
Gokul Rajaram: How to Analyse for Durability and Defensibility in a World of AI
Gokul Rajaram (guest), Harry Stebbings (host)
In this episode of The Twenty Minute VC, featuring Gokul Rajaram and Harry Stebbings, Gokul Rajaram: How to Analyse for Durability and Defensibility in a World of AI explores eight-moats framework for durable SaaS investing amid AI disruption today Rajaram argues public markets are overreacting to AI by treating all software as doomed, while durability depends on specific moats and compounding advantages.
Eight-moats framework for durable SaaS investing amid AI disruption today
Rajaram argues public markets are overreacting to AI by treating all software as doomed, while durability depends on specific moats and compounding advantages.
He proposes an “eight moats” scoring framework—data, workflow, regulatory, distribution, ecosystem, network, physical infrastructure, and scale—suggesting 4+ moats implies strong defensibility.
He distinguishes superficial “bolt-on AI” from real AI-native product reinvention, emphasizing end-to-end UX redesign, tuned models, and rapid iteration as model capabilities change.
He contends vertical AI businesses can still reach $10B outcomes, but typically only by owning the full stack and expanding from software budgets into labor/BPO budgets.
On venture mechanics, he says early price matters less at seed/early A if you’re right, but later-stage entry price can cap returns; he recommends selling based on go-forward IRR and using secondaries thoughtfully.
He predicts switching costs and brand moats weaken as data portability and pixel-perfect cloning get easier, forcing incumbents to commoditize complements and reposition pricing toward outcomes where work is being done.
Key Takeaways
Start with a truly remarkable product; distribution can’t rescue mediocrity.
From Google, Rajaram’s core filter is whether the product is 10–100x better (e. ...
Get the full analysis with uListen AI
Use a multi-moat scorecard; single moats are fragile in an AI world.
He scores companies across eight moats (data, workflow, regulatory, distribution, ecosystem, network, physical, scale) and claims 4+ moats is “pretty damn secure,” while 0–1 is existentially risky.
Get the full analysis with uListen AI
Workflow moats vary by depth; ‘runs the business’ beats ‘helps a team.’
ERP-like embedding (NetSuite) creates higher switching friction and operational dependence than lighter tools (e. ...
Get the full analysis with uListen AI
Most ‘bolt-on AI’ fails unless it changes the product’s frame, UX, and economics.
Adding a thin GPT layer has a ceiling; winners redesign the end-to-end experience around new capabilities (e. ...
Get the full analysis with uListen AI
Brand and classic switching costs weaken as portability and cloning improve.
He expects agents and improved tooling to make migrations easy and experiences replicable “pixel by pixel,” reducing lock-in; incumbents must rely on deeper moats (ecosystems, regulation, distribution, proprietary data) and ship faster.
Get the full analysis with uListen AI
Vertical AI can still build $10B companies—but only by owning the full stack.
One-function vertical agents may be viable but rarely huge; ServiceTitan is cited as a model (dozens of products). ...
Get the full analysis with uListen AI
Pricing will bifurcate: ‘access products’ keep seats; ‘work products’ move to outcomes.
Seat pricing persists for predictability (e. ...
Get the full analysis with uListen AI
For slow-growing, overvalued SaaS, the path is consolidation or ‘burn the boats’ reinvention.
He predicts many high-priced, slowing companies become zombies or PE outcomes; the better play is launching an AI-native new product line, migrating customers aggressively, and abandoning sunk-cost thinking (Intercom/Fin cited as an archetype).
Get the full analysis with uListen AI
Early-stage price matters less than being right; later-stage price can destroy returns.
Rajaram will “pay up” at seed/A with conviction, but warns that at B+ and beyond, buying great companies at too-high multiples can still yield poor outcomes (stagnant valuations despite revenue growth).
Get the full analysis with uListen AI
Sell decisions should be driven by go-forward IRR, not just MOIC narratives.
He argues many firms overweight paper multiples and underweight time; if go-forward IRR at a liquidity window is below what LPs should expect, he believes you owe them partial de-risking (often enabled now by stronger secondaries).
Get the full analysis with uListen AI
Notable Quotes
“The market has decided that since code is becoming free… every software company is going to zero. I think this is one hundred percent an overreaction.”
— Gokul Rajaram
“I call it the eight moats… And I think anything four or more, you’re pretty damn secure.”
— Gokul Rajaram
“You cannot be a single product company.”
— Gokul Rajaram
“Switching costs is just going to go to essentially zero… you’ll have clones popping up left, right, and center.”
