Feb 10, 2025 · 55m · news
Sridhar Ramaswamy, CEO @Snowflake: Deepseek is Not a Threat to OpenAI & OpenAI Beats Anthropic|E1258 · 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, Snowflake CEO Sridhar Ramaswamy discusses high-intensity leadership, talent management strategies, and the competitive landscape of generative AI, drawing on his extensive experience at Google to analyze enterprise innovation, public market discipline, and the future of product-led AI.
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 19.4% of the talking time here. How this is scored →
Harry as informed peer 3.2 Guest teaching 4.2 Guest disagreement 2.2 Harry pushing back 2.4
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 22:03 Model vs product distinction pushback
Sridhar directly rejects Harry's assertion that users will switch to free models like DeepSeek, firmly correcting him that ChatGPT is a complete product rather than just an underlying model.
Hardest push from Harry ▶ 14:48 Challenging demotions vs terminationsHarry explicitly challenges Sridhar's management approach, asking why a demoted employee shouldn't simply be fired rather than given a smaller role.
Biggest teaching moment ▶ 48:09 Google's AOL distribution deal revealSridhar corrects the popular narrative of Google's purely product-driven rise by revealing they paid AOL over 100% revenue share to capture critical early distribution.
Harry holds his own ▶ 39:35 Framing the AI Capex bubbleHarry demonstrates deep market tracking by bringing specific mega-capex figures from Meta and Stargate to force a discussion on market rationalization.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Sridhar's Journey: From PhD to Software Engineer | 1 | 1 | 0 | 0 | Harry opens with a personal angle based on prep from Sridhar's colleague Margo, asking about early career aspirations and advice for young graduates. Sridhar reflects warmly on his transition from academic PhD to engineering leadership and offers advice on adaptability. | |
| How AI Will Impact Software Engineering and Knowledge Work | 2 | 4 | 1 | 1 | Harry asks whether software engineering will be the first profession thinned out by AI. Sridhar reframes AI as a translation layer across all knowledge work and draws historical parallels to how the internet concentrated journalism and music. | |
| Aging, Adaptability, and Sridhar's Limitations | 2 | 2 | 0 | 1 | Harry asks a personal question about what Sridhar has gotten worse at over time and how to bring teams along at high intensity. Sridhar discusses physical decline versus mental adaptability and framing high expectations as opportunity cost. | |
| The Challenge of Scaling Talent and Demoting People | 2 | 3 | 1 | 1 | Harry asks why leaders fail when scaling teams. Sridhar explains that skills enabling early-stage success become inhibitors at scale and discusses the necessity of uncomfortable feedback and demotions. | |
| Navigating Hard Conversations with Humility and Respect | 3 | 5 | 3 | 4 | Harry challenges Sridhar's stance on demotions, asking if keeping a demoted employee is a mistake compared to firing them. Sridhar counters by sharing a concrete Google Shopping case study where splitting a director's overwhelming portfolio led to massive success. | |
| Do Richer Leaders Make Better Leaders? | 4 | 6 | 6 | 5 | Harry pushes Sridhar on consumer loyalty to ChatGPT when free alternatives like DeepSeek emerge. Sridhar forcefully rejects the premise, emphasizing that ChatGPT is an integrated product rather than a commoditized model, telling Harry directly there is a big difference. | |
| Sam Altman's Strategy and the Threat of OpenAI to Startups | 5 | 5 | 3 | 4 | Harry presses Sridhar on competitive threats from Nvidia, cloud incumbents, and Databricks' lead in AI workloads. Sridhar delineates classic machine learning from modern AI and argues Snowflake's agentic framework positions it ahead. | |
| Public vs. Private: How Constraints and Visibility Drive Innovation | 3 | 4 | 3 | 2 | Harry asks whether public or private structure best aids innovation. Sridhar explains how public market discipline forces operational clarity, dismissing competitor AI messaging as three-card Monte. | |
| Enterprise AI Adoption, ROI, and Sridhar's Davos Takeaways | 3 | 5 | 5 | 3 | Harry brings up enterprise CEO complaints regarding a lack of AI ROI. Sridhar bluntly counters that skeptical CEOs simply do not understand the technology, citing specific enterprise workflows and Davos takeaways. | |
| Why Modern Tech Incumbents Are Innovating at Unprecedented Speeds | 4 | 5 | 1 | 2 | Harry notes that tech incumbents are innovating faster than in previous tech cycles. Sridhar provides a historical overview of past platform shifts, explaining how incumbents learned from DEC and SGI to aggressively self-disrupt. | |
| The AI Capex Arms Race: A Good or Bad Bubble? | 5 | 5 | 2 | 2 | Harry cites massive capex numbers from Meta and Stargate, asking where the AI arms race ends. Sridhar uses Nat Friedman's framework comparing telecom fiber overbuilds to short-lived dot-com burn. | |
| Snowflake's Growth Engine and M&A Strategy | 4 | 4 | 3 | 4 | Harry brings public market skepticism about Snowflake's growth potential and M&A strategy. Sridhar defends their organic product expansion model, contrasting product-led engineering with private equity rollups. | |
| AI and Data Applications: Snowflake's Future Revenue Drivers | 4 | 6 | 1 | 2 | Harry asks about future revenue drivers and model consolidation versus verticalization. Sridhar uses Google Search history to illustrate how distribution defaults and universal product expansion conquered niche verticals. | |
| Inside Google's Early Days: Distribution Secrets and Sridhar's Key Takeaways | 4 | 6 | 2 | 3 | Harry asks which early distribution deal mattered most for Google. Sridhar reveals that Google paid over 100% revenue share to AOL to secure user traffic, highlighting the critical importance of aggressive business deals alongside product quality. | |
| Quick Fire: Founder Mode, Parenting, and CEO Realities | 2 | 2 | 2 | 2 | In a rapid-fire sequence, Harry asks about founder mode, parenting, unmade decisions, and personal pride. Sridhar gives concise, grounded answers emphasizing personal presence over managerial trends. |
Statements from this episode (34)
Ramaswamy: Building Startups on Top of OpenAI Is Terrifying
“I think it is terrifying to be a startup building on top of OpenAI.”
Ramaswamy: ChatGPT Users Will Not Switch to DeepSeek
“You're not going to switch from ChatGPT to DeepSeq. It's a product. It's not a model.”
Ramaswamy: Anthropic Lags OpenAI Due to Model-Only Focus
“There's a reason why Anthropic hasn't done as well because it operates sort of at The model level. It's all of the additional things that go into ChatGPD that I think has staying power.”
Ramaswamy: Society Pretends to Value Teaching But Actually Does Not
“I was like, yeah, don't pick these noble professions that we all pretend we value, but in fact, we don't. I was like, being a teacher is really hard.”
Ramaswamy: Snowflake Prioritizes Drive and Malleability in Senior Hires
“I think that core formula of drive and malleability is the advice that I give to a twenty-one-year-old. Honestly, it's the thing that I look for when I am trying to hire you know, take your pick, a senior exec into Snowflake.”
Ramaswamy: AI is a translation layer between structured and unstructured knowledge
“You know, my sort of the simplistic, but fairly accurate description of what AI can do is like, it is this incredible Translation layer between all kinds of knowledge. Unstructured. Structured you can make sense of it all in a way that only humans were able to…”
Ramaswamy: AI Will Not Replace Software Engineering Anytime Soon
“I don't think software engineering is going to disappear anytime soon.”
Ramaswamy: Aging Limits the Ability to Master Brand-New Fields
“Another thing that we have to be realistic about as you grow older is your ability to just sink a huge amount of time into a brand new area is kind of limited. And you have to have the humility about stuff like that.”
Ramaswamy: Executive Strengths Become Inhibitors When Team Size Doubles
“Every time your team doubles in size all of the things that used to make you wonderful for your previous team Are typically massive inhibitors for you to succeed in your new job.”
Ramaswamy: I Have Demoted Executives by Halving Their Responsibilities
“I have demoted people, given them half their job and said, this is better for you.”
Ramaswamy: Postponing difficult workplace conflicts only makes them worse
“Conflicts typically do not resolve themselves. And so, you kind of have to internalize that postponing things is going to make them worse. The more you postpone, the worse it gets.”
Ramaswamy: Leadership Is Selling Employees on Unwanted Role Changes
“Again, that is real leadership. Going and selling something that people don't want, and painting a vision for the future and getting them to stay, that's hard.”
Ramaswamy: Personal Wealth Makes Leaders Callous and Overly Risk-Tolerant
“No. I mean, I think it can make them, ah, callous. It can make them a little too tolerant of massive amounts of risk.”
Ramaswamy: OpenAI will succeed through product experience, not model superiority
“OpenAI, I think, will be successful, not because they are always going to be the best and cheapest at creating foundation models, because they have a product experience that by some accounts has half a billion loyal users.”
Ramaswamy: OpenAI Reached 500M Users Without Paying for Advertising
“Which was the last company that you heard about that got to five hundred million you know, users without really, like, paying for advertising in recent years? Like, That's not a social network. Like none.”
Ramaswamy: OpenAI Strategically Misdirects Competitors on Hard Technical Problems
“It's also now becoming clear that they are good at misdirection in terms of where, you know, where they tell people hard problems are and it is good that people like DeepSeek, you know, went and busted some of those some of those myths.”
Ramaswamy: OpenAI and Anthropic Lead AI; Google Gemini Is Fast Follow
“It's OpenAI and Anthropic that have created the best models on the planet for the last three years. It's not Microsoft. It's not Amazon. It's not Google. It looks like, you know, Gemini is good, but let's face it, it's at best a fast follow.”
Ramaswamy: Snowflake is Ahead of Databricks in Generative AI
“They were first to market. That was their sweet spot, and they've been good at it for a for a while. We were very much in in catch up mode. But AI is a much younger field. Didn't really exist, let's call it pre-ChatGPT. There, I feel incredibly confident that …”
Ramaswamy: Databricks Buys Revenue and Has Double Snowflake's Sales Headcount
“It's very clear that they're buying a bunch of business. That they don't have to worry about things like you know free cash flow. They have doubled the number of sales people that we have.”
Ramaswamy: Databricks Shifts Narratives Between AI and SQL to Mislead
“Even if you look at, you know the competitor that we talked about for a few months, it was all about like amazing. AI is going through the roof. I haven't heard much from them about AI recently. All of a sudden, amazing. It's equal going through the roof. It f…”
Ramaswamy: Enterprise AI Adoption Growth Will Be Gentle, Not Exponential
“I think it is going to be gentler. In terms of just growth. On the other hand I can very confidently tell you that AI is creating value today, and it'll continue to create enduring value.”
Ramaswamy: CEOs Claiming No Value From AI Lack Understanding
“And if some CEO is telling you that they're seeing no value from AI, I'd be like, I'm not sure you know a whole lot.”
Ramaswamy: The AI Capex Arms Race Will End in a Bursting Bubble
“I mean, with a bubble bursting, as all bubbles do”
Ramaswamy: Legal AI Startup Harvey Will Not Be Disrupted by OpenAI
“I think Harvey is a great company. I don't think it's going to get disrupted by open AI.”
Ramaswamy: Snowflake Will Rely on Product Innovation Over Major M&A
“An inorganic kind of unrelated acquisition strategy is not something that I'm excited about. I think it'll be a disruption. Snowflake is a Product-led innovation company. We are not a PE shop. Not that there's anything wrong with being a PE shop. It's just a d…”
Ramaswamy: Snowflake's $150M Acquisition of Neeva Has Paid for Itself
“Snowflake spent some hundred and fifty million dollars for Neva. I think it's more than paid for itself.”
Ramaswamy: AI Will Disrupt Traditional BI Into Direct-to-Consumer Data Apps
“Starting with AI. 100%. I think it's going to be a I think it's going to be a big deal. I think it offers the opportunity for disrupting BI as we know it today, more of a direct to consumer.”
Ramaswamy: Microsoft's Live.com Had Better Image Search Than Google
“It will amuse you to know that Live.com, which is a predecessor to Bing, Had the better image search. They had the best image search. Google did not.”
Ramaswamy: ChatGPT Will Remain the Primary Consumer AI Entry Point
“On the consumer side, I know you disagree, I think ChatGPT is getting to be that entry point, and so if I were a betting person on the consumer side, To the extent that there are variations in models and specializations I think that'll include to the incumbent…”
Ramaswamy: Enterprise AI Will Feature Specialized Models, Not One Winner
“On the enterprise side, I absolutely see there being all kinds of specializations because so far in whatever the last 50 years, it's not the case that there's been an equivalent of Google search.”
Ramaswamy: Google Succeeded Early Because AOL and Yahoo Neglected Search
“Yahoo and AOL. Early on. Those were the entry points. AOL didn't know how to do a search. Yahoo didn't think search was important.”
Ramaswamy: Google Paid AOL Over 100% Revenue Share in Early Deal
“If I remember correctly, the, we paid AOL more money than we made. This was like a hundred plus percent ref share.”