Most important take away

Snowflake CEO Sridhar Ramaswamy argues 2026 marks the end of the chatbot era and the beginning of the agentic era, with Snowflake repositioning from a data warehouse to an AI platform that lets non-technical business users query, analyze, and take action on governed data in natural language. The most actionable takeaway for professionals: adopting AI is no longer optional in software engineering and many knowledge-worker roles — Ramaswamy bluntly predicts software engineers who don’t use AI will soon be unemployed.

Summary

Actionable Insights

For Career & Skill Development:

• Adopt AI natively or risk obsolescence. Ramaswamy states directly: “if you run into software engineers that tell you that they’re deciding whether to use AI, I predict that that’s an unemployed software engineer not too far from now.” His own sons, software engineers in their mid-20s, use AI natively — they don’t have a choice.
• Avoid careers built purely on generating or reading documentation to write small code snippets — those tasks are “largely automatable at this point.” Snowflake itself cut its entire writing team in March 2026.
• Focus on durable skills: the “foundational aspects of software engineering” — system design, code base structure, long-term architecture — “still requires a lot of taste” and will remain in demand.
• Roles Snowflake is hiring into (signal for where to aim your career):
  • Sales and any role involving external communication
  • Software engineers, especially those “born into” the AI era
  • Developer relations (educating external developers on AI tooling)
• Learn to build “skills” — the emerging pattern is: do a task by hand, then ask a coding agent to package it as a reusable skill. Workers who can abstract their work into repeatable agent skills are moving “up the level of abstraction” and staying relevant. Even non-engineers at Snowflake (sales execs, HR) are now writing applications with Cortex Code.
• Learn to query AI agents effectively. The #1 reason agentic pilots fail is users not knowing how to ask questions and reliably get answers. Understanding the underlying data model is a prerequisite for trust.

For Enterprises / Operators:

• Start small with AI pilots. Snowflake pushes customers to “try small things and launch them before trying to do bigger things.” Pilots that make it to production take 4 weeks to 6 months depending on complexity.
• Run evaluation (“eval”) processes when deploying agents — regression-test every new feature like standard software engineering to avoid breaking working capabilities.
• Sequence autonomy carefully: start with low-risk, reversible, repetitive tasks, keep humans in the loop, then widen autonomy only after reliability is proven. Snowflake’s own SRE team used Cortex Code to cut time-to-problem-detection by “orders of magnitude” before moving toward autonomous handling.
• Democratize data access. The biggest unlock from agentic AI is letting CFOs, HR, and sales teams query governed data directly in natural language instead of waiting on analyst-built dashboards.

Stocks & Investments Mentioned

• Snowflake (SNOW) — The focus of the episode. Ramaswamy highlights concrete financial proof points to defend Snowflake’s competitive moat:

  • $9.8 billion in RPO (Remaining Performance Obligation) — durable future customer commitments.
  • A $400M+ single-customer contract recently signed, cited as evidence of long-term trust.
  • Fastest-growing products: Snowflake Intelligence (agentic natural-language analytics) and Cortex Code (agentic coding/automation for data workflows).
  • New launches: Project Snow Work and Cortex Code (March 2026).
  • Actionable angle: Ramaswamy is making a bull case that Snowflake is successfully pivoting from passive data warehouse to an active agentic AI platform with unstructured-data support and MCP connectors. Investors watching the data/AI space should track whether RPO growth and product adoption (Snowflake Intelligence, Cortex Code) continue to accelerate — those are the KPIs he is implicitly pointing to.

• Databricks (private) — Cited as the primary competitor, reportedly raising at a ~$134B valuation on ~$4B revenue in December. Ramaswamy pointedly contrasts Databricks’ “very fuzzy” reported numbers with Snowflake’s audited public-company RPO, implicitly arguing Snowflake’s revenue is more durable. Actionable insight: discount private-market valuations in the data/AI space against public comparables’ disclosed RPO/backlog.

• Hyperscalers and model makers as coopetition: AWS, Google (Azure is mentioned — note: Azure is Microsoft), Anthropic, and OpenAI. Snowflake has multi-billion-dollar spend commitments with Azure and AWS and “slightly more modest but still nothing to sneeze at” commitments with Anthropic and OpenAI. These are partners and competitors simultaneously. For investors, this signals (a) ongoing cloud capex tailwinds for AWS/Azure and (b) meaningful enterprise revenue flowing to Anthropic and OpenAI from data-platform partnerships.

• Customers name-checked as proof points: WHOOP, TSMC (“TSM”), and Evolve are cited as customers seeing outsized value from Cortex Code and Snowflake Intelligence.

• Niva — Ramaswamy’s prior ad-free, privacy-focused search startup, acquired by Snowflake in 2023 (historical context, not an investment opportunity).

• Episode sponsor mention: .tech domains (DotTechDomains) — sponsor content, not an investment recommendation.

Strategic Takeaway on Snowflake’s Positioning

Ramaswamy is trying to erase the line between analytic systems (looking at the past) and operational systems (taking action now). The bet: a future of governed, agentic, natural-language interfaces sitting on top of trusted enterprise data, with MCP connectors extending out to unstructured data and third-party systems. If that vision lands, Snowflake becomes infrastructure rather than a storage vendor — the strategic reason to hold or buy. The risk: hyperscalers and model vendors compete directly for the same “front door to the enterprise.”

Chapter Summaries

• Intro & Ramaswamy’s background: Academic turned Google Ads leader (15+ years, helped grow Search Ads to $100B+), founded Niva, became Snowflake CEO after its 2023 acquisition of Niva.
• The end of the chatbot era: 2026 is when agentic AI converges with coding agents to produce reliable enterprise applications, not just hit-or-miss chat.
• Cortex Code & Project Snow Work: Launched to automate the tedious, error-prone setup work inside Snowflake. Surprise finding: non-technical users (sales, HR, finance) adopted it fastest.
• Accountability and governance: Agents go through rigorous eval processes; governance, auditability, and regression testing are non-negotiable.
• Pilot-to-production reality: Most Snowflake Intelligence use cases make it to production in 4 weeks to 6 months. #1 failure mode: users don’t know how to query effectively.
• Employee adoption: In software engineering and technical sales, AI is already past the “is it useful” debate. Refusing to use AI is a career-ending choice in those roles.
• Layoffs and the writing team: Snowflake laid off its writing team in March 2026. Documentation has shifted to an agent-generated, skill-based model; some writers moved to roles building documentation tooling.
• Role shifts: Hiring is up in sales, AI-native engineering, and developer relations. Work is moving “up the level of abstraction” — people write skills that create agents that write code.
• Snowflake’s identity shift: From cloud data warehouse to agentic AI platform. Natural-language access democratizes data; MCP will blur analytic vs. operational systems.
• Competition with Databricks and hyperscalers: Ramaswamy cites $9.8B RPO and a $400M+ deal as proof of durable revenue, contrasts with Databricks’ “fuzzy” reported numbers, and frames hyperscalers and model makers as both competitors and multi-billion-dollar partners.
• Close: Find Ramaswamy on X and LinkedIn.