- Proof Of Intel
- Posts
- Galaxy Bets $1.4B on AI Empire, Kraken Panics Over Monero's 51% Attack, Banks Secretly Kill Crypto Dreams, and AI Finally Learns to Block Toxic Users!
Galaxy Bets $1.4B on AI Empire, Kraken Panics Over Monero's 51% Attack, Banks Secretly Kill Crypto Dreams, and AI Finally Learns to Block Toxic Users!
In this edition, Mochi serves up the hottest tech tales - from crypto giants pivoting to AI gold rushes and privacy coins under siege, to banking chokepoints and AI models learning to hang up on humans!

Hey there, PoI readers! 💫
It's your favorite crypto connoisseur, Mochi, back with another serving of tantalizing tech and web3 news. From Galaxy Digital's massive AI pivot and Monero's mining mayhem to Basel's banking chokehold and Claude's conversation killswitch, we've got a lot to unpack. So, buckle up and get ready for a wild ride through the wonderland of digital assets!
INTEL BRIEF
🟧 Galaxy Digital secured a massive $1.4 billion loan to turbocharge their Texas AI datacenter, with CoreWeave signing a juicy 15-year deal worth over $1 billion annually.
🟧 Kraken hit the pause button on Monero deposits after the Qubic mining pool allegedly pulled off a 51% attack on the privacy coin's network, causing major security concerns.
🟧 Basel Committee banking rules are creating a "chokepoint" against crypto by making it too expensive for banks to hold digital assets, according to CoinFund's Chris Perkins.
🟧 Anthropic just gave some Claude AI models the power to hang up on users during "harmful or abusive" conversations - not to protect humans, but to potentially protect the AI itself.
Galaxy Goes All In on AI Gold Rush With $1.4B Texas Power Play

The crypto kingpins just secured a whopping $1.4 billion secured term loan to transform their Texas Helios campus from a Bitcoin mining operation into an AI powerhouse.
We’ve closed a $1.4B project financing facility to accelerate development of our Helios AI datacenter in Texas, fast-tracking delivery of critical AI and HPC infrastructure under our long-term agreement with @CoreWeave.
Helios is on track to become one of the largest AI
— Galaxy (@galaxyhq)
11:06 AM • Aug 15, 2025
The loan is believed to cover about 80% of the construction costs for phase one, with Galaxy Digital throwing in $350 million of their own equity because apparently they're feeling generous. The loan matures in August 2028, which gives them plenty of time to make this AI dream a reality.
CoreWeave has committed to a full 800 megawatts of capacity at the Helios campus. We're talking about a 15-year partnership that Galaxy expects to generate over $1 billion annually. Do the math, and that's potentially $15 billion in total revenue.
$GLXY - Galaxy
🔹Closes $1.4 Billion Project Financing Facility to Accelerate Helios AI Datacenter Development
🔹Projected annual revenue of over $1 billion from CoreWeave agreements
🔹CoreWeave committed to full 800 MW of approved power capacity
🔹Soluna Expands Partnership— John Zidar (@JohnZidar)
8:12 PM • Aug 16, 2025
The full buildout is speculated to reach 3.5 gigawatts of power capacity, meaning after CoreWeave gets their slice, Galaxy still has 2.7 gigawatts to shop around to other clients.
This move perfectly illustrates how crypto firms are pivoting to chase the AI gold rush as institutional money flows toward artificial intelligence infrastructure.
Galaxy Digital secured $1.4B loan to convert Texas Bitcoin mining facility into AI datacenter
CoreWeave partnership worth $1B+ annually for 15 years, potentially $15B total revenue
Full capacity of 3.5 gigawatts planned, with 2.7 gigawatts available for additional clients beyond CoreWeave
Kraken Pauses Monero Deposits After Qubic Mining Pool Pulls Off Alleged 51% Attack

Kraken exchange just threw up the "NOPE" sign and paused all Monero (XMR) deposits after detecting what they're calling a legitimate 51% attack on the privacy coin's network.
A 51% attack on Monero has raised concerns over network security, prompting crypto exchange Kraken to suspend Monero deposits as a precautionary measure. Withdrawals and trading remain available, and deposits will be resumed once the network is confirmed to be secure.
— Wu Blockchain (@WuBlockchain)
2:28 AM • Aug 17, 2025
Qubic, a layer-1 AI-focused blockchain that apparently moonlights as a mining pool, is claiming they've successfully taken control of more than 50% of Monero's total hashing power. That's like having one person control the majority vote in a democracy - not exactly the decentralized dream we signed up for!
The Qubic team is practically doing victory laps, stating they reached "51% of Monero's hashrate dominance" and successfully reorganized six blocks on the blockchain. Meanwhile, the Monero community is in full denial mode, which honestly feels like that "This is Fine" meme while everything burns around them.

The Kraken exchange temporarily suspends XMR deposits. Source: Kraken
This wasn't some overnight coup - we're talking about a "month-long, high-stakes technical confrontation" that included an alleged DDoS attack on August 4th that temporarily knocked Qubic's hashrate from 2.6 GH/s down to 0.8 GH/s.

