Makda Fitsum

Brex got acquired by Capital One for $5.15B. One of the largest fintech M&A deals ever. Great that M&A happened. But the number was shocking $5.15Bn. Just a few months ago, Brex was valued at $30Bn. This is a down exit. Preference stacks matter, and common shareholders + employees may see little to nothing. Looks like something happened at Brex that pushed them to get acquired while wiping out 82% of the startup's value. Hoping to get more info on this in the public domain in the days to come.

12

I’ve spoken to over 2 dozen MDs at PE firms I can confidently say that the arb of figuring out how to implement Vertical AI at portfolio companies is very real right now It will fundamentally change underwriting for those who can do it predictably and unlock generational returns. Most are aware they need to act. Very few have.

454

64 Comments

Two weeks ago, Brex and Capital One announced an agreement for Capital One to acquire Brex for $5.15 billion. This is the largest bank-fintech deal in recent history. I've spent the last two weeks thinking about what this means for the PE firms and portfolio companies I work with. The short answer is that everything just changed in terms of what's possible. Brex and Ramp combined had about 3% of the corporate card market. The real fight was never between us and other fintechs. It was always against American Express, JP Morgan and the big banks who control 90%+ of the market. We've been building great products, but we've been doing it with a fraction of their resources and balance sheet capacity. That constraint just disappeared. On day one, we become the #3 corporate card issuer in the US. We gain access to $700 billion in assets, a $6 billion R&D budget, and the ability to offer credit limits that are 10-20x higher than what we could do before. Our AI roadmap just accelerated by 2-3 years because we now have the infrastructure and capital to move faster than anyone else in the market. What makes this work is that Capital One understands how to acquire without destroying what made the company valuable in the first place – or as Pedro has repeated over the last two weeks, “don’t crush the butterfly.” They did this after acquiring ING Direct and they're doing it with us. Pedro stays as CEO, the team stays intact, the culture stays unchanged. Rich Fairbank built Capital One in the 90s by using data and technology to disrupt credit cards. He sees the same DNA in Brex. This isn't about cost cuts. It's about giving us the resources to win the market outright. For the PE firms I work with, this changes the equation significantly. You now get institutional-grade underwriting capacity with the product innovation and service model you've come to expect from us, all backed by a top-10 bank. Same team, same partnership approach, just with Fortune 50 resources behind every relationship. The future of corporate finance isn't American Express or JP Morgan with a better app. It's AI-powered corporate cards from the people who invented it, spend management, real-time compliance, and embedded banking built on infrastructure that can actually support it at scale. That's what we're building now, and we just got a 50x multiplier on our ability to execute. If you're a PE firm interested in how this impacts your portfolio companies, let's talk.

105

2 Comments

Great overview of the Florida VC ecosystem 👇 What stands out is not just the ~$2.85B raised, but how distributed and specialized each region is becoming. • South Florida → crypto, fintech, LatAm access • Tampa Bay → strong early-stage pipeline • Orlando/Space Coast → deep tech, defense, space • North Florida → emerging operator-led funds Tax advantages help, but the real story is growing density of founders, angels, and operators building locally. Still early compared to SF and NY, which is the opportunity. Who are the next Florida-based funds or angels getting active?

4

3 Comments

Startups: New H-1B moves are a stealth “Series A tax” on talent. The headline: the administration just unveiled a $100,000 one-time fee on new H-1B applications (the White House says it applies only to new cases, not renewals), and it’s also pushing to replace the lottery with a wage-weighted selection that favors higher-paid roles. Congress, meanwhile, has an H-1B reform bill in the mix. What this means for early-stage hiring: Runway math changes. A $100k fee per new H-1B could equal multiple months of someone’s salary + options—per hire. Expect boards to ask “can we justify this at Seed/Series A?” Salary bands move up. A wage-based allocation system rewards higher offers; startups will have to either pay closer to BigCo bands or risk losing out. Time-to-fill stretches. Fewer viable visa paths = more rejected offers, more searches, and delayed product roadmaps. Geo strategy gets real. More teams will add near-shore hubs and US-based contract-to-hire while they navigate immigration risk. Visa portfolio thinking becomes standard: for top candidates consider O-1A (extraordinary ability), TN (Canada/Mexico), E-3 (Australia), or L-1 transfers via partner entities. (Talk to counsel—facts matter.) University recruiting/remote fellowships may become the bridge to eventual sponsorship once rules settle. Founder POV: If “talent is the moat,” then policy just raised the toll. Winning teams will (a) budget for the true cost of global hiring, (b) sharpen compensation/leveling, and (c) build Plan B/C pipelines so you’re never one policy away from a missed launch. Curious how this hits your hiring plan? Drop your stage + headcount goals; I’ll sketch an immigration-aware recruiting playbook you can bring to your next board meeting.

