☕ Samsung hit a $1 trillion valuation. Stock up 15% in one day. 400% in one year.

Happy Sunday Crowd!

This week proved something we've been saying for months: the AI boom isn't about software anymore. It's about who controls the physical infrastructure underneath.

Samsung's stock jumped 15% in a single day — the biggest gain in the company's history — pushing its valuation past $1 trillion. Why? AI memory chip shortages. Apple reported its best March quarter ever ($111B revenue) and authorized a $100 billion stock buyback. AMD's earnings sent the stock up 19% as data center revenue hit $5.78B. A Bitcoin mining company signed a $9.8 billion, 15-year lease for AI data center capacity in Texas. And China's government is backing DeepSeek at a $50 billion valuation to build sovereign AI.

See the pattern? Chips. Power. Data centers. Memory. Infrastructure.

The companies making headlines this week aren't the ones building cool AI apps. They're the ones controlling the scarce resources AI needs to exist: semiconductors, electricity, cooling, and compute capacity.

Welcome to Sunday. Let's break down the week that was — and what it means for where the money's actually going. ☕

Jose & Alberto from Founderscrowd

This week's top 5 trending stories:

Founderscrowd: Samsung's stock surged 15% on Wednesday, May 6 — the biggest single-day gain in the company's history — pushing its market capitalization past $1 trillion for the first time. That makes Samsung only the second Asian company to hit this milestone, after TSMC.

The stock has now quadrupled over the past year. In one year. A $1 trillion company grew 400%. Let that sink in.

The numbers:

Why this happened:

AI is starving for memory chips. Every AI model — ChatGPT, Gemini, Claude, the systems running inside every data center — chews through enormous amounts of fast memory. Two types matter:

Samsung makes more of both than anyone else on Earth. And right now, there's a massive shortage.

The bottleneck: Building a new semiconductor fab takes 2-3 years. AI demand exploded faster than anyone could build capacity. Supply will stay tight through at least 2027-2028. That means higher prices and fatter margins for Samsung.

The competitive angle:

Samsung lost the early lead in HBM (high-bandwidth memory — the specialized chips Nvidia stacks into its AI accelerators) to SK Hynix. SK Hynix still controls ~55% of the HBM market vs Samsung's ~25%.

But the gap is closing. In February 2026, Samsung became the first company to mass-produce HBM4 — the sixth-generation standard expected to power Nvidia's next architecture. Customer feedback has been positive.

The ripple effect:

Samsung's rally pulled Korea's entire stock market to record highs. The KOSPI index broke 7,000 for the first time in history. Foreign investors poured $823 million into Korean stocks in a single session.

Why this matters:

When a $1 trillion company grows 400% in a year, it's not speculation. It's supply/demand economics. AI needs memory chips. Samsung makes them. There aren't enough. Prices go up. Margins expand. Stock follows.

For investors: The AI infrastructure thesis is playing out in real-time. Not in software (OpenAI, Anthropic). Not in flashy apps. In boring stuff like memory chips, power plants, and data centers.

Bottom line: Samsung's $1 trillion milestone proves the AI boom is real — but it's benefiting hardware manufacturers, not AI app developers. If you're investing in private markets, follow the constraints: chips, power, cooling. That's where the money is.

Founderscrowd: Apple reported its best March quarter ever on April 30: $111.2 billion in revenue (up 17% YoY), beating estimates by $1.5B. Then the company authorized a $100 billion stock buyback, one of the largest in corporate history.

The numbers:

What's driving this:

iPhone 17 is crushing it. $57 billion in iPhone revenue for a March quarter is unprecedented. The iPhone 17 lineup — particularly the Pro models — is seeing extraordinary demand. This is Apple's strongest iPhone cycle in years.

Services keep printing money. $31 billion in Services revenue (App Store, Apple Music, iCloud, advertising) is an all-time record. Services now generate more revenue than many Fortune 500 companies' entire businesses. And it's almost pure profit.

