$TSLA and $SPCX are in talks with $TSM regarding US-based production capacity. The likely model involves a facility where $TSM retains the technology and holds a majority stake, rather than a scenario where $TSM simply transfers the manufacturing know-how for $TSLA to handle production independently. Actual production is expected to commence around 2030 or later. In short: $TSLA and $SPCX will still rely on $TSM to manufacture their chips, rather than establishing a competing fabrication plant.

แผนการลงทุนปี 2026-2030 ของผม เป็นยังไง มาอัปเดทกันหน่อย -> 2026 - ปีของคนที่มีของ ไม่ใช่คนที่กำลังสร้างโรงงาน -> 2027 - ปีที่ของตึงที่สุด และ Optics เริ่มมา -> 2028 - ตลาดเริ่มหายใจไม่ออก แต่ AI ยังตาย -> 2029 - ตลาดจะให้ premium กับของที่ยังขาดจริงเท่านั้น -> 2030 - AI จะกลายเป็นเรื่องปกติของคน ไม่ได้น่าตื่นเต้นแบบตอนนี้แล้ว ผมเขียนบทความไว้แล้ว อ่านกันได้เต็มๆ ในคอมเม้นเลยครับ

Power + cooling are starting to merge into one stack. Midea just introduced a "power-cooling hyperconverged architecture" combining power systems, energy storage, magnetic-bearing cooling, and liquid-cooling CDUs from day one. As rack density climbs, the bottleneck may shift from: "Do we have enough cooling?" → "Can power + cooling scale together?" Investment implication: Don't just screen for CDU vendors. Look for suppliers sitting at the intersection of both systems: → Power distribution / conversion → Pumps & thermal controls → Heat exchangers → Power + cooling management software → Integrated modular infrastructure Basically, the winner may not be whoever builds the best CDU. It may be whoever owns the most pieces of the power-to-heat stack.

800VDC is starting to look less like a PowerPoint slide and more like an actual supply chain. $NVDA says 80+ companies are now building around the architecture. And we're starting to see the actual hardware take shape: - Power racks → 2H26 - Row power centers + 800VDC busways → 2027 - DC power blocks → next This is where it gets interesting. Everyone already knows the obvious 800VDC names. I'm more interested in finding the "FORM of 800VDC." The boring little components that suddenly become mission-critical when you try to push MW-scale DC power through an AI factory: → DC breakers / protection → busways & busbars → rectifiers → capacitors → high-voltage connectors → DC/DC conversion Basically, I'm looking for the company selling the pickaxe inside the pickaxe. Because once an architecture becomes standardized, the fun part isn't asking whether it'll happen anymore. It's figuring out which component becomes the bottleneck when everyone tries to build it at the same time. That's usually where the weird little stock hiding three layers upstream gets interesting.

Bottleneck Move!! “Time-to-Power” → “Time-to-First-Token” Past: AI ↑ → Data Center ↑ → Power ↑ Now-Future Land → Grid/Substation → Transformer → Switchgear → Microgrid → Rack Power → Cooling → Commissioning → First Token

"Anthropic building chips" And people panic sell $NVDA? Wrong read. Anthropic is exploring its own AI chips Sounds like competition. It’s not. What people miss: - Designing chips ≠ scaling them - AI demand is exploding so fast… - No single supplier can keep up. Custom chips don’t replace $NVDA They fill the gap – $GOOGL, $AMZN, $MSFT already doing this – Now Anthropic joins → demand is real And guess what: - $TSM still builds it - $AVGO / $MRVL still power the stack The play: – $NVDA → still king – $TSM → bottleneck – $AVGO, $MRVL → leverage to custom chips This isn’t $NVDA losing. It’s AI demand going vertical.

Your $NVDA isn’t the bottleneck anymore. And that should scare you. AI is about to hit a wall. Not compute. Bandwidth. Data centers are quietly breaking → copper can’t handle the load The shift has already started: electrical → optical (photonics) Here’s the part no one is talking about: You can have the best GPUs in the world… But if data can’t move fast enough? They sit idle. Burning cash. Doing nothing. Everyone crowded into: $NVDA, $SMCI But the next leg of the trade is forming underneath them: - Optical networking - High-speed interconnects - Photonics This is where AI either scales… or stalls. New AI stack (that most people are late to): - Brain: $NVDA - Body: $SMCI - Nervous system: Photonics Miss the third layer → you miss the next move. The money flow is shifting to: - $AVGO, $MRVL backbone players - $LITE, $COHR, $AAOI pure optical - $CRDO, $ALAB hidden winners - $TSEM, $TSM, $AEHR picks & shovels What people still don’t get: - AI isn’t just a chip story anymore. - It’s a data movement + energy problem. - And both are becoming bottlenecks… at the same time. Translation: - The biggest winners of the next cycle aren’t the ones you already own. - They’re the ones fixing what’s about to break.

