Feb 9, 2026 · 1h 20m · 20vc
a16z, Anish Acharya: Is SaaS Dead? Do Margins Still Matter? Why We Are Not in an AI Bubble? · 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 20VC interview, Andreessen Horowitz General Partner Anish Acharya challenges the "SaaSpocalypse" narrative and discusses AI-driven value migration, consumer companion dynamics, and the rigorous deal-making standards of a16z.
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 24.4% of the talking time here. How this is scored →
Harry as informed peer 5.5 Guest teaching 5.8 Guest disagreement 3.0 Harry pushing back 3.8
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 46:52 Direct Rejection of Series A Hardship Thesis
When Harry asserts that Series A is the hardest investing stage due to inflated valuations and sparse PMF, Anish directly shuts down the premise with a sharp 'I disagree' and argues VCs exist specifically to absorb competitive and pricing risk.
Hardest push from Harry ▶ 11:20 Challenging AI App Revenue DurabilityHarry directly challenges the guest's optimism on AI application layer revenue durability by citing developers abandoning Cursor for Claude Code, arguing Cursor could lose half its revenue.
Biggest teaching moment ▶ 31:38 Reframing Market vs Industry BoundariesAnish corrects Harry's assumption that the support AI market is over-saturated by distinguishing between a narrow market and a $500B broad industry.
Harry holds his own ▶ 11:20 Citing Claude Code Cannibalization of CursorHarry leverages real-time ecosystem knowledge regarding user churn from Cursor to Claude Code to challenge the guest's thesis on application-layer revenue stickiness.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Hook: The Software Market is Oversold | 5 | 5 | 4 | 4 | Harry opens with a firm argument that building outside SF (London) offers talent retention and cost advantages. Anish politely but directly rejects the premise, explaining that cities have network effects and SF provides unparalleled density and commitment. | |
| Defining 'Sufficient' Outcomes in Venture Capital | 5 | 6 | 4 | 4 | Harry asks if public SaaS seat contraction justifies the SaaSpocalypse panic. Anish counters with spend metrics, explaining software is only 8-12 percent of enterprise spend and vibe-coding core back-office tools carries too much downside. | |
| Switching Costs and the End of Software Hostages | 6 | 4 | 2 | 3 | Anish describes how coding agents reduce switching costs, turning software hostages into real customers. Harry quotes Alex Rampell's incumbent vs startup framework to ask who wins the distribution vs innovation race. | |
| Value Migration: The Application Layer Wins | 5 | 6 | 2 | 2 | Harry asks why the application layer will capture more value than foundation models. Anish outlines a multi-model paradigm where apps act as necessary aggregation layers across specialized models. | |
| Revenue Durability: Cursor vs. Claude | 7 | 5 | 4 | 7 | Harry aggressively challenges AI app durability, citing developers leaving Cursor for Claude Code and predicting Cursor could lose half its revenue. Anish responds that market ambition expands faster than supply and developer tools will specialize. | |
| Vertical AI Apps vs. Horizontal Foundation Features | 6 | 6 | 3 | 5 | Harry asks if foundation models offering product features threaten vertical apps, and queries whether AI companions withdraw humans from real interactions. Anish argues model labs lack feature surface prioritization and outlines the psychological benefits of AI companions. | |
| User Experience Paradigms: Saving Time vs. Spending Time | 5 | 6 | 2 | 3 | Harry queries UI paradigm changes and defensibility in AI. Anish draws a distinction between high-agency intent-based UIs and time-spending browse UIs, pointing to live proprietary data as a modern moat. | |
| Do Margins Matter? High-Paying Power Users and Compute Subsidies | 6 | 7 | 3 | 4 | Harry asks whether traditional software margins still matter in AI. Anish educates on compute-credit trial subsidies versus 2021 ad subsidies, noting AI power users pay 10x higher subscription prices alongside consumption fees. | |
| Unpacking AI Metrics: Month-Two Retention and Accurate LTV | 6 | 6 | 3 | 4 | Harry brings up the challenge of calculating CAC to LTV in a fast-moving landscape. Anish recommends treating month-two retention as the true baseline and rejects the AI bubble narrative based on capacity economics. | |
| Swapping SaaS Budgets for Labor Budgets | 6 | 7 | 5 | 5 | Harry highlights 50+ venture-funded customer support AI startups to argue over-competition. Anish explicitly challenges Harry's framing, correcting him that support and legal represent $500B industries rather than monolithic single markets. | |
| The Series A Thesis: Sizing Markets and Founder Inertia | 5 | 6 | 2 | 3 | Harry probes TAM analysis and admits to passing on major winners like Deel and Granola. Anish explains that investor underestimation is common and identifies founder inertia as the ultimate mental model. | |
| Winning Deals, Price Elasticity, and the Zero-Loss Record | 7 | 6 | 5 | 6 | Harry asks about Anish's most painful deal loss. Anish claims he has never lost a deal in six and a half years at a16z, prompting Harry to question if his risk aperture is too conservative. | |
| The Disappearance of the Tweener Growth Round | 6 | 6 | 2 | 3 | Harry asks if $500M growth rounds are disappearing and whether triple-triple-double-double metrics are obsolete. Anish explains that fast-growing outliers leapfrog rounds while area-under-the-curve companies like Figma take longer to compound. | |
| Risk Factors and Competition in Series A Investing | 6 | 6 | 5 | 6 | Harry argues Series A is currently the hardest investing stage due to sky-high ARR multiples and thin PMF. Anish directly disagrees, laying out five distinct VC risk types and insisting VCs are paid to absorb competitive and pricing risk. | |
| Authentic Connection vs. Founder Promiscuity | 5 | 5 | 2 | 3 | Harry asks about founder promiscuity and rapid project pivoting. Anish asserts that irrational commitment to a specific problem domain is necessary to avoid opportunistic jumping between hot sectors. | |
| Repeat Founders: Enterprise Dominance vs. Consumer Beginner's Mind | 5 | 7 | 3 | 3 | Harry queries whether serial repeat founders are universally preferred. Anish differentiates enterprise from consumer, showing that repeat domain experts dominate enterprise while beginner mindsets excel in consumer. | |
| Human-in-the-Loop: Sifting AI Agent Overhype | 5 | 6 | 3 | 3 | Harry asks why Anish perceives AI agent overhype. Anish explains that vague human instructions and complex edge cases mandate humans in the loop to move past local maxima. | |
| Open Source vs. Closed Source Models in the AI Frontier | 6 | 6 | 3 | 4 | Harry cites Jason Lemkin's experience with 11Labs to argue high costs will drive AI model substitution. Anish counters that rapidly expanding model capabilities outweigh token costs during the current expansion phase. | |
| Software's Expansion into Discretionary Spend | 5 | 5 | 3 | 4 | Anish argues software will capture 80-90 percent of consumer discretionary spend across therapy, companionship, and professional development. Harry humorously pushes back on missed spend categories like fashion. | |
| How the Best Founders Leverage VC Resources | 5 | 6 | 3 | 3 | Harry asks if elite founders need VCs. Anish explains that top founders know how to leverage VC platforms aggressively, citing Alex Rampell treating a16z partners as a dedicated sales distribution channel. | |
| Reflecting on Investment Mistakes and PMF Self-Deception | 5 | 5 | 1 | 2 | Harry prompts reflection on past investment mistakes. Anish admits his biggest mistake in 2021 was self-deception regarding product-market fit, assuming traction that was not truly present. | |
| The Hardest Decision and the a16z GP Evaluation System | 6 | 7 | 5 | 5 | Harry cites an insider rumor that a16z tolerates missing deals but not losing them. Anish firmly clarifies that a16z mandatorily expects 100 percent deal visibility and 100 percent win rate across target sectors. | |
| Abstract's Brompton: The Best Seed Investor | 5 | 6 | 1 | 2 | Harry opens quickfire questions on top seed investors. Anish praises Abstract's Brompton as a cold-blooded capitalist and maps the AI wave progression from 2022 models to 2026 native categories. | |
| Scaling Ourselves: The Power of Digital Twins and Moldbook | 4 | 5 | 1 | 2 | Harry asks about the significance of Moltbook. Anish explains how digital twins point toward scalable human replication in dating and productivity, concluding on an optimistic note regarding technology elevating human experience. |
Statements from this episode (65)
Acharya: Software market is oversold and vibe-coding narrative is wrong
“The general story that we're going to vibe code everything is flat wrong, and the whole market is oversold software.”
Acharya: a16z Mandates Seeing and Winning 100% of Target Deals
“I don't think we're allowed to believe in luck at Andreessen. We have to see a hundred percent of the deals in our domain, and that we win a hundred percent of the deals that we go after.”
Acharya: 75% of public SaaS companies raised prices post-ChatGPT
“I looked at the data this morning, and if you look at SaaS, public market SaaS companies, 75% have raised prices since ChatGPT was released.”
Acharya: Coding agents dramatically lower enterprise software switching costs
“But now with coding agents, the complexity of transitioning from SAP to Oracle is dramatically lower, the speed, the risk. So that is how I think coding agents shows up in enterprise software, especially amongst public names. Decrease switching costs, more cus…”
Acharya: Incumbents improve legacy categories while startups own new AI categories
“When you have this product cycle and you have a capable incumbent, what happens is they usually make their product better for their existing categories. So Microsoft will make a better word processor than they've ever made. Google will make a better search eng…”
Acharya: AI movie creation will be won by startups, not Adobe
“If you said something like, you know, software movies or AI movie making or sort of AI assisted movies, that's just not a category in which there is an incumbent, and I'm betting that a native company will actually win that. It probably won't be Adobe.”
Acharya: 80% of foundation model capabilities are interchangeable substitutes
“They're all innovating roughly in lockstep. 80% of what they do, I think that they're actually substitutes for, and then there's the open source models, which also do the same things. And then in the 20%, which arguably is where a lot of the value is, they are…”
Acharya: AI app layer captures value by aggregating multi-model capabilities
“Because you live in this world of multi-model, where for some use cases they're substitutes, for some use cases they're actually specialists, there's a lot of value in having an aggregation layer, and that is the apps company.”
Stebbings: Cursor could lose half its revenue this year to Claude Code
“I think there's a chance that Cursor loses half of their revenue this year with the cannibalization of them by claw code.”
Acharya: AI dev tools Cursor, Codex, and Claude Code will all grow
“The desire and demand for software, both to make it and to consume it, is dramatically more than the supply that we have today, and I think there is a developer and developer adjacent archetype for whom Cursor is going to be perfect, Codex as an app, Codex as …”
Acharya: AI foundation and app markets resemble AWS and GCP, not Uber
“I think the foundation model companies look a little bit like that. And I think in the apps layer, you're just going to have people that want to Consume the code they generate through a rich IDE and those that want to be closer to the metal, and that's probabl…”
Acharya: Seemingly Competing Startups Diverge Within 12 to 18 Months
“I think we're in a part of the market where companies are diverging very rapidly. So even companies that appear to be directly competing today tend to be not competing in, you know, 12 months, 18 months.”
Acharya: AI startups will win by building 'weird' products big tech avoids
“So I think that there is a pocket that startups can really thrive in, which is building these weird products that really touch on many core aspects of humanity that the models can reflect, but the big corporations are uncomfortable.”
Acharya: AI companions foster self-reflection rather than social isolation
“I think it does the exact opposite. I think people are able to be more self-reflective and explore aspects of themselves And human relationships that they often just don't have another person to explore these things with.”
Acharya: Chat and dynamic UIs are overstated in consumer AI
“I think, so voice is amazing for enterprise. I think that one dynamic UIs and two chat UIs are overstated in consumer.”
Acharya: Consumer software users want to spend time, not save it
“Most people don't want to save time. They want to spend time.”
Acharya: Browse-based UIs will remain largely unchanged despite AI
“In a world where we have intent-based and browse-based, browse-based largely stays the same. And perhaps the future of intent-based is chat. I'm still a little skeptical.”
Acharya: Airbnb's network effect remains defensible against AI vibe coding
“Something like an Airbnb, you know, you can have all the vibe coding in the world. Like, their network effect is incredibly powerful.”
Acharya: Commodity AI models with live proprietary data beat cutting-edge models
“Now you can put a relatively commodity model in front of it and get much better results than the most cutting edge model that does not have access to the proprietary or live data.”
Acharya: The current AI tech market is not in a bubble
“For the record, I don't believe we're in that period, but I do think that any time you have this sort of superheated markets, you have some distortion, okay?”
Acharya: AI startup gross margins are worse, but compute subsidies are healthier CAC
“So I do think that the blended margin story for AI native companies tends to be worse. But if you look at the overall sort of form of distortion that's happening, it's a much better one than we had five years ago.”
AI Power Users Pay 10x Higher Subscription Prices Than Traditional Software
“You look at Grok, heavy, it's 300 a month. ChatGPT, 200 a month. Gemini Ultra, two 50 a month. So we're seeing 10 X higher prices paid, and you have consumption revenue on top of it. So for the power users, they're paying incredibly high subscription rates plu…”
Acharya: Evaluate viral AI startups using Month-Two retention as baseline
“Retention really matters, and if you take a look at the best AI products, you know, even if you look at M-II as the new M-I, right, because again, you're getting a lot of sort of tourists who come in at M-I, you're not paying anything for them. Month one, that…”
Acharya: 50% Month-Two retention is solid for viral AI startups
“Like the bigger the better, but certainly 50% is solid, right? And if you're 60, 70%, I mean, we're very, very happy.”
Acharya: OpenAI reached $20B top line by tripling capacity and matching demand
“Look, this is not my area of focus or expertise, but one, you look at OpenAI's recent investment announcement, sorry, which is that there are twenty billion of top line. And the way that they got there is they three X capacity and they three X top line. So eve…”
Acharya: AI customer prices are rising, showing no supply overbuild
“Two, if you actually look at the prices that customers are paying, they're going up. So you're not seeing the sort of price compression that you would get from a typical overbuild of supply, right?”
Acharya: Enterprise spending is shifting from SaaS to human labor budgets
“I mean, we're already seeing it.”
Acharya: Voice agents are the primary entry point for AI in enterprise
“Voice is the wedge into the enterprise. Voice agents are so powerful.”
Acharya: Multi-functional AI bundling will drive 10x enterprise productivity gains
“So the most sophisticated companies are starting to take support, sales, collections, operations, bundle them all together with one broad goal, like CAC improvement. And I think that is going to be the 10 X on productivity more than saying, hey, we're just goi…”
Acharya: Massive AI categories will support dozens of specialized startup winners
“That is an industry, not a market, and you're gonna have dozens of winners that all specialize, just as in legal today, you've got dozens, dozens of specializations. So, I think in many of these markets we are talking about it as if it is one market when it is…”
Acharya: AI legal software value capture will approach $500B labor spend
“I don't think that we're in the eight to 12% anymore, right? Fifty billion legal software traditionally I think we're going to be somewhere between the 50 and the 500, and I think closer to the 500 than the 50.”
Acharya: AI productivity boosts lead to shorter work weeks over job cuts
“It's this thing of like pretty easy to get to 60, 70, 80%. So I do think that's why a 20% productivity increase so far we're seeing it show up more as, you know, a four day work week than 20% less jobs because jobs as bundles of tasks don't set themselves up t…”
Acharya: VCs Underestimate Market Sizes and Overestimate Zero-to-One Ease
“Here's what I think. I think we tend to consistently underestimate how big the markets are and consistently overestimate how easy it is to go from zero to one, right?”
