does Stripe Automate Its Engineers? (5:18) Live Demo: Stripe Minion Updating Docs (8:28) What Is the Agentic Commerce Protocol? (11:28) How Do Shared Payment Tokens Actually Work? (19:14) Will Agents Have Their Own Web Standards? (25:02) Why Build Machine Payments on Stablecoins? (28:40) Will AI Agents Need Chargebacks and Refunds? (31:02) How Do Ephemeral Agents Pay for APIs? (33:48) Live Demo: Paying APIs with Stablecoins (39:31) Will AI Agents Negotiate While We Sleep? (41:15) Did the Autonomous Minion Finish the Code? (41:57) Does Kaliski Even Write Code Anymore? (44:07) Does AI Make Us Faster or Just Busier?

@varshikaARK The Y axis is a “fixed cost” not marginal.

Why did Capital One then move faster than a startup to close a multi-billion dollar deal? (0:00) Largest Bank Deal in History? (1:31) How to Close a $5B Deal in 30 Days? (2:51) Is Silicon Valley Missing the Growth Story? (6:23) Why a $1B Investment in Brex's Strategy? (9:34) The Power of 100M Distribution Accounts? (15:31) Why Lending is the Ultimate Profit Pool? (20:25) Can Capital One Preserve Brex’s Speed? (22:25) Are AI Agents the End of Manual Accounting? (24:30) Why Franceschi Reset Valuation to $4B? (31:20) The Unparalleled Tech Ecosystem of SF?

Garry shipping open source 1030 on a Friday night. SF is fun. Come check it out!

VC IRR decay for 2017 and 2018 has been bad. And this was before the SaaSpocalypse. IRRs driven by big positions held at last round valuations. What would they look like at true fmv?

@WallStreetMav tryhoma.com is the answer.

I am actually surprised that the era of real estate agents and 6% commissions didn't end years ago. What is keeping it alive?

@HarryStebbings I’ll take that bet. $10k says 10 “dog 💩” $30-75 million funds in the 2024 vintage (carta data) will outperform 20 VC‘s 2024 vintage fund. Will settle based on net TVPI in 2030 carta data. @PeterJ_Walker can arbitrate. @HarryStebbings down?

LPs love the idea of the 30M-75M high ownership seed fund. These will be the worst performing funds of this vintage. Average top tier seed rounds are $5M. These funds don’t have enough to lead and have a diversified portfolio. But they are too large to be collaborative and work with the best leads. Adverse selection takes effect. Be small enough to collaborate (sub $20M) or be large enough to lead ($100M+).

all the stats from the convo. thanks Claude! Video: "Inside a16z Growth" — Alex Immerman (GP, a16z Growth) on Sourcery with Molly O'Shea --- Public vs. Private Market Growth - Fewer than 5 public internet software/fintech companies growing north of 30% this year - a16z Growth didn't invest in a single company last year growing only 30% — massive disconnect between public and private growth rates - OpenAI and Anthropic alone added as much revenue last year as half of the public cloud universe (excluding Mag 7) a16z Growth Portfolio & Strategy - ~50% of growth investments are follow-ons to early stage winners - 90% of companies they invest in from the growth fund, the early stage team has already met - Portfolio companies growing on average over 100% - Early growth companies need to be growing several hundred percent for them to get excited - a16z raised 18% of all US venture dollars last year - Growth team has always been small: 7 to 12 people Market Size / Venture Landscape - In 2009 (a16z founding): common wisdom was 15 companies/year would reach $100M revenue; today that's ~150/year (10x) - Top 1% outcome in 2009: ~$1.5B; today: $10B - Top 1% outcome doubles every 5 years — projects to $40B in 8-10 years Waymo - Raised $16B at $126B valuation (Series D) - a16z participated in all 4 external rounds since Series A - 25% market share in San Francisco with very few cars, surpassing Lyft - Only ~1,000 cars in the Bay Area - Scaling to over 20 new cities this year - Ride sharing today: ~$125B market, growing low double digits; 40% in US - Ride hailing is less than 1% of vehicle miles in the US - Ride share market expected to grow by at least an order of magnitude - Tesla FSD 14.1 showed an order of magnitude improvement in disengagements Flock Safety - Solves over 2,800 cases every day - That's 15% of reported crime in America - 50% of murders are unsolved in America ElevenLabs - Just raised $500M - a16z led or co-led the A, B, and C rounds - Enterprise customers include Meta, Salesforce, Deutsche Telecom, Revolut Kalshi - Was #2 in reported volume when a16z invested; now #1 across users, revenue, volume - First CFTC-regulated prediction market - Partners: Robinhood, Coinbase, CNN, CNBC Revolut - 10 different revenue streams, no single one more than 20% of revenue - Already has ~40% ROEs (higher than Nubank) - Over 100% net revenue retention - Primarily transactional/payments vs. lending-focused Robinhood - 11 lines of business doing $100M+ in revenue each Elise AI - Started with leasing assistant wedge product - 2 additional products at 60% attach rate, 2 more at 40% attach rate AI Company Margins - Many AI companies showing up with 0-50% gross margins vs. traditional 70%+ Trillionaire Prediction - 90%+ odds on Elon Musk being first - Dark horse / second trillionaire pick: Larry Page (Alphabet owns ~100% of search, 100% of DeepMind, mid-teens % of Anthropic, mid-single digits % of SpaceX)

