The multi-decade AI infrastructure build-out remains a top investment, as the industry's shift to visual models will require exponentially more computing power. With a five-year outlook, NVIDIA (NVDA) is viewed as undervalued, with its potential revenue from this build-out likely to far exceed current analyst estimates. While remaining cautious on the MAG-7 as a group, consider Google (GOOGL) a potential winner due to its leadership in AI models and drug discovery. The AI drug discovery theme itself presents a massive opportunity, with companies like Eli Lilly (LLY) leading the charge in this space. Finally, consider holding Bitcoin (BTC) as a long-term store of value and a hedge against currency debasement, made more accessible through new ETFs like IBIT.

Artificial Intelligence (AI) as an Investment Theme

• The speaker, Jordi Visser, argues that the demand for AI is infinite and will never end. He compares it to fundamental technologies like electricity and railroads, suggesting we are in the very early stages of a multi-decade trend.
• The current AI build-out is being funded by the MAG-7 companies, which have enormous free cash flow, unlike the dot-com bubble which was funded by low-margin telecom companies taking on debt.
• The build-out is transitioning from Large Language Models (LLMs) which process text, to Visual Language Models (VLMs) which process visual data. This next phase requires 50 to 1000 times more computing power, suggesting the demand for infrastructure is far from over.
• The U.S. government's "Genesis mission" is a key development, merging public and private sectors to ensure American competitiveness in AI against China. This makes the AI build-out "untouchable" from a policy perspective, as it's framed as critical for national survival.
• The biggest long-term risk associated with AI is a deflationary spiral. As AI makes everything more efficient and potentially displaces jobs, it could put significant downward pressure on prices and wages for a large portion of the economy.

Takeaways

• Investors should view AI not as a short-term bubble, but as a foundational, long-term technological shift. The comparison to the dot-com bubble is flawed because the underlying economics and funding sources are vastly different.
• The "picks and shovels" play (investing in the infrastructure builders) is still in its early innings, especially as the industry moves towards more compute-intensive visual models (VLMs).
• Government support solidifies the AI trend, reducing regulatory risk and ensuring continued investment in the sector for geopolitical reasons.
• While the growth is immense, investors should be mindful of the potential for AI to cause significant economic disruption and deflation, which could create losers in other sectors of the economy.

NVIDIA (NVDA)

• The speaker believes NVIDIA is "dirt cheap" on a five-year outlook, using it as a primary gauge for the health of the AI trend.
• He estimates a total global spend of $5 trillion on the data center build-out. Compute infrastructure will make up 30-60% of that cost.
• Taking the conservative estimate of 30%, that's a $1.5 trillion market for compute. NVIDIA currently has a 90% market share in data centers.
• Even if NVIDIA's market share falls to 50% by 2030 (with competitors like Google's TPUs taking the rest), its revenue could reach $750 billion. This is significantly higher than current analyst estimates of $425 billion for 2030.
• The company's moat is its proven track record. In the race to build AI capabilities, companies are unlikely to risk using unproven chips from competitors.

Takeaways

• The speaker's analysis suggests that current market expectations for NVIDIA may still be too low, despite its massive run-up.
• Investors with a long-term horizon (5+ years) might consider the stock undervalued based on the sheer size of the total addressable market for AI compute.
• The key risk to monitor is competition, particularly from other hyperscalers like Google developing their own chips (TPUs). However, the speaker believes NVIDIA's established dominance provides a strong, defensible moat for the foreseeable future.

MAG-7 Companies (e.g., Google, Microsoft)

• The speaker is "marginally bearish" on the MAG-7 as a group over the next few years.
• While they are the primary funders of the AI build-out, he questions whether all of them will be able to generate the revenue needed to justify their massive spending.
• He suggests that over time, some of these companies may start to look and feel more like utility companies—stable, but with lower growth and profit margins as they compete in the AI infrastructure space.
• Google (GOOGL) is highlighted as a potential winner, "breaking out" due to its AI labs (Isomorphic Labs, DeepMind) and its Gemini 3 model. It is also positioned as a key competitor to NVIDIA with its TPU chips.
• Oracle (ORCL) is mentioned as an example of the immense demand; its stock was "getting hit" because it couldn't meet the overwhelming revenue demand for its cloud services, underscoring the capacity constraints in the industry.

Takeaways

• An investment in the MAG-7 is not a monolith. Investors need to be selective and identify which companies are best positioned to monetize their AI investments, rather than just spending on them.
• Look for companies that are not just providing infrastructure but are also developing unique, high-margin applications on top of it, such as Google's work in drug discovery.
• The thesis suggests that the massive capital expenditures of these companies could compress their profit margins in the coming years, potentially making them less attractive than pure-play AI leaders like NVIDIA.

AI Drug Discovery & Longevity

• This is presented as the sector where the "bubble gets the biggest, but it's also where the revenues are infinite."
• The speaker believes AI will effectively "solve aging as a problem in the next five years" and will allow humans to "live technically forever."
• He specifically mentions Eli Lilly (LLY), now a trillion-dollar company, as a prime example of a firm breaking out in this space.
• Google's Isomorphic Labs, run by Nobel Prize winner Demis Hassabis, is another key player mentioned, with the stated goal of curing all diseases.
• The economic impact is massive: Curing disease and extending lifespans would dramatically reduce entitlement spending (e.g., Medicare), which is a primary driver of government debt.

Takeaways

• The intersection of AI and healthcare, particularly in drug discovery and longevity, represents a massive, long-term growth opportunity.
• Investors should pay attention to pharmaceutical and biotech companies that are heavily integrating AI into their research and development processes.
• Companies like Eli Lilly (LLY) and the efforts within Google (GOOGL) are at the forefront of this trend. This theme is in its very early stages and could be a major driver of market performance over the next decade.

Bitcoin (BTC)

• The speaker is very bullish on Bitcoin, viewing it as one of only three things in the world with a true, lasting "moat", alongside gold and religion.
• It is an "accepted store of value" and the primary representation of the new digital economy.
• The investment case is tied to two major trends:
  1. Government Debasement: As governments print money to support failing "zombie" companies and fund social programs, the value of fiat currency erodes. Bitcoin, with its fixed supply, is a hedge against this.
  2. AI Disruption: As AI destroys the competitive moats of traditional businesses (like what happened to Michael Saylor's original business), the only safe havens are assets with un-destroyable moats. Bitcoin is presented as the ultimate digital moat.
• The recent launch of Bitcoin ETFs (like IBIT) represents an "IPO moment" for the asset, broadening its ownership from a small group of early adopters to the wider public and institutional investors.
• Demographics are a key driver. Younger generations who grew up in a digital world understand and prefer Bitcoin over gold. As wealth transfers to them, demand for Bitcoin is expected to rise.

Takeaways

• Bitcoin is positioned as a long-term strategic holding, not a short-term trade. Its value proposition is as a store of value in an increasingly digital and inflationary (in terms of money supply) world.
• The approval of spot Bitcoin ETFs has fundamentally changed the accessibility of the asset for retail and institutional investors, potentially leading to sustained inflows.
• The speaker's framework suggests that as AI disrupts traditional industries, capital will flow towards the few assets with indestructible moats, with Bitcoin being the primary digital beneficiary.