We’re returning to 20VC, featuring Harry, Rory from Scale, and SaaStr’s Jason Lemkin. Discussing GPT-5’s strategic deflation, the ruthless efficiency revolution, and why you may not need half of your company.
Summary
Summary:
- Jason’s Rule: “If AI hasn’t boosted your growth, you’re failing.” If your B2B company hasn’t reignited growth through AI by June 30, 2025, you risk irrelevance. Companies like Oracle and Intercom have succeeded. If you’re still “working on it,” you’re already lagging.
- Rory’s Reality: “Large sums don’t change character; they reveal it.” The AI investment wave ($300-$400B annually) might not deliver rational returns, leading to fewer, major winners reshaping the venture landscape.
- Harry’s Humility: “We overestimate our winner prediction ability.” None of his Fund I projected winners became fund returners, challenging reserve allocation models and provoking brutal honesty about early-stage investing.
GPT-5: The Significant Deflation
While GPT-5 was underwhelming, it may be exactly what the industry needed.
“The underwhelming reaction dampened the techno-optimist narratives of reaching AGI,” noted a host. “We’re focusing on improving business software now.”
The key isn’t the slight capability improvements but the business effects. GPT-5 is 8-10 times cheaper than premium competitors, pressuring rivals like Anthropic. For companies like Cursor, it’s transformational: “It’s the best development for Cursor. We have a competitive product at a quarter of the price.”
The Strategic Takeaway: Shifting focus from “AGI dreams” to business fundamentals benefits operators who need to build sustainable models over pursuing headlines.
The Chrome Acquisition: $34.5B for a Browser?
Perplexity’s $34.5B bid for Chrome may seem absurd until you grasp the strategy involved.
Chrome generates little revenue—Mozilla earns just ~$700M despite its market share. Chrome’s true value lies in distribution control. “Chrome’s true value, if not owned by Google, is greatest to Google, who pays $20B annually for search placement, similar to Apple’s Safari deal.”
For an AI firm like Perplexity, owning Chrome means instant user access. “An AI competitor to ChatGPT would highly desire Chrome’s user base.”
The Reality Check: It’s not about the tech; it’s about distribution in AI’s era. Control the browser, control the default AI experience.
The $3B N8N Deal: AI and Workflow Automation
N8N’s rise from $300M to $3B (reportedly 40M ARR, ending at 80M) illustrates AI’s transformative impact on categories.
“AI has enhanced the value of software by executing more work, beyond just automating workflows,” explained Rory. “AI isn’t just connecting apps now; it’s executing tasks, reshaping infrastructure.”
The Lesson: Identify companies benefiting from AI, focusing on founders who rapidly adapt and deploy LLM-enabled versions.
Datadog’s ~$200M OpenAI Solution
Datadog had its best quarter—$260M net new ARR—despite a 10% stock drop. The concentration risk from a ~$200M annual OpenAI contract partly explains it. AI clients are extensive, but few lead the market.
“With OpenAI growing, it could be $520M in a year unless renegotiated,” said Rory. The real insight: Companies leveraging AI infrastructure spending are capturing immense value.
“Almost all GDP growth is from AI capex. Companies co-attaching to AI infrastructure, like Nvidia, are benefiting significantly.”
The Framework: Identify components for building AI infrastructure—chips, data centers, power, monitoring, security—and invest in these foundational companies.
Palantir: A Rare Growth Phenomenon
Palantir grew from 12% growth at $2B to 45% at $4B ARR. Such enterprise software growth at scale is unprecedented.
The secret? Palantir is “the AI solution for large corporates.” Fortune 100 companies require strong vendors for $10M AI projects, not emerging startups. Palantir delivers