Why the $100B buyback matters:

Apple generates so much cash that it can't find enough high-return investments for all of it. So they're returning it to shareholders via buybacks. A $100 billion buyback reduces share count, which increases earnings per share, which supports the stock price.

This isn't financial engineering. This is "we print so much cash we literally can't spend it fast enough."

The AI angle:

Apple isn't building massive AI data centers like Microsoft, Google, or Meta. But it doesn't need to. Apple's strategy is on-device AI — running AI models directly on iPhones, iPads, and Macs using Apple Silicon.

This requires different infrastructure (chip design, not cloud compute), but the result is the same: Apple's ecosystem becomes stickier, Services revenue grows, and margins expand.

Why this matters:

While everyone else is spending hundreds of billions on AI data centers, Apple is printing record profits from hardware + services and returning $100B to shareholders. Different playbook. Same result: dominance.

For investors: Apple proves you don't need to be "all-in on AI infrastructure" to win. Consumer hardware + sticky services + capital discipline = cash machine.

Bottom line: Apple's $111B quarter and $100B buyback show that while AI infrastructure spending is exploding, the companies with the strongest consumer moats (iPhone, Services) are still printing money. Don't ignore consumer tech just because enterprise AI is hot.

Founderscrowd: Hut 8, a Bitcoin mining company, just signed a $9.8 billion, 15-year lease for 352 megawatts of AI data center capacity at its Beacon Point campus in Texas. This is a Bitcoin miner becoming an AI infrastructure landlord. And Wall Street noticed — the stock surged 25% before settling back.

The deal:

What makes this significant:

1. Bitcoin miners are pivoting to AI infrastructure.

Hut 8 started as a Bitcoin mining company. Mining Bitcoin requires massive amounts of power and cooling — the exact same infrastructure AI data centers need. So Hut 8 is repurposing its assets: land, power contracts, cooling systems, and grid connections.

2. Take-or-pay contracts = guaranteed revenue.

This isn't a "pay for what you use" deal. The tenant (a "high-investment-grade counterparty" Hut 8 won't name) is obligated to pay the full $9.8B over 15 years whether they use the capacity or not. No termination clause. Guaranteed cash flow.

3. Power infrastructure is the new moat.

Hut 8 CEO Asher Genoot said it bluntly: "We start with power." In the AI data center business, securing electricity from the grid is harder than building the facility. Hut 8 has power contracts. That's the bottleneck everyone else is fighting for.

The Beacon Point campus:

Beacon Point will eventually be a 1-gigawatt campus (1,000MW of utility capacity). Phase one is 352MW. Partners include American Electric Power (power), Vertiv (cooling), Jacobs (construction), and Nvidia (architecture — it's being built to Nvidia's DSX reference spec).

The competitive angle:

Hut 8 isn't the only Bitcoin miner pivoting. IREN just announced a $625M deal to acquire cloud player Mirantis. TeraWulf is doing similar AI infrastructure deals. The pattern is clear: Bitcoin miners have power contracts + cooling + land. AI companies need all three. Match made.

The financials:

Hut 8's Q1 revenue jumped to $71M (vs $21.8M last year), driven by $66M in compute operations. But they posted a bigger net loss ($253M vs $134M) due to unrealized losses on Bitcoin holdings.

The company has $1.3B in cash + Bitcoin. After Q1, they closed $3.25B in senior secured notes to fund the River Bend data center build.

The risk:

Building a 1-gigawatt campus by Q1 2027 is ambitious. Supply chain delays, cost overruns, permitting red tape, and grid connection issues could all blow up the timeline. But if they pull it off, Hut 8 becomes a major AI infrastructure player.

Why this matters:

Bitcoin miners pivoting to AI infrastructure is the ultimate "follow the constraints" story. AI companies need power, cooling, and land. Traditional data center operators are capacity-constrained. Bitcoin miners have exactly what AI needs — and they're willing to lease it for $10B+ contracts.

