Jun 20, 2026 · 1h 20m · 20vc
Flexport CEO: Two Questions Every Founder Needs to Ask · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 20VC, host Harry Stebbings sits down with Flexport CEO Ryan Petersen to discuss the realities of scaling a global logistics giant, the transition to AI automation, and his controversial views on remote work.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 30.3% of the talking time here. How this is scored →
Harry as informed peer 4.0 Guest teaching 3.2 Guest disagreement 2.1 Harry pushing back 2.2
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 37:48 Ryan calls remote work white collar fraud
Ryan aggressively rejects conventional remote work arguments, declaring that working from home with young children present is a fantasy and calling remote work white collar fraud.
Hardest push from Harry ▶ 35:25 Harry forcefully pushes back on fundraising adviceHarry explicitly refuses Ryan's cautious framing on fundraising advice, stating that he totally disagrees and asserting that taking discounts for tier-1 VCs is concrete universal advice.
Biggest teaching moment ▶ 24:01 Ryan educates Harry on logistics market proportionsRyan re-educates Harry on market dynamics by explaining that logistics represents 11% of global GDP, dwarf software spending for physical goods companies.
Harry holds his own ▶ 1:02:49 Harry details internal VC associate collusion networksHarry demonstrates insider domain authority by breaking down how junior VC associates run cross-firm WhatsApp groups to pass deal flow metrics and influence fund perceptions.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Motivation: Fear of Losing vs. Thrill of Winning | 1 | 1 | 1 | 1 | Host opens with standard motivational questions about fear of losing versus thrill of winning and sets up a lighthearted game with charity donations for repeated questions. The dynamic is purely conversational and rapport-building with no conflict. | |
| Why VCs Behave Like Herd Animals | 5 | 3 | 2 | 2 | Ryan explains why venture capitalists behave like herd animals due to job security and fear of looking foolish. Harry demonstrates domain familiarity by adding points about 10-year fund management fee guarantees and quoting Keith Raboy. | |
| Financial Security, Risk, and 'The Number' of Happiness | 4 | 3 | 3 | 2 | Harry asks Ryan about his financial number for happiness and ties success in VC directly to financial returns. Ryan gently pushes back on financial motivation, noting that he is driven by power and doing big things rather than money. | |
| Flexport's Revenue Milestones and Paul Graham's Two Questions | 4 | 3 | 1 | 2 | Harry presses Ryan on net revenue numbers, growth rates, and potential concerns around a constrained public market exit landscape. Ryan outlines Flexport's metrics and Paul Graham's framework for sustainable startup growth. | |
| AI Foundation Models: Value Concentration and Operational Dependency | 5 | 3 | 1 | 2 | Harry questions whether value concentration in AI foundation models is dangerous, citing stock market gain statistics and bringing up Robotic Process Automation. Ryan describes Flexport's practical implementation of AI agents for business logic. | |
| AI Costs: Budgeting for LLMs vs. Open Source Deflation | 7 | 2 | 1 | 2 | Harry demonstrates strong knowledge of AI enterprise spending, citing Marc Benioff's $300M Anthropic spend and percentage benchmarks needed to justify trillion-dollar valuations. Ryan details Flexport's LLM spending and potential shift to open-source models. | |
| The Threat of Compute Hoarding and the Future of Workforces | 5 | 3 | 3 | 4 | Harry pushes back on Ryan's hypothetical compute hoarding scenario by noting current AI product limitations and citing productivity skepticism from Uber and Microsoft. Ryan explains why labor-heavy freight forwarding differs from automated tech companies. | |
| Embedded Employees and Capitalizing on Competitor Incompetence | 4 | 4 | 4 | 2 | Harry cites an interview with the founder of Invisible asserting that enterprise sales require forward-deployed employee models. Ryan explicitly rejects the claim as untrue, while acknowledging the competitive advantage of deeply embedded personnel. | |
| Logistics as 11% of GDP: Comparing Physical and Software Budgets | 3 | 5 | 2 | 2 | Ryan educates Harry on the scale of global logistics as 11% of GDP, contrasting physical supply chain spend with corporate software budgets. Harry asks clarifying questions on model usage between OpenAI Codex, Claude, and Gemini. | |
| Open-Source AI, the CCP, and US-China Co-Dependence | 4 | 4 | 4 | 2 | Harry brings up concerns from Keith Rabois regarding CCP-backed open-source AI models leaking Silicon Valley data. Ryan dismisses the threat framing, citing his time living in China and arguing that US-China mutual economic dependence makes military conflict unlikely. | |
| The Series B Preemption Mistake and Peter Thiel's Lifeline | 3 | 3 | 1 | 1 | Ryan recounts his Series B fundraising blunder where an attempted preemption fell through, forcing him to admit the mistake to Founders Fund before Peter Thiel stepped in to lead the round. Harry listens attentively with minor commentary. | |
| Debating Valuation vs. the Value of Tier-1 VCs | 7 | 3 | 4 | 7 | Harry strongly challenges Ryan's hesitance to give generic fundraising advice, citing Brian Chesky and forcefully arguing that tier-1 VCs are always worth taking a valuation discount for. Ryan maintains that generic fundraising advice can be dangerous. | |
| The Debate Over Remote Work vs. In-Person Culture | 3 | 4 | 5 | 2 | Ryan takes a blunt stance on remote work, calling it white collar fraud and explaining why productivity drops at home with family present. Harry agrees on in-person interview dynamics and listens as Ryan describes returning Flexport to an in-office baseline. | |
| Why Executive Hiring and B2B Marketing are Hard | 4 | 4 | 2 | 2 | Harry highlights an example of an enterprise CMO lacking a Twitter presence to illustrate executive hiring pitfalls. Ryan explains why B2B enterprise marketing is exceptionally difficult due to rapid copying of creative breakthroughs by competitors. | |
| Rethinking HR and the Utility of AI Agent Resources | 4 | 3 | 2 | 2 | Harry cites Adam from AppLovin saying every great CEO hates HR, prompting Ryan to provide a nuanced perspective on HR alignment. Harry introduces the concept of agent resources replacing traditional human resources as SaaS contracts face renegotiation. | |
| The Realities of Angel Investing and Revenge Backing | 3 | 3 | 1 | 1 | Ryan shares his early angel investing history in his Y Combinator batch, including hits like Algolia and Rippling, as well as missing Cruise. Harry and Ryan discuss the investment thesis behind backing wronged or vengeful second-time founders. | |
| Honesty in Venture Capital: Direct Feedback vs. Ghosting | 4 | 4 | 2 | 3 | Harry outlines his blunt rejection philosophy as an investor, while Ryan quotes Paul Graham on ignoring why VCs pass. Ryan describes walking out of a pitch where investors incorrectly calculated Flexport's total addressable market. | |
| The Hero's Journey, Corporate Scandals, and Constant Curiosity | 7 | 4 | 2 | 4 | Harry explains how junior VC associate networks collude across firms using WhatsApp groups to exchange intelligence on founder pitches. Ryan shares a story about a Bloomberg article covering a single lost parcel delivered via DoorDash. | |
| Sustaining a Marriage, Pitching Masa Son, and Finding Purpose | 3 | 3 | 1 | 1 | Ryan reflects on marriage, pitching SoftBank founder Masa Son in Woodside, and the life purpose gained from raising children. Harry shares his childhood inspiration for entering venture capital through watching The Social Network. | |
| Shifting Strategy and Embracing Price Competition | 3 | 3 | 2 | 2 | In a quick-fire round, Ryan explains his strategic shift toward becoming a low-cost provider in logistics. He also lightheartedly describes wanting to sponsor German football club St. Pauli as a troll against a competitor. | |
| Aggressive Growth and Artificial Intelligence Milestones | 1 | 2 | 1 | 1 | Ryan details his target for 2026 success, specifying that Flexport needs to successfully operationalize 80% of its planned AI agent workflows. Harry wraps up the interview warmly. |
Statements from this episode (64)
Petersen: Selling SaaS to tech companies will become the toughest market
“I think selling SaaS to tech companies is gonna be the toughest, because we can build stuff for ourselves.”
Ryan Petersen: Revenge and patriotism make a great investment thesis
“Revenge and patriotism is a great investment thesis.”
Petersen: Most VCs collude more with competitors than their own partners
“There's a lot of collusion in VC. Like, I have a feeling that most VCs actually collude more with competitors than with their own partners.”
Petersen: Flexport generated over $2B in revenue last year
“My first Excel my first and only financial model that I made for myself for the business got to one million of revenue, right? Like, we did over two billion last year, so.”
Stebbings: Richer VCs perform better by focusing purely on upside
“I think that richer investors make better investors, and you're like, Sequoia are focused on upside maximization. How big could this be? They are not fearful of LPs coming back for their next fund. They're not fearful of deployment speed compressing, so, you k…”
Ryan Petersen: Driven by power and doing big things, not money
“Yeah, I mean, I haven't been that motivated by money. I'm more into, like power. I want to do big things.”
Petersen: Flexport expects to reach break-even this year
“I think this year we're gonna get to basically break even.”
Petersen: Flexport reaches $450 million net revenue run rate
“We're on run rate to do about four hundred fifty million dollars of revenue, net revenue. Breaking news here, I haven't shared that, but four hundred fifty million of net revenue.”
Petersen: Flexport aims for 30% annual revenue growth for 10 years
“And I think we'll do 600. I mean, the goal is to keep going 30% every year for 10 years.”
Petersen: Flexport holds less than 0.1% of global logistics market
“We're still less than one percent.1%.”
Petersen: Flexport will IPO in a couple of years after hitting a few hundred million EBIT
“I'd like to be a nicely profitable, making a few hundred million of EBIT and then we'll go public. So it could happen in a couple of years. It should, like the way that things are going right now, we should be able to get there.”
Petersen: OpenAI and Anthropic deserve high valuations for miraculous products
“Like they deserve it. They're this, these things are just incredible, miraculous products.”
Petersen: AI agents will enable full end-to-end logistics automation at Flexport
“But agents can go all the way. Like, we think we can automate the full end to end.”
Petersen: Flexport will shift to open-source AI models to lower costs
“We're using OpenAI. We're using Anthropic. We're going to increasingly use open source models to save money, but also because we don't today have a budget constraint on our Anthropic contract.”
Petersen: Flexport spends $5M annually on LLMs, spend doubled recently
“I think we're spending like probably like five million a year or something ramping pretty fast though. Like it's doubled. The last few months.”
Petersen: Flexport will spend $20M annually on LLMs within five years
“Yeah, probably. I mean, I don't know about Anthropic, but on LLMs.”
Petersen: Frontier AI is required for coding, wasteful for basic workflow automation
“For our coding agents and, like, for actually developing product, we probably always want the frontier model, but if you're just, Automating a workflow. If the open source one that's basically free can automate it, then I don't, there's diminishing returns to …”
Petersen: AI providers might cut off customer APIs to train superintelligence
“There's a realistic scenario that they decide the compute that they have is more valuable for training super intelligence than it is for letting customers use it. And they stop, you know, just cut us off.”
Petersen: Flexport headcount will stay flat as AI replaces manual operations
“It might be about the same. We're hiring a lot of salespeople and we're expanding a lot in new markets, but I think you're going to see like a shift from people that today do operations, kind of manual work that I was describing earlier into, okay, you've got …”
Petersen: AI can double salesperson capacity, but sales remains relationship-driven
“And like, there's a limit to how many relationships a human can have. And I think you can probably use those to use AI and use other techniques to like double that ratio or something, but like there's some limit to how many relationships you can maintain. And …”
Petersen: Flexport competitor embedded 130 staffers at hyperscaler, blocking rivals
“I mean, one of the big tech companies that runs like a hyperscaler, let's call it has one of our direct competitors has 130 full-time employees that work at that company that have badges that show up to work every day that run their logistics for them. And we …”
Petersen: Flexport won major client after competitor laid off embedded worker
“And we had this recent, we just had this huge customer win, one of our biggest ever. And I was like, wow, this is great. But I had heard that our competitor makes their Has one of these forward deployed logisticians, like, and runs their RFP, runs their proces…”
Ryan Petersen: Flexport's largest customer spends $150M annually on logistics
“Spend about a hundred fifty million a year.”
Ryan Petersen: Flexport has a dozen customers spending over $50M annually
“We don't have too many like that, but we have probably a dozen that spend over 50, and then a long tail.”
Ryan Petersen: Physical goods companies spend 5% of revenue on logistics
“The logistics is a much bigger market than software. If you, for example, like a typical, like a medium sized company will spend, call it five percent of their revenue, companies spend about five, if they're physical goods companies, they'll spend like five pe…”
Flexport's logistics business serving data centers is growing rapidly
“I mean, by the way, our competitors, like us and others that are scaling, that are serving data center companies, that part of the business is growing like crazy, because there's data center, all this physical, the data centers have to be built.”
