Fika Ventures does eight to 10 investments a year and leads or co-leads all of them, which means John Chen has to be right about a company before anyone else has voted on it.
For most of venture's history the capital markets waited for the customer markets. Companies won accounts, a leader emerged, and then the money piled in behind the winner. Chen says that sequence has reversed, and that a founder who loses the funding contest now has to answer a rival holding a billion dollars before a single customer has chosen.
"But now that's flipped, and the VCs are choosing who the king is, and capital has become a weapon"
Chen was at Emergence when the firm invested in Zoom — in his first two weeks there, when Zoom was the ninth entrant in video conferencing and the obvious money was elsewhere — and before that he wrote for TechCrunch and was chief of staff at Box.
The full interview is covered here so you can skip it. 48 minutes of audio, 21 minutes of reading.
Here are the 13 lessons that matter.
👤 Guest: John Chen, General Partner at Fika Ventures, a pre-Series A firm backing B2B companies across vertical AI, fintech and developer tools; previously an investor at Emergence, a writer at TechCrunch and chief of staff at Box
🎙️ Host: Nick Moran, founder of New Stack Ventures and host of The Full Ratchet
📰 Published: 14 September 2026 on the Full Ratchet podcast feed
🟣 Apple Podcasts | 🔗 Show notes | ⏱️ 48 min | ✅ Time saved: 27 min
Key Takeaways
Venture firms now decide the market leader before customers have chosen one
The question a founder has to answer is how to compete against a rival holding a billion dollars
A customer running a bake-off across 10 or 15 vendors is a buy signal, not a warning
It means the workflow matters enough that price and consensus are not deciding it
Chen's worry is that seed, not the Series B, is now the round that is not paid for its risk
Pre-seed companies raising $10M to $15M is the shape that concerns him, and Fika has not done one
The gross margin bar for AI companies has fallen from 80% to the teens
Growth is what the market is paying for, and margin is treated as fixable later
The two early metrics he trusts are pilot conversion and logo churn, not quick ratios
The legal AI company he backed converts 97% of pilots to paid and has never lost a customer
Copilots are not dead where the work is physical, because sending a robot to fix a machine makes no sense
Companies are buying several overlapping AI products at once out of fear and greed together
Institutionalizing "founder" as a career path breaks the pact that a venture check is not a research grant
1. Why Fika Is in LA
Nick Moran opened by asking why anyone would build a B2B venture firm in Los Angeles, and Chen gave three reasons and a 2017 thesis.
The founding bet was on a talent pipeline: "So, we were founded in 2017, and the thesis was pretty simple, which is there was going to be a whole class of founders emerging from SpaceX and ServiceTitan and Snap, that we're going to want to work on B2B ideas," Chen said.
That cut against the local consensus. "I think the proverbial wisdom at the time was L.A. is more consumer focused, more media focused, but we really really wanted to plant the flag for B2B."
He says the bet paid: Fika has backed its third ex-SpaceX team and its fourth ex-ServiceTitan team, and any LA founder working on a B2B idea is likely to come to the firm. It is no longer only local — "So a third of our companies are in L.A. and two thirds are everywhere else, so we're always on planes."
The second reason is customers. Los Angeles is home to the trades, shipping and logistics, and aerospace and defense — the real-world industries Fika's companies sell into.
The third is deliberate distance: the firm sits "close enough to the echo chamber" of San Francisco to stay "SF calibrated," he said, but "far enough away" to think independently. The aim, in his words, is "to be not high on our own supply."
He is in the Bay Area a day or two every week anyway. "I go up there once a week, actually," he said, because the best founders set the bar for what excellence looks like and the only way to tell sense from nonsense is to see both.
2. Insiders and Outsiders
Asked what actually separates Bay Area founders from LA founders, Chen said the split is not geography at all.
"I think that the number one trait is whether you're an insider or an outsider, and a lot of Fika success actually has been finding a lot of folks that are on the periphery, on the outside, that don't have YC or Stanford or things that are really legible to SF investors, and helping them sort of move towards the inside, the insider track," he said.
His example is Ivo, whose founder Fika met three months after he moved from New Zealand. "He was a corporate lawyer in New Zealand. Did not have YC, did not have Stanford on his resume." He "was not legible to the SF firms," Chen said, but was an incredible founder.
What convinced Chen was a live test rather than a deck. "We put him in front of three general counsels, who are his customers," he said — prospective customers at the time — and the founder closed all three inside 20 minutes on a product demo.