BREAKING: Inside a16z's $22B AUM Growth Fund Waymo, Kalshi, Coinbase, Robinhood, Stripe, Revolut, ElevenLabs, Flock Safety, Harvey, EliseAI, Hebbia, Roblox, Anduril, Sardine With a fresh $6.75B Growth fund, General Partner Alex Immerman (@aleximm) shares how @a16z evaluates category leaders & why firms are rethinking old metrics in the AI era. “If you look at OpenAI & Anthropic alone, those two companies added as much revenue last year as 1/2 of the public cloud universe excluding the Mag Seven.” Even as AI added ~$6T in market value to the stock market in 2025, we discuss the private companies driving the next wave of growth. Topics include: • Marc Andreessen lore • Why the fastest-growing tech companies are still private • Waymo’s expansion & the future of autonomous vehicles • ElevenLabs & the breakout of voice AI platforms • Kalshi & the rise of prediction markets • The four largest tech companies adding 1% to U.S. GDP growth with $400B in spending in 2025 • How a16z grew from a startup VC to capturing 18% of all U.S. venture capital • Why engagement, retention, & product defensibility matter more than ever Full conversation ↓ 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 (00:00) Alex Immerman, GP a16z (00:53) Flock Safety: Most underrated company in America (03:43) Private vs public market growth gap (04:43) Do gross margins matter in AI companies? (06:33) Waymo’s growth & $16B funding round (08:13) The future of ride sharing & autonomous vehicles (10:01) Where the value in self driving tech will accrue (12:27) Tesla FSD vs Waymo’s full stack approach (13:50) The growing brand power of Waymo (15:43) What comes after autonomous cars? (17:41) Are people really worried about AI taking over? (18:55) Inside ElevenLabs’ $500M funding round (22:12) The real world use cases of voice AI agents (22:55) How a16z invested in Kalshi (24:11) Why Kalshi over Polymarket? (26:35) Prediction markets vs g*mbling (29:33) Is fintech making a comeback? (32:44) Why stablecoins are becoming important (33:30) Lessons from Coinbase founder Brian Armstrong (34:19) How a16z evaluates growth companies (36:52) Where startups spend money: R&D vs marketing (38:44) Why startups need a strong Act 2 (40:08) Why venture outcomes are getting bigger & bigger (42:24) Why backing the market leader matters (43:49) The best books for founders (44:59) How a16z built its media machine (47:03) Who will be the first Trillionaire?

@DenverRayburn you're forgetting we spent $300+ billion subsidizing the purchase of heavy machinery in the OBBB. we don't need more subsidies through our various regulatory apparatus whether that's the tax code or banking regs.

1/ American manufacturers pay 3x more to finance equipment than their global competitors. It has nothing to do with pricing or tariffs, it's a regulatory accident from 2008 that no one has bothered to fix until now:

love that @tryramp's landing page for Agent Cards actually has a page for agents! all products need this.

@bscholl Everyone should watch this video. Turbine blades are things of beauty. youtube.com/watch?v=QtxVdC…

🤔