Financial Breakfast — October 21, 2026

Good morning, everyone. Now that the euphoria from last night's earnings has settled, today's conversation revolves around two things: how much money the AI machine will still burn and how much it will still generate. Let's get to what matters.

1. SpaceX blows out the quarter and flirts with Tesla

The fact: SpaceX delivered earnings per share of US$ 0.35 against an expectation of US$ 0.13, with revenue of US$ 14.58 billion (up 161% year over year) and an AI business that has already overtaken Starlink as its largest source of cash.

My verdict: I Would Buy — with reservations. The thesis is solid, but the mention of a merger with Tesla (TSLA) adds an element of unpredictability that calls for caution on position size.

• SpaceX (pre-IPO, exposure via secondary funds)
• Tesla (TSLA) — merger noise is moving the stock
• Satellite and connectivity sector (Starlink as anchor)
• AI and data center supply chain

2. Anthropic and Nscale: another US$ 45 billion in compute

The fact: Less than a month after announcing a US$ 45 billion contract, Anthropic and Nscale are negotiating a second agreement of the same size, signaling that the race for computational capacity is far from cooling off.

My verdict: I Would Buy — demand for AI infrastructure remains the scarcest asset in the market. Nobody wants to be left off the track.

• Nvidia (NVDA) — backer of Nscale, captures the cycle from the ground up
• Data center and cloud sector
• Energy and utilities companies (growing electricity demand)
• Anthropic (indirect exposure via Amazon and Alphabet)

3. Amazon becomes Morgan Stanley's top pick with a US$ 350 target

The fact: Morgan Stanley upgraded Amazon (AMZN) to its top choice, with a price target of US$ 350, and the stock closed up nearly 5% at US$ 262.45, with focus on AWS acceleration.

My verdict: I Would Buy — AWS is the half of the AI thesis that actually generates cash today, and the stock still trades at a relevant discount to the target.

• Amazon (AMZN) — protagonist of the thesis
• Microsoft (MSFT) and Alphabet (GOOGL) — sector multiple read
• Cloud and digital advertising sector
• E-commerce small caps

4. South Korea bets US$ 520 billion on an AI chip hub

The fact: Seoul is planning a semiconductor complex in Gwangju with four AI memory plants from Samsung and SK hynix, with mass production expected by June 2030.

My verdict: I Would Buy — this is a long-term structural move. Memory is the silent bottleneck of AI, and whoever controls supply sets the price.

• Samsung Electronics (005930.KS)
• SK hynix (000660.KS)
• Micron (MU) — DRAM price read
• Asian semiconductor sector

5. BlackRock wants a bitcoin ETF with downside protection

The fact: BlackRock filed with the SEC for an ETF that seeks exposure to the upside of IBIT (or bitcoin), but with a mechanism to cushion declines — a hybrid product that broadens the institutional lineup.

My verdict: I Would Hold — the product itself is interesting and signals maturity, but bitcoin (BTC) remains volatile. Those who already have a position keep it; those who don't should wait for an entry price that makes sense.

• Bitcoin (BTC)
• IBIT and crypto ETFs in the U.S.
• BlackRock (BLK) — fee income
• Asset management sector

6. Strive buys US$ 335 million in bitcoin and prepares a new issuance

The fact: Strive acquired US$ 335 million in bitcoin after investors contributed US$ 369 million in warrants, and now plans to sell US$ 630 million in preferred shares at 15%.

My verdict: I Would Reduce — the "bitcoin vehicle company" model with 15% preferred issuance embeds leverage and expensive cost. Direct BTC exposure via ETF is cleaner.

• Strive (ASST)
• Bitcoin (BTC) — supply pressure from corporate treasuries
• MicroStrategy (MSTR) — model correlation
• Crypto treasury sector

Immediate opportunities

• NVDA and AWS/AMZN: demand for AI compute remains insatiable — Anthropic's two deals are proof.
• Asian memory semiconductors (Samsung, SK hynix): the Gwangju project is a decade-long structural bet.
• SpaceX via secondary funds: robust numbers and an already dominant AI business deserve attention, with a controlled position.

Risks on the radar

• Tesla-SpaceX merger: governance and valuation noise could distort both stocks.
• Leverage in bitcoin treasuries: 15% preferreds and aggressive issuances are a time bomb if BTC corrects.
• AI concentration: any capex slowdown drags down the entire chain — memory, data centers, and energy.
• Geopolitical risk in Asia: the Gwangju hub depends on regional stability and execution by 2030.

Sources: