The new China chip deal removes major uncertainty for NVIDIA (NVDA) and AMD (AMD), making their revenue streams more predictable and strengthening the investment case. Microsoft (MSFT) is a core AI holding as it solidifies its dominance in the lucrative enterprise coding market through its successful GitHub Copilot platform. Conversely, investors should be cautious of companies at high risk of AI disruption, such as research firm Gartner (IT), which has seen its stock plummet on these concerns. Other potentially vulnerable stocks to scrutinize include Wix (WIX) and Shutterstock (SSTK). The most direct investment strategy remains focused on the "picks and shovels" of AI, including semiconductor, infrastructure, and power companies building the foundation for this technological shift.

NVIDIA (NVDA)

• An "unprecedented" deal has been struck where NVIDIA will pay the U.S. government 15% of its revenue from Chinese chip sales in exchange for export licenses.
  • This deal specifically involves the H20 chip, a less advanced semiconductor designed for the Chinese market.
  • Back-of-the-envelope math suggests this could generate around $3.5 billion annually for the U.S. government.
• Wall Street sentiment on this deal is surprisingly positive. An analyst from Futurum Equity noted that it replaces the "binary risk of a ban" with a "predictable, modelable cost," which investors should prefer. It makes the China AI market a "modelable growth stream."
• Risk Factor: In the wake of this deal, China is reportedly urging its domestic firms to avoid using NVIDIA's H20 chips, which could dampen the expected sales. The alternative for China would be to rely on domestic producers like Huawei.
• The company's focus extends beyond chips. It is actively developing "world models" for embodied AI, which involves AI for robotics and physical devices, indicating a broader long-term strategy.
• The podcast notes that people have been calling NVIDIA a "bubble for years," but so far, those betting against the stock have been wrong.

Takeaways

• The 15% fee for China sales is a major development. While it adds a direct cost, some analysts see it as a net positive because it removes the uncertainty of a total ban, allowing Wall Street to better forecast future revenue.
• Investors should monitor sales data from China closely to see if the government's discouragement of H20 chips materially impacts NVIDIA's revenue from the region.
• NVIDIA's work in robotics and other AI software applications is a crucial area to watch, as it represents significant growth potential beyond its core chip business and diversifies its role in the AI ecosystem.

AMD (AMD)

• AMD was mentioned alongside NVIDIA as having to accept the same deal with the U.S. government.
• The company has also agreed to give the U.S. government 15% of its revenue from chip sales to China to secure export licenses.

Takeaways

• Like NVIDIA, AMD's business in China now has a new, predictable cost structure. This reduces the risk of a sudden ban, which could be seen as a positive by investors who can now model this revenue stream more reliably.

Microsoft (MSFT)

• Microsoft is tightening its control over its subsidiary, GitHub, by folding it directly into its core AI engineering team.
• GitHub is no longer being operated as a semi-independent "loss leader" to drive sales to the Azure cloud platform.
• GitHub Copilot, its AI coding assistant, has become a significant revenue driver and is considered Microsoft's main play in the enterprise AI coding market, which the podcast calls the "most important vertical in enterprise AI."

Takeaways

• This strategic shift shows how critical Microsoft views the AI coding market. By integrating GitHub more closely, Microsoft can better leverage it to compete in the AI arms race.
• The success of GitHub Copilot positions Microsoft as a dominant force in enterprise AI adoption, making it a key stock for investors looking for exposure to the application layer of AI.

AI Disruption & At-Risk Companies

• A key investment theme discussed is identifying and betting against companies that are likely to be disrupted by AI.
• Bank of America created a list of 26 companies most at risk. The podcast specifically named:
  • Wix (WIX) - Web development firm
  • Shutterstock (SSTK) - Stock image company
  • Adobe (ADBE) - Software conglomerate
• This group of "at-risk" stocks has underperformed the S&P 500 by 22 percentage points since mid-May 2024.
• Gartner (IT), a research and advisory firm, was highlighted as a prime example of AI disruption.
  • After Gartner cut its revenue forecast, its stock fell 30% in five days.
  • The stock is down 50% so far this year, with Morgan Stanley analysts stating the results "added fuel to the AI disruption case."
• The podcast quotes an investment CEO saying, "Any company where you're paying someone to do something that AI can do faster and cheaper will be wiped out."

Takeaways

• Investors should consider the "second-order effects" of AI, which includes significant risks for established companies whose business models can be automated.
• The dramatic fall of Gartner's (IT) stock serves as a real-world example of how quickly the market can punish a company once the threat of AI disruption becomes apparent.
• When evaluating companies, especially in the tech and services sectors, a key question to ask is how vulnerable their core business is to being replaced or devalued by AI tools.

Investment Theme: AI Infrastructure & Coding

• A successful AI investment strategy highlighted in the podcast involves buying semiconductor, infrastructure, and power companies that are essential for the AI build-out.
• The scale of this build-out is massive, with spending on data center construction now nearly matching the spending on general office construction.
• The AI coding market is described as a central battleground. The intense competition between Anthropic's Claude, OpenAI's GPT-5, and Microsoft's GitHub Copilot is driving innovation and adoption.
• Demand for AI coding tools is so high that some platforms are reportedly losing money on each customer (negative gross margins) because users are consuming so many resources. This is seen as a sign of incredible product-market fit, not a lack of interest.

Takeaways

• The most straightforward way to invest in the AI theme is through the "picks and shovels"—the companies building the foundational infrastructure. This includes not just chipmakers but also data center operators and power companies.
• The AI coding sector is experiencing explosive growth. While the business models are still evolving, the immense demand signals a massive market opportunity for the companies that win this space.

Private Companies & The AI "Arms Race" (Anthropic & OpenAI)

• While these are private companies and not directly investable for the public, their dynamics offer crucial insights into the broader market.
• Anthropic has shown spectacular growth, reaching $5 billion in Annualized Recurring Revenue (ARR), primarily from its AI models being used for coding.
  • Risk Factor: This revenue is highly concentrated. Half of its API revenue ($1.4 billion) comes from just two customers: Cursor and GitHub Copilot.
• OpenAI is directly challenging Anthropic's dominance in the coding space with its new GPT-5 model, which was released with "extremely competitive pricing."
• The podcast suggests that if major customers like Cursor and GitHub Copilot switch from Anthropic to OpenAI, it could cause a "reversal in the market."

Takeaways

• The AI "arms race" is shifting. While model capability is still key, cost and efficiency are becoming major competitive factors, as shown by OpenAI's aggressive pricing.
• Anthropic's reliance on a small number of large customers highlights a key risk in the AI application layer. The platforms that are building tools on top of these foundational models can switch providers, creating volatility. This dynamic is important to understand even when investing in public companies that partner with or compete against these players.