Acharya: Underwrite Startup Investments on the Assumption of Founder Inertia
“I think when you have a formidable founder and they're showing a lot of early momentum in a market, inertia is the best mental model. In my mind, inertia is the most powerful force in the universe. So everything that is happening today is going to default happ…”
Acharya: Fast-growing startups are skipping $500 million growth funding rounds
“Yeah, that's right.”
Acharya: Traditional SaaS growth benchmarks remain viable despite AI speedup
“I don't think so. I mean, I think that a lot of it is dependent. It's calibrated to your part of the market. So product velocity plus business velocity. I do think that you have to be top quartile compared to your peer set, right? I think there are some market…”
Acharya: Software markets are shifting from execution tools to thinking tools like Figma
“That, by the way, is sort of ahead of where I think the market is going in terms of moving from products focused on execution, which today are being subsumed by coding agents, to markets focused on thinking, right? And I think a lot of the thinking work is goi…”
Acharya: VC culture lionizes hyper-growth over long-term defensibility
“So it has its own idiosyncrasies and difficulties, but I think often some of the most significant companies are these area under the curve companies and they're, look, they're these like 20 year overnight success stories. And those are, I think that those are …”
Stebbings: Series A is currently the hardest stage for venture investing
“And I say that Series A is the hardest place to be investing right now. Because essentially you have a million in revenue, very little signs of product market fit, honestly, at a million in revenue. You're paying a hundred to 200 XAR and it's incredibly compet…”
Acharya: Founders need irrational domain interest to endure market cycles
“You have to be a little bit irrationally optimistic to do it. I think you also have to be irrationally interested in the domain in which you're working, because these things get hot and cold all the time, you know?”
Acharya: Kriya founders pitched a16z wearing matching kimonos drinking Celsius
“I mean, the Korea guys to me are a great example. You know, they come into our first pitch, first meeting with everyone in the room, Thanksgiving holiday. And, you know, I think they're both wearing matching kimonos. They're both drinking Celsius's, you know, …”
Acharya: Willingness to be embarrassed is a competitive advantage in consumer tech
“I actually think conversely in consumer, having a beginner's mind and a high willingness to be embarrassed is a competitive advantage because so many consumer products feel embarrassing and, you know, they're immediately dismissed as embarrassing or impossible…”
Daily Product Usage Is Essential for VC Investors, But Most Don't
“Well, I think the number one thing, and here's my free advice to other investors, but also founders, is just like, you have to use the products today more than ever. You know, and I think the investing landscape of five or seven years ago when there was a ton …”
Acharya: Autonomous AI agent overhype ignores the need for human exception handling
“I think that the extremist view that we are going to have autonomous agents that simply do everything over incredibly long time horizons, like maybe we'll get there someday, but I do think that at a minimum, you need humans in the loop for exception handling.”
Acharya: Execution and expertise are no longer constraints in the AI era
“Because I don't think execution or expertise is any longer a constraint.”
Acharya: Offshore BPO call center jobs are prime targets for AI automation
“So if you look at BPOs, you know, business process outsourcing, they're the areas in which there's the least ambiguity, where the job is literally a series of tasks, where people in offshore call centers take a task off the queue. Those things are very well se…”
Acharya: Computer vision models have fallen behind hype and expectations
“I mean, I think RPA is super interesting, but vision models haven't nearly kept up with the sort of, you know, the way that we've talked about them.”
Acharya: Token cost for GPT-4 has dropped 100x since release
“The cost of actually a token on GPT-IV has, you know, gone down a hundred X since the model was released.”
Acharya: AI capability gains enable price increases that outpace model costs
“As the models get better, downstream players' ability to take those capabilities, productize them, and raise prices, Has outstripped the raising costs, right? So the incremental cost increase potentially, and in many cases not a cost increase, but it's, you kn…”
Acharya: High AI model costs enforce essential early business model hygiene
“The fact that these companies have costs actually forces a business model hygiene that I don't think existed across the board 10 years ago, and that's a good thing.”
