Investors should prioritize the "Application Layer" of AI by targeting companies like Eleven Labs and Leyora, which are transitioning from simple tools to autonomous agents capable of replacing high-cost human labor. Monitor legacy incumbents Relx PLC (RELX) and Thomson Reuters (TRI) for potential downside risk as AI-native startups disrupt their traditional "billable hour" and data-access business models. Oracle (ORCL) remains a high-conviction infrastructure play as the primary provider for these hyper-growth AI companies to train and deploy their models at scale. Look for opportunities in Disney (DIS) as they lead the "Voice IP" economy by licensing legacy character assets for AI-generated content and digital monetization. For private market exposure, focus on "lean" hyper-growth firms that automate internal operations to achieve massive revenue-per-employee ratios, similar to Eleven Labs' $600M ARR milestone.

This podcast episode features a deep dive into two high-growth AI companies: Eleven Labs (Voice AI) and Leyora (Legal AI). The discussion focuses on how AI is transitioning from a "productivity tool" to an "autonomous agent" capable of replacing high-cost human labor in trillion-dollar industries.

Eleven Labs (Private)

• Eleven Labs is a research-led AI company specializing in human-like text-to-speech, speech-to-text, and voice orchestration.
• Revenue Growth: The company has seen an extraordinary ramp, reaching $600 million in ARR (Annual Recurring Revenue) in roughly 40 months.
  • It took 20 months to hit $100M, 10 months to hit $200M, and only 5 months to hit $300M.
• Business Model: Shifting from consumer "fun" tools to enterprise-grade "AI Agents" for customer support, sales (SDRs), and operations.
• Key Partnerships: Working with major brands like Disney (Darth Vader/James Earl Jones estate), Epic Games (Fortnite), Masterclass, and Headspace.
• Market Expansion: They have paid out over $22 million to voice actors who license their voices on the platform, creating a new "Voice IP" economy.

Takeaways

• The End of "Voice Jail": The "press 0 for operator" era is ending. AI agents now handle complex, real-time interactions with high fidelity, allowing users to interrupt and speak naturally.
• Voice as Liquid IP: Celebrities and voice actors can now "digitize" their identity to earn passive income across multiple languages simultaneously.
• Internal Efficiency: The company operates with 600 employees and zero Product Managers (PMs). They embed engineers in every department (Legal, HR, Talent) to automate workflows, suggesting a new blueprint for "lean" hyper-growth startups.

Leyora (Private)

• Leyora is an AI-native platform disrupting the legal industry, specifically targeting the "billable hour" model.
• Growth Metrics: Sustaining 50% quarter-over-quarter growth for the last seven quarters. It is one of the fastest enterprise software companies to scale from $1M to $150M in revenue.
• Market Opportunity: The legal services market is $1 trillion, but software spend is only $40 billion (4%). Leyora aims to capture the 96% currently spent on human labor.
• M&A Strategy: The company is aggressively acquiring legacy legal tech businesses (4 acquisitions this year) to integrate their data and customer bases.

Takeaways

• Disruption of Junior Lawyers: AI is now capable of performing "Level 1" associate work—document review, due diligence, and contract analysis. This forces law firms to move toward fixed-fee or success-fee models rather than hourly billing.
• The "Data Moat" Myth: While legacy players like LexisNexis and Westlaw have massive historical databases, Leyora argues that being "AI-native" and moving faster is more important than owning the archives.
• In-House Power Shift: Large enterprises are using tools like Leyora to bring legal work in-house, significantly reducing their reliance on expensive external firms like Kirkland & Ellis.

Investment Themes & Sectors

1. The "Agentic" Era of Software

• Insight: We are moving past "Copilots" (where AI helps a human) to "Agents" (where AI does the work and the human manages the output).
• Actionable Trend: Look for companies that are "vertically integrated"—meaning they don't just provide a chat box, but handle the entire workflow (e.g., Leyora for Law, Eleven Labs for Voice).

2. Legacy "Tax" and Infrastructure

• Mentioned: Airwallex (Private)
• Context: Mentioned as a "first principles" financial platform. The theme is that companies built on "legacy rails" (old banking systems, old legal databases) are being replaced by "AI-native" infrastructure that doesn't charge a "legacy tax."

3. Frontier Model Competition

• Sentiment: Neutral/Competitive.
• Context: Both CEOs use "Frontier Models" (like OpenAI’s GPT or Anthropic’s Claude) but view them as "commoditized" engines.
• Insight: The real value is moving to the Application Layer. Owning the specific customer data (legal cases or labeled audio) is seen as a stronger long-term moat than the AI model itself.

4. Public Companies Mentioned (Indirectly)

• Relx PLC (RELX) & Thomson Reuters (TRI): Owners of LexisNexis and Westlaw. The transcript suggests these incumbents are "getting crushed" or priced with "AI uncertainty" because they cannot pivot fast enough.
• Oracle (ORCL): Mentioned as the infrastructure provider (OCI) for major AI companies to train and deploy at scale.
• Disney (DIS): Highlighted for its forward-thinking approach to licensing "Legacy IP" (voice) for future AI-generated content.

Risk Factors

• Data Leakage: For legal and financial sectors, the risk of proprietary data "seeping" into general AI models is a major barrier to adoption.
• Regulation: Mentioned specifically regarding European/French laws on privacy and "Right to Publicity" (voice cloning).
• Impersonation/Fraud: Eleven Labs noted the need for "AI Watermarking" to prevent scams and unauthorized use of celebrity voices.