Investors Look for Comfortable AI Valuations

Key Concepts:

• AI as a profound technology shift
• Valuation of AI-related companies (Palantir, Taiwan Semiconductor, Nvidia, Intel)
• AI's impact on various sectors (tech, financial services, consumer, healthcare, industrial)
• Infrastructure constraints for AI deployment (power, cooling)
• AI pilot project failure rates
• Dot-com era comparison
• Semiconductor supply and demand
• Intel's challenges and competition with Taiwan Semiconductor

1. AI as a Transformative Technology

• The speaker emphasizes that AI represents the most significant technological shift of our time.
• He draws parallels to the cloud, social, and mobile revolutions of the past 25 years, suggesting AI's potential for even greater value creation over the next 20 years.
• The fund's launch is based on the long-term positive trajectory of AI, irrespective of short-term market fluctuations.

2. Valuation Considerations in the AI Market

• The speaker acknowledges that some AI-related companies are trading at speculative and highly valued multiples.
• Example: Palantir is cited as a company with strong fundamentals but a very elevated valuation.
• He contrasts this with companies like Taiwan Semiconductor (TSMC), trading at a teens earnings multiple, and Nvidia, which, despite its growth, still trades at a reasonable multiple.
• The key takeaway is that valuation depends on the specific company and its earnings growth rate relative to its stock price.

3. Broad Impact of AI Across Industries

• The fund's strategy is to invest in companies across various sectors that will be enabled, enhanced, or benefit from AI.
• Examples: The fund includes holdings in financial services, consumer, healthcare, and industrial sectors, in addition to technology.
• The inclusion of Eaton and Blackstone suggests investments in power and real estate, recognizing the infrastructure needs of AI.

4. Infrastructure Challenges for AI Deployment

• Power is identified as the single biggest constraint to deploying data centers at the scale and density required for AI.
• Cooling data centers is another critical aspect, creating an entire ecosystem of specialized companies.
• The fund aims to participate in this infrastructure build-out over the long term.

5. AI Pilot Project Failure Rates and Lessons from the Dot-Com Era

• The speaker addresses concerns about the high failure rate (95%) of AI pilot projects, referencing an MIT research report.
• He draws a parallel to the early days of the commercial Internet, where many failures occurred before the emergence of successful companies.
• The speaker suggests that failures are a natural part of the learning process, leading to improved technology and better deployment strategies.
• Quote: "You're generally going to, you know, fail a little bit out of the gate and you learn from that."

6. Intel's Valuation and Semiconductor Market Dynamics

• Intel's high valuation is attributed to collapsed earnings, not necessarily market enthusiasm.
• The speaker finds SoftBank's investment in Intel curious, given their existing investments in ARM, Stargate, and Ampere.
• This investment is viewed as a potential hedge against semiconductor undersupply.
• The US government's involvement is questioned due to Taiwan Semiconductor's established leadership and superior process technology.
• Quote: "Intel's not going to get fixed overnight and there's nothing the government can do to change the the disparity between Intel's process technology and Taiwan Semiconductor."

7. Semiconductor Supply Constraints

• Semiconductor content and power are identified as significant constraints in the AI landscape.
• SoftBank's strategy of selling traditional holdings (e.g., T-Mobile) and redeploying capital into AI infrastructure is noted.

8. Conclusion

The speaker presents a bullish long-term outlook on AI, despite acknowledging short-term market volatility and challenges. He emphasizes the importance of considering valuations carefully, investing across various sectors, and addressing infrastructure constraints. The comparison to the dot-com era suggests that failures are a necessary part of innovation, and the focus on semiconductor supply highlights the critical role of hardware in the AI ecosystem. The speaker is cautiously optimistic, recognizing that Intel's challenges will not be resolved quickly and that Taiwan Semiconductor remains a dominant player.