Petersen: Anthropic is a better investment than OpenAI at equal valuation
“I think I'm going Anthropic. No offense, because Sam's a friend of mine, but. Yeah, the enterprise business seems great. And the team, I think there's a lot of value of like this cohesive team that's been together forever, and OpenAI's had ups and downs, and p…”
Petersen: Peter Thiel offered Flexport $300M valuation after Petersen asked for $275M
“Best offer I got was two 75, but the firm wants board control. So if you'll just do it without taking the control of the board, you guys can have it. And Peter just offered 300 instead of two 75, which wasn't as good as my 500, but he didn't need to do that. I…”
Petersen: Generic fundraising advice is dangerous for startup founders
“It's very hard to give generic fundraising advice. I think it's very dangerous. Like each founder asked me for advice. I'm like, you know, your situation much better than I do.”
Stebbings: Tier-one VC backing is worth taking a valuation discount
“Tier ones are worth it. Take a discount. They will help you get a next round unwaveringly.”
Stebbings: Top talent joins startups based heavily on VC brand signaling
“The amount of great operators who you're like, well, they wouldn't be, I'm so, I don't mean this rudely, dumb or naive enough to just join a company because of a VC. Do just join a company because of a VC. They'll go, oh, it's a Sequoia-backed company, oh, it'…”
Petersen: Only 4% of Flexport's workforce is in San Francisco
“Like, our team in San Francisco is maybe four percent of the companies based in San Francisco, even though that's where we were founded and where I'm based.”
Petersen: Flexport will push employees to move back to San Francisco
“So I'm like pushing, I'm probably, Gonna start pushing more of our employees like, hey, you should come move back to San Francisco.”
Petersen: Flexport mandates a five-day in-office baseline requirement
“We're five day a week in office as the baseline assumption at our company.”
Petersen: Work-from-home ultimately benefits employers through global labor arbitrage
“The idea that work from home is going to benefit highly paid employees is sort of a total fantasy. Like, work from home done correctly, you should be hiring the world's greatest geniuses. There should be a labor arbitrage here, where you're finding really, rea…”
Petersen: B2B marketing strategies decay immediately when competitors copy them
“In enterprise marketing. The moment something works, it stops working because the competitors will copy it.”
Petersen: High marketing spending serves as a quality signal in enterprise B2B
“And it's expensive as a feature in enterprise marketing. Cause it's like, you wouldn't spend a lot of money on something promoting something if it wasn't good. So it must be good.”
Petersen: Startups do not need a CMO to execute B2B event marketing
“B to B marketing is events marketing to some extent, and like, that, that's like plain vanilla, but by the way, you don't need an exec, you don't need to go hire a CMO to run events.”
Stebbings: Every great CEO hates HR
“Oh, I fucking hate HR, and every great CEO hates HR, and like a complete dick that I am, I thought I'd front run his show and take every great CEO hates HR, and I'd tweet it myself. Oh my god, I got killed By everyone for this tweet, which was essentially Adam…”
Petersen: HR leaders must represent the CEO, not act as employee union reps
“I mean, I think the thing that you cannot have is them being like union representatives of your employees. They're not. They're there to represent the company, and therefore the CEO, and so if there's, if you don't have that as a CEO, like a trusted HR Leader …”
Stebbings: AI agents will largely replace humans, spurring 'agent resources' roles
“That you mine and agents will in large part replace humans, and so the really interesting job function that could be created is like agent resources.”
Petersen: Salesforce's Slack acquisition will prove to be great
“Slack is also going to be a great, turn out to be a great acquisition, because like, that's pretty sticky. Nobody wants to replace, build their own Slack.”
Petersen: Threatening to build in-house tools yields 20% SaaS discounts
“I think you get like 20% out of almost everybody.”
Petersen: Flexport will focus engineering on replacing labor, not SaaS
“I don't want to dedicate that much of our engineering resources to replacing SaaS. Like I want to build, I want to replace work, like expensive labor, and I want to build our product and make it way better. So I don't, I want to, like, focus on our core busine…”
Petersen: Almost all logistics competitors have suffered major cyberattacks
“Almost all of our competitors have had major hacks in logistics. Like, it's been a major target area for cyber, and there's been a lot of, like, ransoms.”
