The Twenty Minute VCEventbrite Sold for $500M, Databricks $5B Raise at $134B Valuation & Why SaaS is Like Japan
CHAPTERS
OpenAI–Thrive partnership in context: “Code red” and doubling down on the core
The panel opens on Thrive’s partnership with OpenAI, but quickly reframes it as yesterday’s story compared to OpenAI’s newer “code red” posture. They discuss what the partnership signals about OpenAI’s priorities and how much it’s really about strategy versus optics and access.
Power-law venture behavior: going deeper with the one or two true winners
Jason and Rory zoom out from the partnership to a venture reality: most outcomes come from a tiny number of breakout companies. They describe why firms concentrate time, relationships, and capital on the rare winner—and why that can feel “soul-crushing” to founders.
Databricks’ rumored $5B raise at $134B: pricing growth vs. multiple
The conversation shifts to Databricks’ rumored financing and whether 32x forward sales is justified. Rory frames a core valuation question—how much multiple to pay for incremental growth—using Snowflake as a comparable.
Re-acceleration at scale and the “infinite value” trap in models
They focus on what’s unusual: Databricks allegedly re-accelerating at very large scale. That breaks typical assumptions about deceleration and forces investors to rethink valuation frameworks—similar to how foundation-model companies get repriced during re-acceleration phases.
Snowflake vs. Databricks: oligopolies, agentic data access, and a new architecture
Rather than expecting one winner-take-all outcome, they predict an extended enterprise slugfest similar to past software oligopolies. The discussion then pivots to how agents accessing enterprise data may reshape architectures and the role of CRMs vs. data platforms.
Security backlash: platform lock-downs, vendor risk, and “revenge of the enterprise”
Jason raises a major counterweight to rapid agent adoption: security and data residency. Using examples of SaaS vendors being locked out of Salesforce ecosystems after breaches, they argue enterprises may become more restrictive—possibly favoring incumbents and bundled solutions.
Eventbrite’s $500M sale, PagerDuty at ~2x revenue: the harsh reality of low growth
They analyze the sobering valuations of slower-growing public SaaS companies and why being public can force decisions. Rory notes that acquirers may believe they can create growth where the public market no longer gives credit, but Jason questions how often “add AI and win” actually works.
The TAM Trap: why many SaaS companies can’t grow out of their markets ("SaaS is like Japan")
The discussion broadens into a diagnosis: many public SaaS firms are stuck in finite markets and struggle to find a second act. Rory argues it’s not CEO incompetence but saturation and crowded adjacent categories—leading to the “TAM Trap” and a need for tighter pricing discipline.
AI pricing and the seat-based model crisis: value, usage, and shrinking headcount
They explore why seat-based pricing is threatened as AI reduces headcount and boosts ARR per employee. The conversation compares seat-based models to usage/value-based pricing and highlights the difficulty of measuring value delivered—especially as competition erodes “labor-savings” pricing power.
Getting to $100M ARR faster: why hypergrowth and efficiency can coexist in AI apps
They distinguish between mature public SaaS optimizing for free cash flow and early AI startups where growth can overwhelm the need (and even the ability) to hire. Rory outlines three buckets—public incumbents, model builders, and AI app companies—each implying fewer people are needed per dollar of output.
Incumbents cloning fast: Google’s Lovable/Replit-like launch and defensibility in apps
They react to Google shipping a competitor to vibe-coding tools, noting how quickly incumbents now clone emerging categories. Rory argues model providers won’t (and can’t) copy everything, and OpenAI’s renewed focus suggests peak fear of “models will do every app” may be fading—except in obvious areas like coding.
AI in wealth management as a wedge into huge markets—without falling into another TAM Trap
A debate breaks out over investing in wealth/financial automation: Harry questions whether massive outcomes are realistic, while Rory and Jason argue AI can unify fragmented services (taxes, trusts, estate planning) and expand access down-market. They stress market segmentation, long adoption cycles, and the importance of proving traction before scaling spend.
Venture’s “relevance game,” compounding, and choosing where to compete
They close with a meta-discussion about modern venture incentives: fast follow-on rounds, media, and momentum vs. conviction-driven compounding. Rory argues the primary filter is whether a truly big company can be built, while acknowledging that winning the hottest deals requires paying up or losing to top-tier competitors.
Quickfire: Supabase ($5B) vs. Lovable ($6B) — infra defensibility vs. category capture
In a final quickfire, they debate whether they’d rather own the database layer (Supabase) or the front-end vibe-coding product (Lovable). Rory prefers the category winner at the surface layer if vibe coding is real, while Harry prefers the “harder problem” and stickier infrastructure bet amid rapid cloning risk.
