The AI application layer is a key investment theme, with a focus on companies that have a defensible moat through proprietary models and data. OpenEvidence, a private AI company for doctors valued at $6 billion, is a prime example due to its proprietary technology and adoption by 40% of US physicians. Its specialized model, which achieved a 100% score on medical board exams, creates a significant competitive advantage. For investors seeking high-growth FinTech, private company Revolut is a notable opportunity, recently valued at $75 billion in a tender offer. However, be aware that Revolut's success is tied to a strong macroeconomic environment and its ability to grow in the competitive US market.

Revolut

• Revolut is a private FinTech company, primarily operating as an online bank in Europe with a growing presence in the United States.
• The company recently conducted a tender offer at a $75 billion valuation.
• Strong Growth Metrics:
  • Customer base grew to 35 million active users, a 35% year-over-year increase.
  • Revenue in 2024 was $4 billion, representing a 72% year-over-year increase.
  • Revenue is forecasted to be $5.9 billion in 2025 and $9.3 billion in 2026.
• Valuation: The $75 billion valuation represents a 12.7x multiple on the company's 2025 forecasted revenue of $5.9 billion. The speakers noted this multiple is "totally not outrageous."
• Staying Private: The tender offer is part of a strategy to stay private for longer, providing liquidity to employees and early investors. This is a growing trend among large private tech companies like SpaceX, OpenAI, and Stripe.
• Sentiment: The sentiment is mixed but leans positive on the business performance.
  • Bullish: The company is executing a successful global playbook, similar to a successful neobank in Brazil. Its growth in customers and revenue is described as "enormous."
  • Bearish / Risks: FinTech valuations are highly cyclical and dependent on the macroeconomic environment. A downturn could negatively impact the company's valuation. The business model has low margins ("pennies on the dollar") and requires massive scale. The company is reportedly struggling to penetrate the competitive US market.

Takeaways

• Investing in Revolut is a bet on a high-growth, global FinTech company that is successfully scaling its operations, but it comes with risks tied to the overall economy.
• The trend of large private companies using regular tender offers to provide liquidity instead of going public is seen as a "super healthy" development. For investors, this provides a way to get a regular, institutionally-backed valuation mark on their private holdings.
• The key risk for Revolut is its dependence on a strong economy and its challenges in the US market. Investors should monitor macroeconomic trends as a key indicator for FinTech performance.

AI Applications (Investment Theme)

• The podcast hosts believe the AI application layer is moving past the "too early" stage and is now showing "glimpses of clarity." They feel this is where "investors get made."
• Two primary strategies for building AI apps were discussed:
  1. Building on Top of Existing Models: Using foundational Large Language Models (LLMs) from companies like OpenAI, Anthropic, and Meta. Sierra AI is an example of this.
  2. Building Proprietary Models: Creating your own specialized, in-house LLMs for a specific purpose. OpenEvidence is an example of this.
• The discussion favored the second approach, highlighting that a proprietary model and data set create a stronger, more defensible moat.

Takeaways

• The AI application layer is becoming a major investment theme. Investors should start identifying promising companies in this space.
• When evaluating an AI app, a key question is whether it has a defensible moat. The speakers lean towards companies with proprietary models and/or proprietary data sets as being more defensible long-term.
• This is because companies simply building on top of major LLMs risk being disrupted when the platform provider (e.g., OpenAI) decides to release a similar feature.

Sierra AI

• Sierra AI is a private AI company that creates AI agents for customer service.
• The company is valued at $10 billion and is backed by top-tier venture capital firms including Greenoaks, Iconic, Sequoia, and Benchmark.
• Leadership: The CEO is Brett Taylor, who is also the Chairman of OpenAI. This connection is viewed as a significant strategic advantage.
• Technology: Sierra builds its solutions on top of existing LLMs from OpenAI, Anthropic, and Meta.
• Business Model: The company uses a "consultative approach" rather than a simple software subscription. They work closely with large enterprise clients to build custom solutions, which creates deep integration into the client's workflow. This deep integration is considered a potential moat.

Takeaways

• Sierra AI represents an investment in the "AI agent" space, betting that its consultative, enterprise-focused model creates a durable business.
• The investment thesis hinges on the belief that this deep client integration and workflow knowledge is a strong enough moat to defend against competitors, even without a proprietary foundational model.
• The leadership connection to OpenAI is a major bullish factor, suggesting a privileged relationship with a key technology provider that could mitigate platform risk.

OpenEvidence

• OpenEvidence is a private AI company providing a medical-focused Large Language Model (LLM) solution for doctors.
• The company is valued at $6 billion and is backed by Kleiner Perkins and Google.
• Market Dominance: The platform is already used by 40% of US doctors, which is considered a massive moat due to the trust and network effects it creates.
• Technology: Unlike Sierra, OpenEvidence has built its own proprietary LLMs. It uses a more capital-efficient "deep seek" approach, creating smaller, highly specialized models trained for specific medical use cases.
• Performance: The model's accuracy is extremely high, reportedly achieving a 100% score on medical board exams.
• The Moat: The company's defensibility comes from several factors:
  • Its proprietary, specialized model.
  • Its massive user base (40% of US doctors).
  • It solves a major pain point for doctors: reducing administrative work like paperwork and note-taking. This is compared to the healthcare software giant Epic, which dominates its market through workflow integration, not necessarily superior technology.

Takeaways

• OpenEvidence is presented as a prime example of a successful "vertical AI" company with a strong, multi-faceted moat.
• Investors looking at AI applications should seek out companies like OpenEvidence that target a specific industry, solve a clear and high-value problem, and have a defensible advantage through proprietary technology and deep market penetration.
• The company's success demonstrates that building smaller, specialized AI models can be a more efficient and effective strategy than competing to build massive, general-purpose models. The high adoption rate and accuracy validate this approach.