A Once in a 100 Year Investment Opportunity Just Started (What and Where I'm Buying)

Core thesis

Tom Nash argues that a sharp market selloff, Nvidia’s leadership change, and competition from in-house AI chips are creating an unusually attractive long-term buying opportunity. He believes AI demand and company backlogs remain strong, so investors should focus on business fundamentals rather than alarming headlines. He expects volatility and further “fear cycles,” but believes patient investors could benefit over a 10–15-year horizon.

These are the speaker’s opinions and reported figures, not independently verified here. Some numbers and names in the auto-generated subtitles appear garbled or internally inconsistent.

Market and AI-industry case

The speaker describes a major market drop around the U.S. midterms, along with anxiety over Nvidia’s CEO departure and the “chip war.”

He argues that Google and Amazon developing or selling their own AI chips does not necessarily undermine Nvidia:

• Cheaper computing could expand overall AI use and demand.
• Nvidia could have a smaller share of a much larger market and still grow revenue.
• Amazon reportedly plans to deploy more than 1 million Nvidia GPUs while developing its own Trainium chips.

He cites AI-related backlogs and commitments as evidence that demand is real:

• Google Cloud backlog: $560 billion.
• CoreWeave backlog: $132 billion, up 27% quarter over quarter.
• Anthropic commitments reportedly include $110 billion to Google and AWS and about $30 billion to Microsoft.

The speaker says CoreWeave revenue grew 170% year over year and that the company raised its full-year guidance.

He disputes headlines that Anthropic had lost $80 billion, saying about $72 billion was a non-cash accounting effect related to convertible notes. He says Anthropic revenue rose from $1.1 billion in the first half of the prior year to $16.3 billion in the first half of the current year.

Company views and figures cited

• Nvidia (NVDA): The speaker believes operations, customer commitments, product plans, and demand remain intact after CEO Jensen Huang’s departure. He views interim CEO Colette Kress’s immediate tasks as delivering the next earnings report, handling regulatory and export-control matters, and retaining senior employees. He emphasizes that a permanent CEO will need expertise in chip design, finance, supply chains, and dealmaking with suppliers such as TSMC and memory makers. He speculates that Huang could return, but says Nvidia can succeed without him.

• Google / Alphabet (GOOG/GOOGL): The speaker is bullish on Google’s cloud and TPU opportunity. He says Google Cloud revenue grew 103% to $30.7 billion in the cited quarter, with more than 100 significant TPU customers. He says the stock fell from about $400 in May to below $300 amid concerns about capex and negative free cash flow, which he views as investment in capacity rather than maintenance spending.

• Amazon (AMZN): According to the speaker, Amazon’s Trainium chips are expected to be offered outside AWS around late 2027. He cites Amazon’s plan to deploy more than 1 million Nvidia GPUs as evidence that custom chips and Nvidia products can coexist.

• Bloom Energy (BE): The speaker presents Bloom Energy’s on-site fuel cells as a solution to data center power constraints. He says Bloom supplied power to an Oracle Utah site in 55 days and prepared an initial power block for a New Jersey site in about two months, compared with years for a grid connection. Figures he cites include:

  • Quarterly revenue up 138% to $1.24 billion.
  • EPS of $0.94, compared with an expected $0.69.
  • Backlog of $17.8 billion, up from $9.4 billion a year earlier.
  • 2026 revenue guidance of $4.4–$4.6 billion.
  • Nearly $200 million of negative free cash flow in the quarter.
  • A roughly 40% decline from recent highs and a 132× trailing-earnings valuation.

His assessment is that Bloom is attractive but highly speculative, and that investors should monitor operating cash flow.

• Palantir (PLTR): The speaker cites quarterly revenue of $2.3 billion, up 96% from $1.2 billion, about $10 billion in cash, and no debt. He gives an operating margin of 47%, an adjusted operating margin of 62%, and net income of $1.13 billion on $2.31 billion in revenue. Using adjusted margin, he calculates a Rule of 40 score of 158; using reported operating margin, 138. He says U.S. commercial revenue is guided to grow at least 143% for the year and full-year guidance is about $8.5 billion. He also argues that government business offers resilience in a downturn.