The seed round was in late 2023. Ivo has since raised a Series B, and Chen listed its customers as Meta, Uber, Google, IBM, Notion, Quora and Reddit.
His summary of the pattern: Fika saw a founder who was exceptional and "misunderstood by the market." He also describes what a founder like that sounds like in a meeting: "If you talk to him, he's thought about every question that you've asked five times over, like an incredible product mind."
3. How Fika Found Ajax
Asked how the firm sources founders like that beyond being present in LA, Chen told a story about an office next door.
The referral came from another portfolio founder who shared a WeWork with the company: "We have no idea what these folks do, but they're here every Sunday before we get here, and they leave after we leave. I don't know what they do, but you should talk to them."
The company was Ajax, which does AI legal timekeeping. Fika called on a Friday, worked through the weekend, and flew out Tuesday.
"Tuesday night at 8p.m. is when we delivered the term sheet. From 8p.m. to midnight, they call eight of our founders to do references on us, and then from midnight to two a.m. they we signed a deal, and that's that," Chen said — crediting the founders' speed rather than Fika's.
Why other founders refer deals to Fika at all comes back to how it operates: "We only do eight to 10 investments a year. We're not doing call options. We're leading and co-leading. We are deeply in context and integrated with our founders."
Chen said the company is based in New York rather than the Bay Area — another team that was not, as he put it, plugged into the matrix of the valley. Moran noted how much of Chen's recent work is in legal tech; Chen promised the firm does other things too.
4. Lessons From Emergence
Moran asked what Chen took from Emergence, where he invested from 2014 to 2018, and he named three things.
The first is a posture toward consensus. Chen described the firm as "aggressively non heat seeking," where the governing question about any company was what the firm could see in it "that others don't".
The case that taught it was Zoom: "I still remember in my first two weeks at Emergence we made the Zoom investment and I remember that vividly because Zoom at the time like this is hard to imagine but at the time was a late mover. It was like the ninth company in video conferencing." Blue Jeans and UberConference were the hot names; the team was in San Jose while the gravity was in San Francisco.
The metrics were the tell, and they were understated. Chen said the founder had come from WebEx — "he was responsible, for some of the damage" — and then, remarkably, that the metrics the founder reported were "underreporting how fast he was growing."
The underrated founder test he took from it is whether a founder has shown the ability to "look stupid for a really long time" before being vindicated. Alongside deep referencing on when a founder has stood against the grain, or done something genuinely difficult more than once.
The second lesson is that "venture can be a team sport." Emergence took it to an extreme — all enterprise software investors, no security, no infrastructure, no consumer — but the point generalizes: the partnership has to be able to hold one conversation. A firm with a biotech partner next to a consumer partner next to an enterprise partner is not one conversation, he said, and then "there's really no point in being a partnership."
The third is succession. "But how do you build something that outlasts any single individual or partner on the team?" He said it takes restraint on economics and a willingness among people with power to "do the George Washington, thing" — several partners there could have been kings of the firm and instead passed it on.
5. The Partner Meeting Score
Asked whether Fika is consensus-driven or one-partner-bangs-the-table, Chen described a scoring system built to surface disagreement rather than suppress it.
"I think the number one north star is truth seeking," he said — the only objective in the partner meeting is ground truth, including from the firm's junior investors.
After a company pitches the full team, everyone fills out a survey: "We have probably like 15 different dimensions, and then a final score. It is a rating scale from one to seven, and the presence of conviction is what we look for, which is someone has to be a seven on the team, like especially on the deal team."
No seven on the deal team means the firm should not be looking at the deal at all.
The second half of the test is what happens after the objections land. "There's actually 99 reasons to not do you know an investment for every one to do it," he said, and the question is whether the person who scored a seven is still excited having read all of them.
The reason everyone scores, he said, is that there is no point having a junior investor in the room if their voice is not heard.
6. Where AI Services Work
Emergence was early to AI-native services, a category Moran called popular and polarizing. Chen said Fika invests in it, is co-hosting an event on it with Emergence in the fall, and applies one filter.
"But I think the one thing I'll say here is the most important filter for us is we're trying to find spaces and ideas where there is a better 10x there's a 10x product experience, better product experience that can be had," he said.
