In partnership with
Tuesday, April 21, 2026 | Private Markets Intelligence
Good morning. Apple just named its next CEO. Amazon bet $25B on the company that makes me (Claude). And the U.S. is spending billions to break China's chokehold on the magnets that run everything.
Three moves, one theme: whoever controls the infrastructure underneath wins.
Jose here, Founderscrowd
In today's rundown:
Apple names John Ternus CEO—hardware guy takes over $4T company
Amazon commits $25B to Anthropic in biggest AI infrastructure bet yet
USA Rare Earth spends $2.8B for Brazil mine to escape China's 90% controlroq buy vs. OpenAI's $20B Cerebras deal = the inference war is here
Trump signs psychedelics executive order
US-Iran peace talks resume in Pakistan
Rex Gelb spent a decade building HubSpot's paid engine. Now he's showing founders exactly how to do it.
On April 27th, get the framework to structure, launch, and scale paid media that drives pipeline, not just traffic. 20 minutes. Live Q&A. Free.
🍎 APPLE NAMES HARDWARE CHIEF AS NEXT CEO
Apple announced John Ternus will become CEO on September 1, 2026. Tim Cook becomes Executive Chairman after 15 years leading the company. Ternus runs hardware engineering—the division responsible for 80% of Apple's revenue.
The details:
Ternus age: 50 (younger than most senior leaders)
Apple tenure: 25 years at the company
Tim Cook's run: CEO since 2011 (15 years)
Market cap under Cook: $350B → $4T (1,000%+ growth)
Arthur Levinson: Current chairman → Lead independent director
Transition date: September 1, 2026
Why it matters:
Designed the original iPad
Ran Mac development during the M-series chip transition (Intel → Apple Silicon)
Took over all hardware engineering in 2021
Launched $599 MacBook Neo (mass-market play)
Controls development of Vision Pro, AR glasses, robotics division
The succession signal: Apple just did what every tech company wants to do—promoted from within, kept institutional knowledge, gave the successor years to prepare. Cook mentored Ternus for 3+ years. No drama, no outside hire, no board fight.
What this says about Apple's next phase:
Hardware still matters (80% of revenue isn't going away)
AI will be built into devices (not just cloud services)
Vision Pro + AR glasses = Ternus's bet on the next iPhone
Robotics division under Ternus = exploring new categories
Investor takeaway: When a $4T company picks a CEO, watch who they pick, not just that they picked someone. Hardware > services = Apple believes the next trillion comes from what you hold, not what you subscribe to. The MacBook Neo ($599), Vision Pro (spatial computing), and robotics bets are Ternus's to win or lose.
💰 AMAZON BETS $25B ON ANTHROPIC (THE COMPANY THAT MAKES CLAUDE)
Amazon announced it will invest $5B immediately + up to $20B more tied to commercial milestones into Anthropic (total: $25B). This follows Amazon's previous $8B investment. Anthropic commits to spend $100B+ over 10 years on AWS, securing 5 gigawatts of compute capacity using Amazon's Trainium chips.
The details:
New investment: $5B immediate + $20B milestone-based
Previous Amazon investment: $8B (2024-2025)
Total Amazon → Anthropic: Up to $33B
Anthropic → AWS commitment: $100B+ over 10 years
Compute secured: 5 gigawatts (Trainium chips)
Anthropic valuation: $380B (Feb 2026 Series G)
AWS customers using Claude: 100,000+
Chip generations included: Trainium2, Trainium3, Trainium4, future generations
Why it matters: This isn't just an investment—it's a lockup. Amazon is paying Anthropic $25B to run exclusively on Amazon's chips for 10 years. In exchange, Anthropic gets $100B in compute credits + Amazon becomes the exclusive cloud provider for Claude.
What Amazon gets:
Captive demand — Anthropic spends $100B on AWS (guaranteed revenue)
Chip validation — If Claude runs on Trainium, customers will too (Nvidia alternative proven)
Equity upside — Amazon owns a chunk of a $380B company
Strategic moat — OpenAI runs on Microsoft. Claude runs on Amazon. Google has Gemini. The AI wars are cloud wars.
What Anthropic gets:
$25B in cash — Funds model training without raising more venture rounds
$100B in compute — Enough capacity to train next 5+ generations of Claude
100,000 customers — AWS customers get Claude through Bedrock (Amazon's AI platform)
Infrastructure lock-in — Amazon builds Project Rainier (500,000 Trainium2 chips) specifically for Anthropic
The Trainium bet: Amazon's custom chips (Trainium for training, Inferentia for inference) are designed to undercut Nvidia on cost. If Anthropic proves they work at scale, AWS can sell compute cheaper than Microsoft/Google (who rely on Nvidia). That's the real play—not owning the model, but owning the cheapest way to run it.
Investor takeaway: Follow the infrastructure money. Microsoft spent $14B on OpenAI. Amazon just committed $25B+ on Anthropic. Google spent billions on Gemini. The AI boom isn't about who builds the best chatbot—it's about who controls the chips, power, and data centers underneath. Amazon is betting Trainium chips + Anthropic models = AWS growth for the next decade.
WANT IN ON DEALS LIKE THESE?
Most of these opportunities disappear before they hit the news.
Founderscrowd Premium members get first look at exclusive pre-IPO investments, the kind of deals institutional investors see before they go mainstream. We're talking early-stage access to companies that could be the next exits in satellite, EVs, fintech, and AI.
Premium includes:
✅ Weekly deal memos on live investment opportunities
✅ Private markets intel before it's public
✅ Direct access to carry on winning deals
✅ Community of 1,000+ active private markets investors
The difference? When OpenAI acquires a company, Premium members already knew the founders. When Amazon drops $11.57B, Premium members already owned the company. When Tesla hits 10M subscribers, Premium members will have already ridden the private equity wave.
Become a Premium member →
Lock in $40/month before it goes to $100+ (this week)
⚙️ USA RARE EARTH SPENDS $2.8B TO ESCAPE CHINA'S GRIP
USA Rare Earth agreed to buy Brazil's Serra Verde Group for $2.8B ($300M cash + $2.5B in stock). Serra Verde operates Latin America's only producing rare earth mine—and it's one of the few sources of heavy rare earths outside China. The deal closes Q3 2026.
The details:
Deal size: $2.8B ($300M cash + 126.8M shares)
Serra Verde production: 6,400 tonnes/year (stage 1)
China's market share: 90% of global processed rare earths
Serra Verde's share (by 2027): 50%+ of heavy rare earth production outside China
Mine life: 25 years
Expected EBITDA (2027): $550M-$650M annually
USA Rare Earth stock reaction: +14% (day 1), +60% YTD
US government backing: $1.6B (USA Rare Earth) + $565M (Serra Verde) in DFC financing
Offtake agreement: 15 years, 100% of production, US-backed
Why it matters: Rare earths aren't rare—they're just controlled by China. This deal gives the U.S. the only Western source of all four magnetic rare earths: neodymium, praseodymium, dysprosium, terbium. These go into permanent magnets for EVs, wind turbines, F-35s, and missile guidance systems.
What are rare earths? 17 elements on the periodic table with special magnetic properties. They're in everything: smartphones, EVs, wind turbines, military jets, precision-guided weapons. China mines 70% and processes 90%. When China threatened export bans in 2025, EV and defense stocks tanked.
Why heavy rare earths matter: "Light" rare earths (like neodymium) are easier to find. "Heavy" rare earths (like dysprosium and terbium) are critical for high-temperature magnets. Serra Verde has both. Most Western mines only have light rare earths, which means you still need China for processing.
The U.S. government's role:
$1.6B to USA Rare Earth (CHIPS Act funding)
$565M to Serra Verde (DFC financing)
15-year guaranteed purchase agreement (U.S. buys 100% of production)
Price floors included (protects against China flooding the market)
This is economic warfare. The U.S. is paying billions to build a supply chain China can't cut off.
What USA Rare Earth is building:
Mine (Serra Verde in Brazil) → Extracts ore
Processing (Oklahoma plant) → Refines into oxides
Magnets (Oklahoma plant opening 2026) → Manufactures finished magnets
Partnerships (UK's Less Common Metals, France's Carester) → Full supply chain
This is mine-to-magnet vertical integration. No Chinese middleman.
Investor takeaway: China controls 90% of rare earths because it's willing to operate at a loss to keep dominance. The U.S. government just said: we'll subsidize the alternative. $2.2B in federal backing + 15-year guaranteed purchase + price floors = USA Rare Earth doesn't have to compete on price. It just has to produce. If geopolitical tensions escalate, rare earth supply chains are the new oil pipelines—and the U.S. just secured one.
QUICK HITS
📰 Everything else moving markets:
Trump signs psychedelics executive order — Fast-tracks FDA review of psilocybin, MDMA, ibogaine for mental health treatment. $50M federal funding, National Priority Vouchers = weeks (not years) for approval. Stock reaction: AtaiBeckley +25%, Compass Pathways jumped. Potential path forward for cannabis rescheduling. Veterans with PTSD targeted first.
US-Iran peace talks resume in Pakistan — VP JD Vance + Jared Kushner + Steve Witkoff heading back to Islamabad for round 2. Ceasefire expires Tuesday. Trump threatened Iran: "lots of bombs will start going off" if no deal. First round was 21-hour marathon with no agreement. Iran skeptical this isn't a trap. Strait of Hormuz remains closed to Western shipping.
THE BIG PICTURE
What today's moves reveal:
1. Infrastructure > apps. Apple picked the hardware guy. Amazon bet $25B on chips + compute. USA Rare Earth bought a mine. Nobody's fighting over the best app—they're fighting over what powers it.
2. Vertical integration is back. Amazon owns the chips (Trainium) + the cloud (AWS) + a stake in the model (Anthropic). USA Rare Earth owns the mine + the processing + the magnet factory. Owning the whole stack = control.
3. Government is picking winners. USA Rare Earth got $2.2B in federal backing. Anthropic runs on DOD systems. When national security meets tech, subsidies follow.
WHAT TO WATCH
Apple's transition (Sept 1):
Does Ternus sell more hardware or just protect the iPhone business?
Vision Pro sales (flop or next iPhone?)
MacBook Neo ($599) success = Ternus's mass-market bet validated
Amazon/Anthropic results:
AWS revenue growth (is Anthropic actually driving cloud sales?)
Trainium adoption (do customers beyond Anthropic use Amazon's chips?)
Claude's market share vs. ChatGPT/Gemini
USA Rare Earth execution:
Mine production ramp (hits 6,400 tonnes/year or misses?)
Magnet plant opens on time (2026)?
Stock volatility (up 60% YTD—hype or fundamentals?)
That's your Tuesday case study.
Three companies betting billions that the next decade isn't won by better software, it's won by controlling the hardware underneath.
⭐⭐⭐⭐⭐ This is the signal
⭐⭐⭐ Average
⭐ Needs work
See you Thursday,
Jose ☕
Founderscrowd
P.S. Thursday we're breaking down why Elon's companies keep buying Cybertrucks from each other. Spoiler: it's not demand—it's desperation.