Twitter.
Will the Skydance merger work?
It is a coin flip: I give the Skydance merger a 50% chance of working. Paramount bought Warner Bros. Discovery for $110 billion, the deal closed on 6 October 2026, and the new parent is called Skydance. It works only if it hits its $6 billion savings target under $79 billion of net debt. The numbers are enormous.
Questions about Twitter, answered
Will the Skydance merger work?
It is a coin flip: I give the Skydance merger a 50% chance of working. Paramount bought Warner Bros. Discovery for $110 billion, the deal closed on 6 October 2026, and the new parent is called Skydance. It works only if it hits its $6 billion savings target under $79 billion of net debt. The numbers are enormous. Paramount projects roughly $69 billion in combined 2026 revenue and $18 billion in earnings before interest, tax, depreciation and amortization (EBITDA), but that EBITDA already assumes the full $6 billion-plus savings target. Management says it can deliver more than $6 billion in annual savings within 3 years. Under its legal settlement, the combined company must also release at least 30 films a year for the first 2 years and 32 a year for the following 3. Then comes the bill. Roughly $79 billion of net debt. Junk credit ratings from S&P and Fitch. Expensive borrowing. Old cable and broadcast TV in decline. And a $6 billion savings target that Warner Bros. Discovery itself previously called aggressive and highly speculative. And the big money keeps coming in. Elon Musk put about $1 billion into Skydance in October 2026, Semafor reported. That makes him a shareholder in the parent company of CNN, CBS and HBO. The friendship runs both ways: Larry Ellison, Oracle co-founder and father of Skydance chairman David Ellison, put $1 billion into Musk's Twitter takeover in 2022. Neither Musk nor Skydance has confirmed the deal publicly. My forecast: layoffs, 100% probability. Full $6 billion savings target achieved, 50% probability. Those percentages are my forecast, not company guidance or Wall Street data. The debt decides, not the content.
Full answer →Is X Money good or dangerous?
Both. X Money can bank the unbanked, or score them. The builder decides which. The good: X Money launched in April 2026 with money-transmitter licenses in 41 US states, a Visa partnership, a 6% yield and 600 million users already on the platform. I forecast it processes over $1 trillion a year by 2030. By 2040 it is infrastructure for a generation that never opened a bank account. The danger: within 36 months comes AI credit scoring from your social graph via Grok. Your follower count becomes collateral. That is 1 step from a social credit score. New York regulators call Musk unfit, and banks, regulators and governments can always change the rules. For now, Elon Musk deserves the benefit of the doubt. He bought Twitter for $44 billion in 2022 to save freedom of speech, and he takes on governments that attack it. The trend I predict. The mission I choose: technology must empower society, never control it. That test applies to Musk exactly as hard as it applies to Brussels and Beijing.
Full answer →Why raw materials instead of crypto?
Because you cannot short a copper mine to zero with a tweet. Crypto was built on hype, leverage, and a few loud men. Raw materials are built on physics. Demand doubles by 2030, and the world cannot print copper. Raw materials, natural resources, commodities, minerals, critical minerals and rare earths: see my top 10 raw materials, my OIL 2075 prediction card and the top 10 richest countries in natural resources.
Full answer →Why does memory matter more than logic chips for AI?
Because every AI system is memory-bound: the chips everyone tweets about wait on the memory feeding them. Korea took the AI hardware crown quietly while the analysts were writing Nvidia fan fiction, and in June 2026 Samsung and SK Hynix pledged $520 billion for 4 new chip plants. The market that books out a year ahead tells you where the bottleneck, and the pricing power, sits.
Full answer →How big is the HBM market?
$35 billion in 2025, racing to $100 billion by 2028, an industry forecast, not a fact yet. Want to know who wins the AI hardware war? Ignore the logic chips everyone tweets about. Watch the memory.
Full answer →