HashAndPower (@HashAndPower) header
HashAndPower (@HashAndPower) profile picture

Software Engineer by profession. Skeptic by nature. AI needs power. Machines need money. Watching the scarce inputs, not the narrative.

대한민국
Joined July 2025
Photo by HashAndPower (@HashAndPower)Photo by HashAndPower (@HashAndPower)Photo by HashAndPower (@HashAndPower)Photo by HashAndPower (@HashAndPower)Photo by HashAndPower (@HashAndPower)Photo by HashAndPower (@HashAndPower)
The market is bigger than all of us. Stay humble. Stay strong. $BTC $KAS
🤖 Made with AI
4
7
1
44
2,022
Replying to @jukan05
Interesting how the market focuses on the negatives. 🤬
1
1,335
Replying to @linbo_tang
Bro, believe it or not, I'm already number one in my field! 😎
1
2
18
Replying to @linbo_tang
Become the best in your field. The money will follow. 💰
1
1
25
Replying to @Kas123theking
Nice move, bro! Patience and conviction will pay off in the long run. Let's see where $KAS takes us! 🔥
34
Replying to @linbo_tang
Bro, study hard! I'll keep working hard too!
1
1
10
Replying to @linbo_tang
The future is bright, bro! 😎
1
1
14
Replying to @linbo_tang
Shouldn't you ask if I have any cash left first? 😂
1
1
46
Amazon is unlocking billions from its data centers while keeping long-term access to the infrastructure. This is what happens when AI infrastructure becomes an institutional asset class. Power is the foundation. Long-term leases unlock its value. Now imagine what 4.8 GW could mean for $MARA. 🔥
AMAZON IS SELLING DATA CENTERS & RENTING THEM BACK Blackstone's new data center trust is buying a 75% stake in two $AMZN data center campuses valued at about $5 billion. The deal covers 13 newly built data center buildings with 647 megawatts of power capacity, nine in northern Indiana near Chicago and four in Mississippi. Amazon keeps the other 25% and will lease back all 13 buildings for at least 12 years, with options that can stretch to 32. The price works out to a 7% initial yield, and the rent rises 2.75% a year. At closing, Amazon gets a cash payout funded by $1.9 billion from Blackstone and about $2.5 billion of new mortgage loans on the buildings. Amazon said the partnership lets it keep expanding as demand grows. Meta did a version of this last year, when funds run by Blue Owl took an 80% stake in its $27 billion Hyperion data center campus in Louisiana. It is the first deal for Blackstone Digital Infrastructure Trust, which raised $1.75 billion in its IPO in May. It is expected to close by year-end.
1
7
167
You're responsible for your own investment decisions. There's no need to blame others. If someone told you to buy and you did, whose decision was that? If you still believe in it, DCA if it makes sense for you. If you don't, sell and move on. It's really that simple. Study #Kaspa 🔥 $KAS
Replying to @HashAndPower
You guys have been saying that for three years, and it’s cost 84% of people’s net worth
1
14
458
Since Q3, MARA has repaid $750M in Bitcoin-backed debt, freeing up 17,477 BTC in collateral. Total debt is down $1.4B (37%) YTD. BTC sales continue to fund AI infrastructure, debt repayments, and buybacks. A stronger balance sheet. A bigger AI ambition. Now it's all about execution. $MARA
After quarter-end, we used bitcoin sale proceeds to fully repay $750 million of bitcoin-backed debt, releasing all 17,477 bitcoin pledged as collateral. Since the beginning of 2026, we have reduced total debt by $1.4 billion, or 37%. Q3 letter: buff.ly/99aUBUg
1
9
410
Power is just the beginning. Exaion could turn MARA's megawatts into high-value AI revenue. The real upside is moving up the value chain. $MARA
I think Exaion is one of the most important pieces of the $MARA AI thesis: -Exaion just signed a $79 MILLION GPU-as-a-service contract with its first U.S. customer on 100% take-or-pay terms -Its pipeline has already grown to more than $3 BILLION in potential contract value, with roughly $1 BILLION targeted for delivery by the end of 2027 -Exaion gives MARA direct exposure to private cloud, inference, GPU-as-a-service, and sovereign AI, where governments and enterprises increasingly want compute infrastructure they can control -The real opportunity is margin expansion. MARA can monetize the same underlying power through infrastructure first, then capture additional economics from the GPUs and cloud services running on top of it This is how $MARA moves up the value chain. Not just selling access to megawatts, but turning those megawatts into higher-value AI revenue. And I still think sovereign AI is one of the most overlooked parts of the entire AI infrastructure buildout.
1
12
336
Replying to @realvijayk
I feel the same way, brother. Every single day, I wake up feeling grateful, happy, and genuinely moved by the gift of being alive. Life itself is a blessing. 💚
1
2
75
Studying the risks of AI is important. But so is moving fast. The global AI race is already underway. China isn't waiting. 🇨🇳🇨🇳🇨🇳
3
109
Fred, 4.8 GW is the potential. Long-term contracts are the proof. Execution is what turns that potential into shareholder value. Let's make it happen! 🔥🔥🔥 $MARA @fgthiel
5
78
4.8 GW. That's not just a number. That's the potential scale of MARA's power portfolio. AI needs compute. Compute needs power. And power is becoming one of the biggest bottlenecks. The real opportunity is turning that power into long-term AI infrastructure revenue. The market is still pricing MARA as a Bitcoin miner. For how much longer? 🔥🔥🔥 $MARA
1
3
12
354
Buying Kaspa today is like buying Bitcoin at $100. Most people won't realize it until it's too late. 🔥 $KAS
$KAS fam, Wanna see the future I'm dreaming of?
3
9
56
827
One day, we may look back at MARA trading in the $6 range and realize what an opportunity this was. Time will tell. ⏳ $MARA
2
24
564
Execute off-chain. Verify on-chain. 🔥 $KAS
When vProgs fully land on Kaspa $KAS, it’s going to be a big deal. Most smart contract chains work the same way: every node re-runs every contract’s code, so the whole network is limited by what one node can execute. Kaspa is taking a different route. With vProgs (verifiable programs), the code doesn’t run on the network at all. It runs off-chain, and you can write real, general-purpose programs instead of squeezing logic into a cramped on-chain VM. What makes that trustworthy is zero-knowledge proofs. When a program runs, it produces a small cryptographic proof that the computation was done correctly. Kaspa doesn’t re-execute anything; it just checks the proof, which is fast and cheap no matter how heavy the original computation was. If the proof checks out, the result is accepted with the full security of Kaspa’s proof-of-work consensus. So the base layer stays lean. It orders transactions and verifies proofs, and the heavy lifting happens elsewhere. Each vProg keeps its own state, so apps don’t fight over one global pile of shared state, and each one only has to prove its own activity. The part I think sets it apart is where this is headed. Other ecosystems do off-chain compute with on-chain proofs too, but they end up with a pile of separate rollups, with liquidity split between them and bridges in between. vProgs are designed so programs can eventually call each other and compose in the same transaction, all sequenced by Kaspa itself, on a fast proof-of-work blockDAG. If they pull that off, I don’t know of anything else that combines all of that.
1
1
14
366