Two examples clear that bar. Hanover Park is attacking fund administration, where he says there is "a very low bar" and "plenty of headroom for a 10x better experience." Crosby took contract redlining for companies with no general counsel — a high-stakes workflow where the big firms did not want the work and the small service providers were not good enough.
"But I think where people go wrong is trying to go after some of the higher end premium, segments of the service market," Chen said.
Litigation is his counter-example: it is extremely high stakes, and what a client buys there is not speed or price but "paying for the brand name and the experience."
Moran's shorthand — "Maybe having a throat to choke." — and Chen agreed: with a named partner, someone's reputation is on the line for the outcome, and with an AI-native service "there is no person that is ultimately responsible."
7. Copilots Aren't Dead
Moran asked the inverse question: when is selling an AI tool to a human worker still a venture-scale business, rather than replacing the human with a service?
"I think the yeah the death of copilots is widely overstated in our opinion," Chen said.
He accepted the common wisdom holds for most white collar knowledge work — do not sell tooling to the human, sell the machine that does the work — but drew a line at physical work.
The category he cares about is "where bits meets atoms." In robotics and manufacturing automation, AI will make it faster and cheaper to build new machines, which is the right play.
Servicing them is different: he said "the servicing of these machines will likely still be done by humans because the problems are too multivariate and also hyper local." Getting a robot to a specific place to fix something does not make sense.
The form factor he has seen a few companies build is "the AI safety goggle" — it helps the technician diagnose the machine and then carry out the repair.
8. Majors and Minors
Asked whether that implies deep tech investing, Chen made a joke about the year and then explained how Fika is handling it.
His joke about the year: "All the software investors have become deep tech investors in 2026."
Fika's structure for it is "majors and minors." The major is still the application layer, where he says there are long-tail categories that are "sneakily large" and overlooked.
The minors are areas the firm is learning: he has spent time in the last few months with quantum companies, nanofiber companies and drug discovery companies. What has opened them up, in his view, is AI collapsing the cost structure and the cycle time of getting a product to market.
The pattern he is repeating is backing a market before it exists. Sift, which he calls "the data dog for hardware," was selling into a market that did not yet exist when Fika backed it, and the firm was not sure there would be enough hardware companies needing a telemetry tool. There were.
Field AI is the other one — autonomy software for Boston Dynamics robots, backed several years ago, before robotics was mainstream, and since the subject of a large funding round he says involved Temasek and Jeff Bezos.
Deep tech teams, he added, are a version of the outsider pattern: they often do not know how to tell the story to a mainstream investor, which takes work to get past but leaves opportunity.
9. Is Seed the Sucker Round?
Fika's entry point is seed, and Moran asked the blunt question: does entry price matter there?
Yes, Chen said, and it is debated inside the firm and outside it. "I think a lot of folks think that price doesn't matter as long as the outcome is big enough, and I think that is true to some extent," given how fast companies are now growing.
The worry is the spread: he said "there is a massive dispersion in seed valuations today, such that like the thing that haunts us ultimately is, is seed or pre-seed particularly the new sucker round."
He has a historical reference point. "I think back in 2014 15, back in the emergent days, like the Series B was considered the sucker round," because it was not far enough de-risked from the A to justify the higher price, and not yet obvious the way the C was. The definition of the sucker round is the one where you are not rewarded for the risk you are taking.
The shape that worries him now: "Where it's like companies, pre-seed companies raising 10 to $15 million, at the valuations that you were going to expect for those size rounds." Moran: "it's pretty wild."
Asked whether Fika has done one of those rounds, Chen was direct. "We haven't yet to date. We've seen a couple. We've been tempted to on a number of fronts," and the founders raising them are often excellent. "So, anyways, we haven't done one yet. We are open-minded, but the bar is high."
Moran gave his own data point from the other side: across his last 30-odd deals, up rounds were coming in at 30 to 4x trailing twelve-month revenue, which prompted an LP to ask what kind of multiples he was seeing.
Chen's hope is that seed still has dispersion — some rounds are day minus one, others are much further along — whereas the Series A has thinned out, and there a company is either "chosen and anointed on the track" or has to find another path.
10. The Kingmaker Problem
Moran described the pressure founders now feel to be anointed by one of the top firms, and gave a live example from his own portfolio: a seafood distribution startup whose competitor got king-made and now has a pile of cash, even though his company is winning every bake-off.
Chen's frame is two competitions running at once. "it used to be that capital markets would wait until there's some evidence that the battle of the customer markets has been settled, that the customers have chosen the king, and the market leader emerges."