8

1 Comment

Boston doesn't just breed startups. It breeds operators who've seen the inside of billion-dollar exits and still come back hungry. That's the DNA behind Jellyfish. Founded in 2017 by Andrew Lau, David Gourley, and Philip Braden, the company set up at 225 Franklin Street with one mission: bridge the gap between engineers building and executives betting. The founders lived that disconnect at Endeca, the Cambridge outfit Oracle scooped up for $1B+. They knew the gap wasn’t abstract. It was personal. Jellyfish was their answer. Eight years later, Jellyfish is back with a launch that pushes its platform closer to an operating system. On Sept 9, 2025, the company released its Model Context Protocol server as a Claude Desktop extension. That may sound technical, but the shift is anything but academic. What used to require API integration is now a desktop download. Ask Claude “Where are we investing engineering effort this quarter?” and the answer arrives conversationally, in real time, safeguarded by Meta's #Llama #PromptGuard2. This isn’t a gimmick. It’s a structural change in how engineering intelligence gets delivered. Jellyfish already powers 500+ organizations, Mastercard, PagerDuty, Priceline, ZoomInfo, Hootsuite among them, but this extension lowers the barrier for every stakeholder. Finance can query allocation. Product can explore delivery. Leadership can ask about throughput. No dashboards, no gatekeeping. Just data in plain language. The MCP push carries Nicholas Arcolano, Ph.D.’s fingerprints. Jellyfish’s Head of Research, a Harvard University-trained mathematician, has published lessons from MCP experimentation since Anthropic unveiled the protocol in 2024. Others are validating too: Jamie Newcomb at Ably has shared how MCP integration evolved from basic retrieval to analysis that shaped decisions. The official blog post was signed by Sophie, a Jellyfish research intern. Light on detail, heavy on intent: take the intelligence already humming inside Jellyfish and make it available in workflows people adopt daily. Momentum backs the move. Jellyfish has raised $114.5M through Series C, tripled revenue and headcount in 2021, launched an AI Impact Dashboard in 2024, and landed on Forbes’ Next Billion-Dollar Startups in 2023. Competitors like LinearB, Swarmia, and Waydev can’t match the same focus. Jellyfish has always been less about vanity metrics, more about business clarity. Now that clarity speaks back. The extension is more than a technical win. Once your engineering intelligence plugs seamlessly into AI workflows, it stops being optional, it becomes indispensable. The only real question is how many teams will be waiting when Jellyfish expands to more platforms. #AI #EngineeringIntelligence #EngineeringLeadership #SoftwareDevelopment #SoftwareEngineering #Integration #Data #Infrastructure #Innovation #TechEcosystem #StartupEcosystem #TechNews If software engineering peace of mind is what you crave, Vention is your zen.

1

1 Comment

From early hire to CBO, Art L. shares how Brex scaled from a corporate card for startups to a global spend platform built for enterprises, with lessons on building trust, navigating M&A, and leading through change. Catch it all on the latest Sourcery drop 🎙️⬇️

14

1 Comment

Opendoor was a $15B SPAC rocketship. Then it crashed 98%. Now it’s 4× in three weeks… and might have just meme-stocked its way back to life. Let me explain, because this story is completely unhinged: Opendoor went public at the peak of the SPAC boom, riding on Chamath tweets, algorithmic home-flipping hype, and the idea that real estate agents were obsolete. At one point it was trading at $35 a share. Then the market did what markets do. Rates went up. Housing slowed down. And Opendoor kept buying homes like it was still 2021. They bought 35,000 homes and lost a billion dollars doing it. CEO got the boot. Stock fell to $0.60. They were about to be delisted from Nasdaq. They even scheduled a reverse stock split just to survive. It looked like another SPAC carcass on the side of the road. Then WallStreetBets showed up. Reddit declared: “Opendoor is the next Carvana.” Trading volume exploded. $OPEN went from $0.60 to $2.50+ in weeks — at one point spiking 90% in five days. Even touched $4 intraday. And it wasn’t just memes. A hedge fund manager (Eric Jackson) dropped an $82 price target, called it a 100-bagger, and outlined a full turnaround thesis: • Cost-cutting and layoffs • Better unit economics • New model that sends leads to real estate agents • Debt refinanced • $700M+ in cash on hand • EBITDA breakeven in sight Suddenly, Opendoor isn’t a bankruptcy story. It’s a meme stock with a potential redemption arc. And here’s the wild part: The Reddit rally may have actually saved the company. With the share price back above $1: • They’re no longer at risk of delisting • They might shelve the reverse split • They can raise new capital or do deals using equity • They bought themselves time to prove the turnaround is real It’s GameStop and Carvana all over again. But there’s one big question left: Is this a real comeback, or just a sugar high? The fundamentals haven’t changed much. Still unprofitable. Still at the mercy of the housing market. Still has to convince customers to sell their homes to a website. But for now, they’ve got attention. They’ve got cash. And they’ve got a second chance. What they do next will either make this the comeback story of the decade… Or just another chapter in the SPAC collapse Wikipedia page. 👇 Would you bet on OPEN turning this around? Or is this just a Reddit fever dream?