For investors: This validates the "infrastructure >> software" thesis. A Bitcoin miner signing a $9.8B lease for AI capacity proves the real money is in boring, physical infrastructure, not flashy AI apps.

Bottom line: When Bitcoin miners are signing $10B AI data center leases, you know the infrastructure bottleneck is real. Power and cooling are the new moats. Software is commoditizing. Hardware and real estate are printing money.

Friday's Premium deal memo:

A Series D enterprise software company doing $180M ARR, 85% gross margins, preparing for 2027 IPO. Already profitable (rare). Raising final private round at $2.8B pre-money. Minimum investment: $100K. Existing investors: Andreessen Horowitz, Sequoia, Salesforce Ventures.

Why it matters: This is one of the last "sleep well at night" enterprise SaaS deals before IPO. Profitable, growing 60% YoY, customers include 40% of Fortune 500. When enterprise software IPOs come back (they will), this exits at $5B+.

This week, Premium members also got:

Next week:

A biotech company with FDA breakthrough designation. Phase 2 trials showing 73% efficacy. Raising Series C before Phase 3. High-risk, high-reward. If Phase 3 works, this is a 10x. If it fails, it's zero.

$40/month. Lock in before price increases to $100/month in May.

Upgrade to Premium →

Founderscrowd: DeepSeek, the Chinese AI startup that shocked Silicon Valley last year with ultra-cheap AI models, is raising its first-ever funding round at a $50 billion valuation. China's national AI fund is leading. Tencent is participating. This is Beijing betting big on sovereign AI.

The raise:

Why this is a big deal:

1. DeepSeek was self-funded until now.

DeepSeek's founder, Liang Wenfeng, funded the company through his quant hedge fund (High-Flyer) and refused outside capital for years. This allowed DeepSeek to operate like a research lab, not a startup.

But the AI arms race forced a strategy shift. Competing with OpenAI, Anthropic, ByteDance, and Alibaba requires billions in compute, infrastructure, and talent. Self-funding doesn't scale.

2. China is treating AI as strategic infrastructure.

China's National AI Industry Investment Fund was created in January 2025 specifically to back domestic AI champions. Beijing sees AI the same way it sees semiconductors: too important to depend on the U.S.

DeepSeek leading the charge as China's flagship AI company.

3. DeepSeek proved cheap AI is possible.

DeepSeek broke into global consciousness in early 2025 by launching models that trained on a fraction of the compute and cost of OpenAI/Anthropic models — while staying competitive on benchmarks.

This terrified Silicon Valley because it suggested AI doesn't require infinite money. It requires smart engineering.

The competitive pressure:

DeepSeek is losing ground to rivals with deeper pockets:

DeepSeek needs capital to compete. Hence the pivot.

The talent war:

AI talent in China is expensive and scarce. DeepSeek has already lost key researchers to competitors. The funding round specifically allocates money for "employee benefits and talent retention."

Translation: Pay people more so they don't leave.

The geopolitical angle:

This raise comes as:

DeepSeek raising at $50B with government backing is China saying: "We're building our own AI stack, independent of the U.S."

The Huawei connection:

DeepSeek recently launched its V4 model optimized for Huawei Ascend chips instead of Nvidia GPUs. This is critical. If China can't access Nvidia chips (due to export controls), it needs domestic alternatives. Huawei is that alternative.

DeepSeek proving you can build competitive AI on non-Nvidia hardware is a strategic win for Beijing.

Why this matters:

DeepSeek's $50B valuation validates two things:

For investors: Sovereign AI is becoming a theme. Countries are realizing AI is too strategically important to outsource. Expect more government-backed AI champions in Europe, India, Japan, and the Middle East.

Bottom line: DeepSeek's $50B raise isn't just about one company. It's China declaring AI independence. When governments start backing AI companies at $50B valuations, AI has officially become infrastructure — not software.

Samsung's momentum:

AMD's scaling:

Hut 8's execution:

DeepSeek's funding:

Infrastructure broadly: