Google (GOOGL) is presented as a deeply undervalued AI powerhouse, with some analysts projecting a potential future valuation of $10 trillion. The company's custom TPU chips offer a significant cost and performance advantage over NVIDIA (NVDA), creating a massive new revenue stream as Google begins selling them to AI labs like Anthropic. In contrast, Apple (AAPL) is executing a capital-light strategy by focusing on running AI models locally on its devices, avoiding the expensive AI infrastructure race. This approach leverages partnerships, such as its deal to use a custom Google Gemini model, to create a uniquely private and personalized user experience. Investors should consider these two tech giants as distinct but powerful ways to gain exposure to the future of AI.

Google (GOOGL)

• The hosts present a strong bull case for Google, arguing it is one of the most undervalued companies in the world, despite its multi-trillion dollar valuation. One host suggests Google is an "AI giant currently masquerading as a simple, humble search engine" with a potential future valuation of $10 trillion.
• Dominant AI Models: Google's AI models are described as best-in-class, consistently outperforming competitors like ChatGPT on benchmarks.
  • Gemini: Their flagship Large Language Model (LLM).
  • VO3: A text-to-video model capable of creating "cinematic masterpieces."
  • NanoBanana: An advanced image generation and editing model.
  • Alpha Gemini: A series of science-focused AI models credited with contributing to two Nobel Prizes.
• Hardware Advantage (TPUs): Google has developed its own custom AI chips called Tensor Processing Units (TPUs), which are presented as the only real competitor to NVIDIA's GPUs.
  • TPUs are described as hyper-specialized for AI tasks, much like a "production line for sandwiches" is more efficient than a master chef who can cook anything (the GPU analogy).
  • They offer 2x the performance per watt compared to NVIDIA's A100 chips and are 1.5x to 2x more cost-effective per token.
  • While historically used in-house, Google is beginning to sell TPUs, citing a potential $50 billion deal with AI company Anthropic. This opens a massive new revenue stream.
• Unmatched Distribution: Google has a massive, built-in user base to deploy its AI products to, giving it a significant advantage over competitors like OpenAI who have to build an audience from scratch.
  • Key platforms include Gmail, YouTube, Android, and Google Search.
  • The new AI Overviews in Google Search are already monetizing at the same rate as traditional search links, proving they can integrate AI without hurting their core business.
• Strategic Investments: Google is a savvy investor, holding significant stakes in other major tech players.
  • Owns 14% of AI competitor Anthropic.
  • Owns 6% of SpaceX.
• Strong User Retention: Data shows Google's AI products are sticky. Gemini has a 90% three-month retention rate, which is considered staggering for a consumer product.

Takeaways

• The investment thesis for Google is that the market misunderstands it as just a search and advertising company, while it is actually a vertically integrated AI powerhouse with superior models, proprietary high-performance chips (TPUs), and the world's largest distribution network.
• Google's ability to design and use its own TPUs provides a massive cost and efficiency advantage, insulating it from NVIDIA's market dominance and supply chain constraints. The potential to sell these chips to others could unlock a new multi-billion dollar business line.
• Despite a rocky start with its public AI releases (like Bard), the return of co-founder Sergey Brin has helped refocus the company, leading to the development of market-leading models. This suggests a successful cultural and operational turnaround.
• Investors should look beyond the search engine and value Google based on its complete AI ecosystem—from the chips that power AI to the apps that deliver it to billions of users.

Apple (AAPL)

• The hosts address the major disconnect between negative public sentiment (Siri sucks, no AI investment, failed car project) and the stock trading at all-time highs. The bull case is presented as an "accidental" but brilliant strategy.
• The "Local Inference" Strategy: Apple's strength lies not in cloud-based AI, but in running AI models directly on its devices (iPhones, Macs).
  • Compute that happens on a local device is free, whereas cloud-based AI costs money for every query.
  • Apple's custom M-series and A-series chips are highly optimized for AI tasks, making them capable of running powerful models locally without needing a data center.
• Immunity from the AI "Bubble": By not investing trillions in building AI data centers, Apple is shielded from the immense capital expenditure (CapEx) and financial risk that companies like Google, Microsoft, and OpenAI are taking on.
  • This strategy allows Apple to "cherry-pick" the best AI models from partners and focus on what it does best: creating a seamless user experience.
• Strategic Partnership with Google: A new deal is highlighted where Google will create a custom Gemini AI model specifically for Apple's ecosystem.
  • This is compared to the existing search deal, where Google pays Apple $20 billion annually to be the default search engine on the iPhone.
  • This allows Apple to get a state-of-the-art AI model without the development cost and risk.
• The Ultimate Personalized Assistant: The true power of Apple's strategy is its access to user data in a private way.
  • An AI running on your iPhone has access to all your texts, emails, calendar, and context, which is far more personal data than a cloud service like ChatGPT has.
  • This enables the creation of a truly personal assistant that can manage your life, a promise that Siri failed to deliver on.
• Future Monetization Power: The hosts speculate that the power dynamic could flip. As AI "agents" become more common, companies like Google may need to pay Apple to get their agents access to Apple's highly engaged and valuable user base, similar to the current search deal.

Takeaways

• Apple's perceived failure to invest in AI is reframed as a strategic advantage. The company has avoided the costly "AI infrastructure wars" and is instead positioned to dominate on-device AI.
• The investment case is built on Apple's control over its hardware (chips), software, and, most importantly, the end-user. They can offer a uniquely personalized and private AI experience that competitors cannot replicate.
• The partnership with Google is a masterstroke, outsourcing the high-risk, high-cost model development while retaining control over the user experience and integration.
• For investors, the thesis is that the market is finally recognizing that Apple's path to AI dominance is not through building the biggest models, but by creating the most useful and integrated AI experience for its billions of users, leveraging its existing ecosystem.

NVIDIA (NVDA)

• NVIDIA is consistently mentioned as the dominant force in the AI chip (GPU) market and the primary benchmark against which Google's TPUs are measured.
• The hosts acknowledge NVIDIA's strengths, including its faster innovation pace and its "monopoly on the supply chain," particularly its access to cutting-edge manufacturing in Taiwan.
• However, the discussion posits that Google's TPUs are a legitimate competitor, especially for specific AI training tasks where they are more efficient and cost-effective.
• The sentiment is that while Google may not displace NVIDIA entirely, its entry into selling TPUs could "knock off a trillion or two off of NVIDIA's market cap" if they can scale production, representing a significant new competitive threat.

Takeaways

• While NVIDIA remains the undisputed leader in AI hardware, investors should be aware of emerging competition from vertically integrated giants like Google.
• Google's ability to produce and potentially sell its own high-performance TPUs could challenge NVIDIA's market share and pricing power in the long term, especially for large-scale AI companies looking for specialized, cost-effective solutions.

Other Investment Themes & Companies

• Anthropic: Mentioned as a major AI lab and a key competitor to Google and OpenAI. Their significance in this discussion is a $50 billion deal to purchase Google's TPUs, validating Google's hardware and marking the first major AI chip sale outside of NVIDIA's ecosystem.
• AI Infrastructure Race: A major theme is the massive capital expenditure (CapEx) required to build AI data centers. The hosts question if this is a bubble. Google is a participant in this race but has a cost advantage with its own TPUs. Apple is highlighted for strategically avoiding this race entirely.
• Microsoft (MSFT): Mentioned briefly alongside Google as a company with a major cloud platform (Azure) that is successfully signing large deals for AI services, indicating the broad demand for AI infrastructure.