Petersen: Identifying terrible founders in YC is easier than picking winners
“It wasn't obvious who was great. It was obvious who was terrible, though, and I think if you can eliminate the bottom half, like, you're doing pretty well inside of a YC batch.”
Ryan Petersen: I was the first angel investor in Rippling
“I was the first investor in Rippling. Put a big check into that. Did really well.”
Petersen: Most VCs ghosted Flexport after pitch meetings
“I had, like, a lot that seemed like they went well, and then they never invested, and the number of investors who ghosted me, by the way, after the pitch, the vast majority, I would say. Where I pitched him, it seemed great, and then it's just like, never hear…”
Ryan Petersen: Founders Should Ignore VC Explanations for Rejections
“PG said his advice to founders was always, you should hear the no, but ignore the why, because the odds of them telling the truth multiplied by the odds of them being correct are so low that you're just like very little signal in what investors tell you about …”
Harry Stebbings: Multi-Stage VCs Obfuscate Rejections to Retain Optionality
“Especially when they're multi-stage funds, when there's absolutely no reason for them to ever kill optionality. Like, I want you to think that I'm amazing because in case you do inflate, I want you to come back.”
Stebbings: Avoid Investors Who Need Education on Your Market
“If you need to educate an investor on the market, they're not the right investor for you.”
Petersen: Sam Altman Was One of Flexport's Earliest Investors
“Sam was one of our early, earliest investors”
Stebbings: Passing on Great Companies Over Market Size Is My Biggest Mistake
“So my biggest mistake is when I focused on market at all. All of my biggest mistakes in investing have been actually not really like, you know, losing money. It's been turning down great companies, but because I thought that I was smarter than the market.”
Petersen: 3X Angel Returns Have Zero Impact Compared to 500X Outliers
“Once you see the world from the perspective of the angel investor, you realize that we really don't care. Even a three X, like if you take my spreadsheet and this investment made three times the money, like the founder is very proud. Like, Hey, I made you thre…”
Petersen: Mark Angel Investments to Zero Immediately to Avoid Anxiety
“And when I write a check as an angel investor, I would just mark it to zero. I don't care. I just assume that it's a zero. And then that way I don't worry if it fails.”
Petersen: Flexport will never need to raise capital again
“I don't need to raise any money ever again.”
Petersen: Founders should cherry-pick their best metric when pitching VCs
“You should never share your metrics because what you want to do when you pitch a VC is take, you want to cherry pick whatever metric looks the best and pitch that as the important metric for your firm. But once you've shared a metric, you're now committed to t…”
Petersen: SoftBank led Flexport's $1B round and urged undercutting rivals by 10%
“Yeah, well, they led the round of a billion, yeah. And he was like, whatever the price of freight is, you'd just be 10% cheaper than everybody, and then if someone matches you, you'd just be 10% cheaper than that, which is like a terrible strategy. I did not d…”
Petersen: Founder press presence directly helps sales teams close deals
“Yeah, I do. And I've, I, as measured by when I tell my sales team I'm going to do less of it, they always push me to go do more press. Like they say that it helps them get deals and close deals and stuff.”
Petersen: Flexport must become low-cost leader to win logistics market
“For the last decade, I wanted to be, I didn't want to compete on price, and I wanted to be the premium value, premium provider in logistics and thought it would be bad to compete on price. And now I'm like pretty convinced that we need to be the low cost leade…”
Petersen: A passive board is preferable to an active, 'helpful' board
“I don't really want a more active board. They're not super active in the business. They're not like telling us what to do. They don't help that much, but they don't ever hurt. And like, I would much rather have that than like a really helpful board. I don't th…”
Ryan Petersen: Flexport has deployed AI agents for 5 of 100 targeted workflows
“We've got about, well, Probably like a hundred core workflows that are costly that we have AI that we're building AI agents for right now. And like five of them are live and working and saving us money and 95 that are under development.”
Shorts cut from this episode
▶ "Remote Work is White Collar Fraud" · 20VC with Harry Stebbi (@0:00)
▶ San Fransisco's Loyalty Problem · 20VC with Harry Stebbings (@36:37)
▶ "Remote Work is like White Collar Fraud" · 20VC with Harry S (@0:07)
▶ $8BN Founder on Pitching to VCs · 20VC with Harry Stebbings (@1:04:27)
▶ Two Questions Every Startup Needs to Ask on Growth · 20VC wi (@10:26)