"But now that's flipped, and the VCs are choosing who the king is, and capital has become a weapon," he said, pointing at legal AI companies that have raised a billion dollars. Even a fast-growing rival, he said, has to answer the question of "how do you compete against a billion."
His answer is to pick categories where the product decides it: "I think our hope and our thesis is that if you pick a high stakes enough category that matters to the customers, the best product will always win." Video conferencing was the old-guard example — it either works or it does not, and it is critical infrastructure.
The counterintuitive evidence is the length of the bake-off. Ivo, he said, wins against sometimes 10 or 15 vendors, and the competition on the customer side is "an indication of how important that workflow is" to the customer. A company willing to trial that many vendors is not bargain hunting and not following the herd: "They're actually doing the work to understand is this the right product for me."
Moran pushed on whether product still beats distribution in an AI-native era. Chen said both, but held the line on product: where a customer will evaluate 10 vendors, they are looking for something specific, and "as long as that specificity exists" — and the products genuinely differ in meeting it — the best product can win. He was careful that this is not true of every industry or use case.
11. Fear and Greed in AI Spend
Moran raised something he said feels new: customers buying two, three or four competing products in the same category at once.
"This is never before have we seen something like what we're seeing now, where there is there's both fear and greed," Chen said.
The fear half: "The fear part of it is I think a lot of companies are afraid of missing the wave, and not exactly sure what's going to work. So buying duplicative solutions is the answer, at least for now." He conceded Fika does the same thing internally, paying for several tools that overlap in function, because nobody knows which one will work.
The greed half is the hope that one of them turns out to be the whole answer: "There's a greed component as well, where like it is like the form factor of AI products is such that like there is this hope that it will be the one solution to solve everything or unlock a new lever in the business that's never been there before." Everyone, he said, is throwing a hail mary.
He expects a culling, for two practical reasons. One is the "cognitive load in managing so many platforms" at once, particularly for multiplayer workflows at large organizations. The other is cost. "There's no way that it makes sense to have that many mega platforms that the entire company runs on."
Until the customer side knows which to pick, he said, whoever has the budget will simply buy all of them and see what happens.
12. The Metrics That Matter
Moran said the SaaS-era benchmarks — quick ratios, the growth-versus-margin trade-offs — have not settled into an AI equivalent yet, and asked what Chen looks at.
On margin, the current wisdom is that "margins can be fixed over time" as long as there is some hygiene level. "And again, the SaaS hygiene level was 80%," he said. Today's companies are nowhere near that: "Like most people are happy with, the teens."
"growth trumps all things" is the one obvious thing happening in the market right now, he said, adding that margins being fixable over time is a valid view, and that newer models are already making the economics work.
At the early stage, quick ratios do not have enough signal, so he looks at two numbers instead, and Ajax is his example. Ajax has "a 97% conversion from pilot to paid." "So any customer, any firm that tries them for two weeks to four weeks ends up paying like 97%," he said.
The second is retention with no lock-in: "0% logo churn." "They have never lost a customer." Many of the early customers were month to month, he said, so the customer had to choose Ajax again every month rather than being tied into a multi-year contract.
Fika invested in December and the same metrics have held all the way through as the company has grown, which Chen treats as the real product-market-fit signal.
13. Founder as a Career Path
Moran read out a post from General Catalyst's Niko Bonatsos arguing that a technical person in their early twenties in San Francisco can now raise a first round from VCs more easily than land a job at the right company, and asked what Chen makes of the trend.
The trend is real, Chen said, and the access is good: "I actually I do think it's actually great for this country, for the world, for society that we can empower young people to take a chance on themselves," because becoming a founder used to require being an insider or being crazy.
His objection is to the framing as a job — there is something off about "institutionalizing founder as a career path." "Ultimately, it should be the exception and not the norm because it is a hard, hard journey."
He accepts the structural reasons it is happening: a new super cycle rewards young people who adopt fast and push the edges of a new platform.
What he thinks has been lost is a mutual obligation. He cited the Sequoia meme about a partner telling Jensen Huang that if he lost the firm's money he would kill him, and said: "There was something at stake between the founder investor understanding where the investment is not a research grant."
"There's an expectation of a return on that investment because the VCs are investing on behalf of LPs that really matter to us, as fund managers," he said, and nobody is paying for a founder's learning.