— Gokul Rajaram
“Most early-stage firms get wrong… they just focus on MOIC, they don’t focus on IRR.”
— Gokul Rajaram
Questions Answered in This Episode
In your eight-moats scoring, how do you operationally measure a ‘data moat’ versus marketing claims (e.g., what qualifies as truly proprietary and compounding)?
Rajaram argues public markets are overreacting to AI by treating all software as doomed, while durability depends on specific moats and compounding advantages.
Get the full analysis with uListen AI
You rated Atlassian ~3 moats and Monday ~1; what concrete product or GTM moves could add a fourth moat for Atlassian (or Monday) in the next 24 months?
He proposes an “eight moats” scoring framework—data, workflow, regulatory, distribution, ecosystem, network, physical infrastructure, and scale—suggesting 4+ moats implies strong defensibility.
Get the full analysis with uListen AI
If switching costs trend toward zero via agents and migration tooling, which categories of SaaS become structurally unattractive, even if they have strong retention today?
He distinguishes superficial “bolt-on AI” from real AI-native product reinvention, emphasizing end-to-end UX redesign, tuned models, and rapid iteration as model capabilities change.
Get the full analysis with uListen AI
What’s the clearest ‘tell’ that an AI feature is thin bolt-on versus a real reframing of the product’s core job-to-be-done?
He contends vertical AI businesses can still reach $10B outcomes, but typically only by owning the full stack and expanding from software budgets into labor/BPO budgets.
Get the full analysis with uListen AI
You excluded brand as a moat; what evidence would change your mind, and are there B2B categories where brand still functions as distribution?
On venture mechanics, he says early price matters less at seed/early A if you’re right, but later-stage entry price can cap returns; he recommends selling based on go-forward IRR and using secondaries thoughtfully.
Get the full analysis with uListen AI
Transcript Preview
I call it the eight moats. Data moat, workflow moat, regulatory moat, distribution moat
We're on number five. I'm loving this. [laughs]
Ecosystem moat, network moat, physical infrastructure, and the eighth one I would say is scale moat. And I think anything four or more, you're pretty damn secure.
[air whooshing] I'm thrilled to welcome one of the best operator turned investors of the last two decades, Gokul Rajaram.
You cannot be a single product company. Vertical products, you've got to really own full stack. It's harder otherwise to be a 10 plus billion dollar company.
What's the biggest miss? Is it Banter?
Most recently, Quince, but to be honest, even bigger miss than that in some ways. It's not a miss in terms of investing, it's that-
[clapperboard snaps] Ready to go? [upbeat music] Gokul, I've wanted to do this for years, and you have... I mean, you've coyly played hard to get, let's put it mildly, uh, after my continuous WhatsApp messages. But thank you so much for joining me today, man.
It's my pleasure to be here, my friend. Thank you again.
Now, I wanted to start with how some of your prior companies that you've worked at have shaped your investing mind specifically, and I wanted to start with Google. When you reflect on your time with Google, how did that shape your mindset for the types of companies that you like today?
I think the best way to think about my, the Google experience is Google taught me that ultimately the best companies have a remarkable product at their core. Google was a remarkable product. Google was definitely a philosophy of build it remarkably and they will come. Uh, GTM was not Google's specialty, but what Google was really good at was building amazing products. Sometimes the go-to-market worked, sometimes it didn't work. So but at the core was a remarkable product. So I think I... Ultimately, my core investing thesis is that if there is not a remarkable product, all the go-to-marketing distribution in the world will not save you. So that's-- I look for what the remarkability is in the core product or value proposition of the company. Is it 10X, 100X better than the alternative? I'll tell, I'll tell you a story at Google. When I joined in 2003, there was a project going on called, um, Caribou internally. I was like, "What the hell is this?" This was web email which gave one gigabyte free storage, and back then Yahoo Mail offered 10 megabytes of storage, so it was 100X. I thought it was truly... I was like, "There's no way it's possible." And it turns out it was, and it was released on, if you remember, April 1st, 2003, and that people thought it was an April Fool's joke, but that was Google literally taking something that was unbelievable and making it reality. And so that's, that's the kind of products I like, something remarkable, something unique, something powerful.
I like it. It reminds me of actually Neil Mater, who talks about kind of jaw-dropping customer experience as one of his kind of core monikers for thinking about companies and investments. [laughs] Next, we have Facebook. How did Facebook impact the types of companies that you like?
Install uListen to search the full transcript and get AI-powered insights
Get Full TranscriptGet more from every podcast
AI summaries, searchable transcripts, and fact-checking. Free forever.
Add to Chrome