Acharya: Software will eventually capture 80% to 90% of discretionary spend
“We're going to asymptote to 80 to 90%, I believe, for consumer spend and enterprise spend. The way we do that is by pushing the frontier, not by Cutting. The sort of discretionary spend. Of course, there's going to be, you know. Rent and food. But yes, dude, I…”
Acharya: YC community provides a distinct sales advantage to enterprise startups
“YC is an awesome place to start an enterprise startup that sells to other enterprise startups. And they've got these sort of like good vibes within the community that makes it easier to sell into even much bigger, more established YC companies.”
Acharya: VCs cannot perform kingmaking on inferior products
“So I think that the right investor can be a catalyst, but I don't think that you can take a product that would not otherwise be the winner and anoint them the winner.”
Acharya: Deel's Alex Bouaziz treats a16z partners like an extended sales team
“I mean, I basically have a sales quota with Alex, you know, and DG would say the same thing. And Ben, he's even calling Ben saying, Hey Ben, can you help make this introduction XYZ? Like, He knows how to get the best out of Andreessen Horowitz, and all of the,…”
Stebbings: Deel Co-Founder Alex Bouaziz Handled $1,000 Deal Lead Personally
“I pinged Alex on a Sunday morning seven a.m., Saying, hey, they want an intro to a sales rapper on your team. Who's the best person? He's like, intro to me, please. I'm like, dude, it's like a thousand dollar deal. He's not worth your time. He's like, no, no, …”
Acharya's Main 2021 Investment Mistake Was Assuming Product-Market Fit
“If there's a mistake that I've made, it's been being a bit too casual about product market fit, and this was more of a 2021 mistake, which is assuming something had product market fit, and perhaps it didn't, and perhaps the founder had a super credible theory,…”
Stebbings: Investors Know If an Investment Works Within Three Months
“I think, you know, a good investment or a bad investment in the first three months.”
a16z Spent Nine Months Trying to Reach Kriya's Founders
“We, they'd been so mysterious. We'd been unable to get ahold of them for nine months.”
Acharya: Most Business Books Are Full of Shit
“Hard Things was the first honest business book. Okay, and I always say about business books, like the business model of business books is selling business books. It's not making you better. Most business books are full of shit, and Hard Things is the first one…”
a16z Evaluates GPs via Founder 360s Every Two Years
“And he's like, well, in the near term, we don't measure you based on returns. We measure you by going and talking to every one of your founders every two years, doing a three 60 on you. And if your founders say you're telling them the truth, you're showing up,…”
Acharya: Seed stage is the hardest venture stage to systematize
“The seed stage is the hardest stage at which to invest, because it's easy to make one great seed investment, but it's hard to have a system, I think, for doing great seed investing, because there is, even when the people are amazing, There just isn't anything …”
Anish Acharya: Early AI market leaders from 2023-2024 have maintained their dominance
“In this product cycle, what's actually interesting is a bunch of the early leaders from 23 and 24 have maintained their lead. You know, we talked about Harvey. That's a really impressive company. You know, Gamma is a really impressive company. You know, Cursor…”
Shorts cut from this episode
▶ What makes Marc Andreessen Different? · 20VC with Harry Steb (@1:10:05)
▶ We don't believe in luck at a16z · 20VC with Harry Stebbings (@0:26)
▶ Browse-based vs chat-based UI · 20VC with Harry Stebbings (@20:53)
▶ We are not in an AI bubble · 20VC with Harry Stebbings (@28:25)
▶ Marc Andreessen's Best Advice · 20VC with Harry Stebbings (@37:49)
▶ "We are going to vibe code everything" is BS · 20VC with Har (@0:00)
▶ "The whole market is oversold software" · 20VC with Harry St (@0:00)