• Tesla (TSLA): The speaker cites a 2.6% operating margin in the latest quarter and $25 billion of capex planned for the year. He views this as investment in autonomy, robotics, data centers, chips, AI, and energy—not just cars. He sees possible synergies or a future merger with SpaceX and attributes a potential $4 trillion combined valuation to an RBC report.

• SpaceX: The speaker describes SpaceX as three businesses: launch services, Starlink connectivity, and AI. He cites quarterly revenue of $14.6 billion, versus $5.6 billion a year earlier (161% growth), and net income of $4.4 billion, versus $1.1 billion. He gives segment figures of about $7.5 billion for AI, $5.25 billion for Starlink, and $1.8 billion for space.

Other figures he cites include:

- Starlink had **13.6 million customers**, up from **7.2 million** a year earlier, and service on roughly **4,000 aircraft**.
- More than **80%** of global mass launched to orbit since 2023, about **99%** Falcon mission success, and **35 launches** in the latest quarter.
- Starship flight 14 carried 26 Starlink V3 satellites and could add roughly **10×** the download capacity of a Falcon 9 Starlink launch.
- About **$20 billion** of capex in the recent quarter and a plan for roughly **$40 billion** in Q4, much of it for AI.

He says SpaceX went public on June 12 at $135, rose to about $225, and later traded around $120–$124. His plan is to accumulate slowly over 10–15 years, as he considers it too risky for investors with a five-year horizon.

Portfolio and trading framework

Allocation

• Keep roughly 40–50% in the S&P 500 as the portfolio’s core.
• Hold growth companies with established revenue and scale in a middle tier.
• Limit speculative holdings such as Bloom Energy and SpaceX to 5% or less of the portfolio each.

Dollar-cost averaging

The speaker advises against investing a lump sum. Instead, set a monthly amount and vary the pace according to price movements.

• With $1,000 available monthly, invest $500 at normal speed and keep $500 in dry powder, such as short-term Treasuries or a money-market fund.
• If a stock is 20% or more below its 52-week high, increase to “double-down” speed. The example is to deploy $1,500, using the month’s $1,000 plus $500 of dry powder. For the S&P 500, use a 10% decline threshold.
• If an investment rises 20% or more within 30 days, slow purchases to $250 and keep the remaining $750 in dry powder.

Trimming and process

As a position appreciates, the speaker recommends trimming:

• 10% after a 50% gain.
• 20% after a 100% gain.
• 30% after a 150% gain.

After each trim, reset the ladder and put proceeds back into the DCA pool.

The process is to dollar-cost average, check each holding against its 52-week high, choose the purchase pace, apply the trim ladder to winners, and keep the speculative sleeve within its limit. The speaker’s stated principle is to follow rules rather than predictions, panic, or headlines.

Explicit recommendations and cautions

The speaker favors gradual accumulation in the named companies, particularly when prices are well below 52-week highs, but emphasizes position sizing and patience. He characterizes Bloom Energy and SpaceX as speculative, volatile holdings, recommends slow DCA rather than lump-sum buying, and warns against exceeding the 5% portfolio cap for speculative positions. He also says AI investing will involve substantial volatility.

Disclosures and sources

• No explicit spoken “not financial advice” disclaimer appears in the subtitles. The speaker promotes the Tom Nash Academy/Patreon community near the end.
• Named companies and organizations include Nvidia, Google, Amazon, CoreWeave, Anthropic, Microsoft, Oracle, TSMC, Bloom Energy, Palantir, Tesla, SpaceX, Starlink, and RBC.
• The speaker also references Peter Lynch, Alan Iverson, Steve Jobs, Jeff Bezos, Sergey Brin, Bill Gates, Bill Ackman, and Elon Musk.
• Presenter: Tom Nash.