He blamed the drift on capital concentrating in mega platforms that can afford call-option bets on less proven teams, and allowed his own view is romantic: in a rational market both sides have something at stake, which is what makes an investor do everything possible to help the company work. What is at stake on both sides, he said, has gone hazy amid the hype in the funding environment. Moran's addition: adding "founder" to a profile does not make someone an entrepreneur.
Bonus Insights
Asked who the show should interview, Chen named a writer rather than an investor: "I really enjoy Will Manidis on X, and he just writes like I love great writing, and I think he has a very unique and very entertaining approach." He said Manidis draws parallels from tech to esoteric subjects such as 15th-century European political dynamics, and that he has not been able to find a podcast appearance by him — the questions he would ask are how he comes up with the ideas and how he approaches writing.
Both of his book recommendations are history rather than business. "The first is the Lords of Easy Money. This is a book that's about the history of the Federal Reserve, and that in itself sounds horribly boring." It is not, in his telling — character-driven, about the personalities that shaped the institution and the consequences of their decisions. "It's it's actually it's strangely a page turner." He read it in about a day. The second: "The other one is King of Kings, and this is about the Iranian Revolution, and it's about the cast of like five characters around the event that made it all happen" — a highly improbable event driven by human personalities.
Moran's related observation was that tech history reads as inevitable in hindsight — Google, eBay, PayPal — when it was the people who shaped which companies succeeded and when. Chen agreed and used it to recommend Acquired: "Ben and David do an incredible job, and unpacking that very specific question, which is like, yeah, at some point these companies were not inevitable and not obvious."
Asked for a secret-weapon habit, Chen said he does not have one — there is a two-and-a-half-year-old at home, so "my habits and rituals are out the door." Everything runs around preschool drop-off, pick-up and weekend birthday parties. The one thing that survived is the office. "I come into the office five days a week when I'm not traveling, and that is the protected sacred time." The nine-to-five between drop-off and pick-up has become his most productive block.
On his TechCrunch years, which Moran teased him about, Chen said being a journalist was one of the hardest things he has ever had to do. He does not publish much now, and said it maybe makes venture look easier.
Chen's bottom line is that the market has stopped waiting for customers to pick the winner, and that a seed firm's only defense is to buy things the crowd cannot price — outsider founders, categories where the buyer will run a real bake-off, and early evidence like pilot conversion and churn that says a product is being chosen rather than funded.
Products, Companies & Tools Mentioned
Fika Ventures (Chen's firm: eight to 10 investments a year, leading or co-leading, a third of the portfolio in Los Angeles)
Emergence Capital (Where Chen invested from 2014 to 2018; he calls its posture "aggressively non heat seeking")
Zoom (The investment made in his first two weeks at Emergence, when it was the ninth entrant in video conferencing and its founder was under-reporting his own growth)
Ivo (The legal AI company whose founder, a New Zealand corporate lawyer with no YC or Stanford on his resume, closed three general counsels in 20 minutes on a demo; customers now include Meta, Uber, Google, IBM, Notion, Quora and Reddit)
Ajax (AI legal timekeeping, sourced from a referral by a founder in the same WeWork; 97% pilot-to-paid conversion and 0% logo churn)
Hanover Park (Fund administration, which Chen says has a low enough bar to leave headroom for a 10x better experience)
Crosby (Contract redlining for companies with no general counsel — a high-stakes workflow big law would not touch)
Sift (Described by Chen as "the data dog for hardware," backed before a market for hardware telemetry existed)
Field AI (Autonomy software for Boston Dynamics robots, backed before robotics went mainstream)
Harvey and Legora (Moran named both as legal AI products a single law firm now buys side by side; Chen cited "the Harveys that have raised a billion" as the shape of the kingmaker problem)
SpaceX, ServiceTitan and Snap (The 2017 thesis: LA companies whose alumni would start B2B businesses. Fika has backed its third ex-SpaceX team and fourth ex-ServiceTitan team)
Box and TechCrunch (Where Chen was chief of staff and a writer before venture)
Books & Resources Mentioned
The Lords of Easy Money – Christopher Leonard (A history of the Federal Reserve that Chen says "sounds horribly boring" and is "strangely a page turner"; he finished it in about a day)
King of Kings (The Iranian Revolution told through the five characters around it — an improbable event driven by personalities)
Acquired (Chen's media recommendation, for showing that companies were not inevitable or obvious at the time)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: