If I wanted to build $1M+ AI startup in 2025, I'd do this

Key Concepts

• Sustaining Innovation
• Moats (Data Advantage, Network Effect)
• Distribution as a Moat
• Vibe Coding
• AI Agents
• Data Advantage
• Network Effect Advantage
• High-Value Niche
• Distressed Assets
• Model Context Protocol (MCP)
• Secure MCP
• AI-Based Lending
• Format Businesses
• Personal Monopoly

Opportunities and Challenges in the AI Landscape

The conversation begins with a discussion about the current AI landscape and the challenges of building sustainable businesses in this rapidly evolving field. The speakers highlight that while it's easier than ever to start a business leveraging AI, it's also easier than ever to lose equity value due to intense competition.

• Sustaining Innovation: AI is often a sustaining innovation, meaning it drops costs but doesn't necessarily create a long-term competitive advantage. For example, using AI to automate outbound recruiting might provide a short-term margin boost, but competitors will quickly adopt the same technology, driving prices down.
• Moats Matter More Than Ever: Traditional moats, like being in a niche market too small for venture capital, are no longer sufficient. Real moats are now data advantages and network effects. An example of a real moat is software that allows funeral homes to track and trade inventory, creating a network effect that's difficult to replicate.
• Distribution is the New Moat: The ability to distribute a product or service is now a critical competitive advantage. Anyone can create something, so getting it in front of the right people is key.
• Vibe Coding: The speakers joke about "vibe coding," which refers to quickly creating functional but potentially insecure code using AI tools. They caution against using vibe-coded code in production due to security risks.

Tactical Approaches and Business Ideas

The discussion then shifts to more tactical approaches and specific business ideas that could leverage AI for success.

• Data Advantage or Network Effect: To build a sustainable AI business, you need either a data advantage or a network effect. This means focusing on acquiring unique data or building a community that creates a barrier to entry.
• Media Businesses as a Starting Point: Buying or incubating media businesses can be a good strategy. The idea is to first build a network effect by capturing attention, then acquire data, and finally build AI-powered tools to monetize that data.
  • Example: Buying a yachting website with a large community, then building data tools to track yacht prices and maintenance costs, and finally selling targeted services to the community.
• High-Value Niches: Focus on high-value niches rather than broad markets. Mr. Beast, for example, has a huge audience, but they are low-value because they are mostly children with limited spending power.
• Distressed Assets: Look for distressed assets with strong brands and valuable niches. TechCrunch, for example, has a great brand but may be struggling, presenting a potential acquisition opportunity.
• Event Series: Acquiring a sought-after event series like South by Southwest or TED could provide a platform for building a community and developing AI-powered tools.
• Cal.ai Example: Cal.ai is a scheduling app that has achieved significant ARR. The speakers suggest that Cal.ai should focus on building a network effect to avoid being outcompeted by similar apps. They suggest that Cal.ai could become the modern Weight Watchers by building a community around health and wellness.

Automations and Tools

The speakers discuss the automations and tools they are currently using.

• Gum Loop: A platform similar to Make or Lindy that allows for more intelligent automation than Zapier. It can be used to score sales leads and add context to CRM data.
• Manis AI: A tool for scoring candidates based on resumes and other data. It can be used to quickly identify the best candidates for a job.
• Vercel v0: A tool that can quickly generate website code based on a description.
• Lindy: A platform for building AI agents that can automate tasks like labeling emails and drafting responses.

MCP (Model Context Protocol) and Secure Data

The conversation delves into the concept of MCP and its potential for building secure and valuable AI applications.

• MCP Explained: MCP is a way of piping data or APIs into an LLM, allowing the LLM to query your data set.
• Personal Accounting Example: One speaker trained a Claude project on his personal accounting data from Xero, allowing him to ask questions about his finances and get tax optimization advice.
• MCP Server for Mac: The speakers propose building an MCP server for Mac that would make it easy to integrate APIs and manage data for use with LLMs.
• Security is Key: The competitive advantage for an MCP server would be security. The goal would be to become the trusted and secure platform for managing sensitive data.
• Secure MCP Branding: The speakers suggest branding the business as "scpmcp.com" to emphasize security.

Additional Business Ideas

The speakers brainstorm several additional business ideas that leverage AI.

• Automated Agency for Website Redesign: An automated agency that scrapes local businesses, identifies those with ugly websites, redesigns them using AI, and then automatically reaches out to the business owners to sell them the new website for a monthly fee.
• Message Maxing: An app that analyzes your text messages to identify toxic relationships, negative behaviors, and neglected friends.
• AI-Based Lending: An AI-powered lending platform that automates the diligence process for loans, making it faster and easier to get funding.
• Bank of Vibe Coding: An AI lending platform that provides funding to vibe coders in exchange for a percentage of their profits.

Conclusion

The conversation concludes with a discussion about the long-term prospects of AI businesses. While the speakers acknowledge that many AI businesses will likely fail due to intense competition, they also believe that there are opportunities to build valuable and sustainable businesses by focusing on data advantages, network effects, security, and distribution. They emphasize the importance of building a personal brand and audience to create a competitive advantage. The speakers end on a note of excitement about the potential of AI, while also acknowledging the challenges and risks involved.