25

5 Comments

Most private equity firms hire operators post-acquisition and give them "equity." But it's usually not real equity. It's phantom equity, options with impossible vesting schedules, or profits interests that only pay out if the fund hits a certain IRR. The economic reality is that the operator is an employee with upside optionality, not an owner. That misalignment creates problems. When the operator doesn't have meaningful skin in the game, their incentives diverge from investors. They optimize for salary and bonus, not long-term value creation. And when the business hits challenges, they're more likely to leave than fight through it. Our model is different. The operator receives a percentage carry at close. Day one, before the business generates a dollar of profit. That's real ownership. If the business delivers a set net IRR, the operator earns an additional carry. At the top end, they own a healthy percentage of the total upside. This isn't a "nice job" bonus. It's generational wealth creation tied directly to investor returns. That alignment changes everything about how operators run the business. They're not thinking about their next job. They're thinking about how to drive EBITDA growth, improve margins, and position the company for a premium exit. When investors win, operators win in actual dollars that hit their bank account. Most PE firms talk about alignment. We've structured it into the cap table and waterfall mechanics in a way that makes it impossible to misalign. ✍️ If you are an accredited investor and actively investing in the lower-middle market... Every week, I breakdown what's actually happening in operator-led buyouts and lower-middle-market M&A. Subscribe to "Operator-Led" (link on my profile).

76

21 Comments

With most most the conversation positioning AI as a rocket ship, here is a new visual for you. And an accurate one at that. The AI Wrecking Ball The discussion is how Generative AI is rapidly eroding the moats of traditional vertical software and other businesses. Most of it is real. Some of it is an overreaction. David Sambur Pradeep Aradhya

4

Early fund success isn't just momentum. It can be a mathematical lock. Consider the following math problem, known as Polya's Urn: - Start with an urn containing one red ball and one blue ball. - Draw a ball at random, return it to the urn. - Whichever color you drew, add one more ball of the same color to the urn. - Repeat 100 times. Now think about this question: if your first two draws are red balls, what’s the probability by 100 draws that the fraction of red balls in the urn is below 50%? Let’s walk through the math... Before your first draw, the probability of red or blue is 50%. The expected value of the mix over time is also 50/50. Now draw your first ball. Draw red, put it back and add another red ball. The Urn is now 2 red, 1 blue. Next-draw probability for red: 67%. Expected long-run red rate: also 67%. Draw again and also draw red. The urn is now 3 red, 1 blue. Next-draw success rate: 75%. Long-run expectation: 75%. The fraction of red balls in Pólya's urn is what mathematicians call a martingale — a process whose expected future value always equals its current value. Wherever the early draws push you becomes the system's new expectation for where it will end up. Early draws have massive influence. The long-run destination is being written in the first few chapters. Now let’s extend this to fund performance. Imagine that each draw represents the performance of an investment. Red means success; blue bad result. Follow the same draw mechanics and you can quickly see that a couple of early successes start to heavily tilt your long-term probabilities. Specifically, if the first two balls you draw are red (successes), the probability that after 100 draws you are less than 50% red (successful) goes to only 12.5%. In just the first two moves, you have greatly shifted the eventual end state probability. This is often what plays out in VC. Ramana Nanda at Harvard and Olav Sorenson at Yale studied investment-level data across hundreds of VC firms in 2020 and found that each additional IPO among a firm's first ten investments predicted an 8% higher IPO rate on all subsequent investments. The mechanism wasn't superior investment-picking. Early success improved reputation expanded co-investment networks, and unlocked deal flow that wasn't previously available. Each success literally added a red ball to the urn. The persistence data confirms the pattern. Harris, Jenkinson, and Kaplan at Virginia, Oxford, and Chicago found that top-quartile VC firms repeated top-quartile performance 48% of the time: nearly double the 25% you'd expect if outcomes were independent. Bottom-quartile firms stayed bottom quartile 40% of the time. Once the urn tilts, it tends to stay tilted. Most models treat portfolio outcomes as independent draws from a fixed distribution. Pólya's urn shows why that's wrong. In systems where success breeds access and failure breeds friction, early results don't wash out. They compound. Cheers!

12

4 Comments