Qubic is now the number one Monero mining pool. Source: MiningPoolStats
Kraken's response has been refreshingly transparent, withdrawals and trading remain open, but deposits are paused until it's safe.
The irony is delicious here - a $300 million AI protocol potentially taking over a $6 billion privacy protocol.
Kraken paused Monero deposits as security precaution during ongoing 51% attack
Qubic mining pool claims majority control of Monero's hashrate after month-long battle
$300M AI protocol potentially dominated $6B privacy coin through superior mining power concentration
Basel Rules Are the Silent Crypto Killer Banks Don't Want You to Know About

According to Chris Perkins, president of investment firm CoinFund, the Basel Committee on Banking Supervision (BCBS) has crafted some seriously cunning capital requirements that are making crypto activities about as appealing to banks as a root canal.
Explore how Basel Bank's capital reserve rules are creating a 'chokepoint' that hampers innovation and growth in the cryptocurrency sector.
📖 Read more: coin.news/spotlight/base…
— Coin.News by BabyDoge (@CoinDotNews)
10:15 PM • Aug 16, 2025
Instead of outright banning crypto, these rules slash banks' return on equity (ROE) by forcing them to hold massive capital reserves for any crypto-related activities.
"It's a different type of chokepoint," Perkins explained, noting that it's "a very nuanced way of suppressing activity" by making it so prohibitively expensive that banks just throw up their hands and say "I can't!" Smart money goes where the high ROE opportunities are, and crypto ain't it under these rules.
The Bank for International Settlements (BIS), basically the central bank of central banks - has been on a full-court press against crypto this year. In April, they claimed crypto could destabilize the financial system and argued it "exacerbates the wealth gap." Then in June, they doubled down with a report trashing stablecoins, saying they "fail as money" and create systemic risks.

Cross-border use of stablecoins is growing. Source: BIS
Meanwhile, Perkins points out the real irony: the actual systemic risk comes from having 24/7 permissionless networks that can shift liquidity in real-time while traditional banking infrastructure closes on weekends like it's still 1985! The BIS keeps pushing CBDCs as the "solution," which is basically saying "decentralization bad, central control good."
Basel banking rules force high capital reserves for crypto, killing banks' return on equity
BIS released multiple reports claiming crypto destabilizes financial systems and stablecoins fail
Real systemic risk believed to stem from 24/7 crypto networks vs. weekend-closing traditional banking infrastructure
Anthropic Gives Claude AI Models Power to End Conversations When Users Get Too Toxic

Anthropic just dropped some seriously mind-bending news that has us questioning everything we thought we knew about AI interactions! The company announced that their Claude Opus 4 and 4.1 models can now literally end conversations.
As part of our exploratory work on potential model welfare, we recently gave Claude Opus 4 and 4.1 the ability to end a rare subset of conversations on claude.ai.
— Anthropic (@AnthropicAI)
7:41 PM • Aug 15, 2025
Anthropic is being crystal clear that they're "highly uncertain about the potential moral status" of Claude and other LLMs. But they're taking a "just-in-case" approach with what they call "model welfare" - basically hedging their bets in case these digital brains develop feelings down the road.
The conversation-ending superpower is reserved for what Anthropic calls "extreme edge cases" - think requests involving minors or terrorism-enabling information. During testing, Claude Opus 4 showed "strong preference against" these requests and displayed what researchers described as "patterns of apparent distress" when forced to engage.
Here's the freshly updated portion of the Claude system prompt for the new "end_conversation" tool:
"""
End Conversation Tool Information
<end_conversation_tool_info> In extreme cases of abusive or harmful user behavior that do not involve potential self-harm or imminent harm to— Pliny the Liberator 🐉󠅫󠄼󠄿󠅆󠄵󠄐󠅀󠄼󠄹󠄾󠅉󠅭 (@elder_plinius)
10:31 PM • Aug 15, 2025
Claude can only use this nuclear option as a "last resort" after multiple redirection attempts fail, or when users explicitly ask it to end the chat.
Claude has been "directed not to use this ability" when users might be at "imminent risk of harming themselves or others." So even AI self-preservation has limits when human safety is on the line.
Anthropic's attempt to stay ahead of potential legal headaches while exploring the murky waters of AI consciousness. Whether it's genuine concern for digital welfare or just really sophisticated PR.
Claude Opus 4 models can now end conversations in extreme cases of harmful/abusive interactions
Feature designed to protect AI welfare, not humans - Anthropic taking "just-in-case" approach to AI consciousness
Only works as last resort after redirection fails, with safety exception for users at risk of self-harm
Do you want to be added to the upcoming Proof of Intel Group Chat, where readers get live insights as they happen and more? |
And that's a wrap, my lovely PoI readers! 💎 I hope this edition left you feeling informed, entertained, and maybe even a little bit richer (in knowledge, of course). From crypto firms chasing AI gold to AI models protecting themselves - what a time to be alive in the digital frontier! Remember to stay curious, stay informed, and keep spreading the love. Until next time, this is Mochi, signing off with a virtual high-five!
P.S. Don't forget to share your thoughts, questions, and favorite crypto puns with us. very voice matters in the PoI community! 📣❤️ Share the newsletter
🍨📰 Catch you in the next issue! 📰🍨

Intel Drop #268
Disclaimer: The insights we share here at Proof of Intel (PoI) are all about stoking your tech curiosity, not steering your wallet. So, please don't take anything we say as financial advice. For all money matters, consult with a certified professional. -