How many times does retail need to miss the obvious? AWS CEO drops the bomb: Every single latest Anthropic model was trained on Trainium not $NVDA Trainium 2 sold out. Trainium 3 nearly gone. Anthropic is scaling hard on Amazon’s own silicon. Retail still treating $AMZN like a boring cloud stock. This is exactly how you build a real alternative to Nvidia’s monopoly, one proven frontier model at a time. Custom chip narrative just got a massive upgrade.

$AEHR just hit a new all-time high. Retail kept shorting because they only looked at this quarter’s revenue. Meanwhile, hyperscalers are quietly qualifying for volume ramp $37M+ bookings last quarter with 3.5x book-to-bill says it all. Qualification cycles aren’t instant revenue. They’re the calm before the explosion. Still early.

Day 17/100 of One Person Business Attention is the new currency. “Distribution is everything.” You can build something amazing… But if no one sees it, it doesn’t matter. The best product doesn’t win. The best distributed one does. Learn to write. Learn to tell stories. Learn to capture attention. Because once you have attention, you can sell anything.

Day 16/100 of One Person Business If you’re still getting paid by the hour, you’re capping your own income. “Don’t sell time. Sell outcomes.” Time is limited. Outcomes scale. Clients don’t care how long you work. They care about what they get. More leads. More revenue. Less pain. That’s what they pay for. Shift your thinking: From “How many hours?” → “What result do you want?”

If you spend every single dollar you earn... You're still broke. Doesn't matter if you make $5k, $10k, or $50k a month. High income with $0 left at the end of the month = still poor. Poverty isn't about how much you make. It's about how much you keep.

Day 15/100 of One Person Business People don’t pay for your time. They pay for acceleration. No one cares how hard you worked. They care how fast they win. Effort is invisible. Speed is obvious. 10 hours of work? That’s your problem. 10 minutes to the result? That’s what they pay for. Speed is leverage. Outcomes are currency. Don’t sell effort. Sell time saved.

Day 14/100 of One Person Business You don’t get rich by working more. You get rich by working differently. “Wealth = specific knowledge + leverage.” Anyone can work hard. Not everyone builds leverage. Specific knowledge = what only you can do Leverage = code, content, capital That’s how you scale without trading time. Stop chasing more hours. Start building things that work for you even when you’re offline.

Day 13/100 of One Person Business If your price feels “safe,” it’s probably too low. Charge more than you’re comfortable with. Low prices attract people shopping for cheap. High prices attract people paying for results. Price changes perception. People don’t buy effort. They buy outcomes. The right clients don’t want cheap. They want certainty, speed, and trust. Stop pricing to be liked. Start pricing to be respected.

Everyone wants the next 10x stock. But most fortunes were quietly built in $MSFT and $AAPL Here’s the uncomfortable truth: Most people are not bad at picking stocks. They’re bad at staying in the game. Everyone looks like a genius in a bull market. Every chart goes up. Every dip gets bought. Every portfolio is green. Then the market drops 10-20%. Suddenly the question becomes: “Do I still believe in this… or should I sell?” That’s where most people lose. Not because the companies changed. But because their emotions did. The market doesn’t reward excitement. It rewards discipline, conviction, and time. You do not need to chase the fastest-moving names if they keep you awake at night. Sometimes slow money is the smartest money. Because getting rich slowly with companies that have already proven themselves is far better than chasing fast gains and panic-selling at the first sign of pain. The real game is simple: survive long enough for compounding to do the heavy lifting. That’s how wealth is actually built. Not in days. Not in weeks. But over years.

Day 12/100 of One Person Business Most people don’t fail because they’re lazy. They fail because they build something nobody wants. “Make something people want.” A beautiful product means nothing if it solves no real pain. Before you build, ask: Does anyone actually need this? Will this save time, make money, reduce pain, or create status? If the answer is no, it’s not a business. It’s just a hobby.

Day 11/100 of One Person Business Simple sells. Confusion doesn’t. “Simple sells.” If people need 30 seconds to understand what you do, you already lost them. One clear problem. One clear promise. One clear outcome. That’s it. The best businesses aren’t always the smartest. They’re the easiest to understand. If you can’t explain it in one sentence, it’s too complicated.

Day 10/100 of One Person Business Your idea is worth $0 until the market sees it. Ship first. Learn later. The internet pays creators who publish, not thinkers who plan.