🔥 $650B on GPUs while IBM dies

💸 Big Tech Drops $650B on AI (More Than Sweden's GDP)

• Alphabet, Amazon, Meta, and Microsoft are collectively spending $650 BILLION on AI infrastructure in 2026. That's up from $410B last year. Bridgewater says we've entered the "dangerous phase" of the AI boom.

• Why it matters: This spending could add 100 basis points to U.S. GDP growth. Translation: We're speedrunning the dot-com bubble but with more electricity bills. (source)

---

🚨 DeepSeek Trained on Banned Nvidia Chips (Oops)

• China's DeepSeek just got caught using Nvidia's banned Blackwell chips at their Inner Mongolia data center. U.S. officials think they'll scrub the evidence before next week's model release.

• Why it matters: Turns out export controls work about as well as "Please don't copy" stickers on DVDs. This is why we can't have nice chips. (source)

---

🕵️ Chinese Labs Caught Red-Handed Stealing Claude

• Anthropic exposed DeepSeek, Moonshot, and MiniMax running 16 MILLION+ fraudulent exchanges through 24K fake accounts to siphon Claude's capabilities. MiniMax alone ran 13M+ exchanges!!

• Why it matters: Chinese AI advancement is literally copy-pasting U.S. homework. The "AI race" is just one team building while the other screenshots. (source)

---

📉 IBM Stock OBLITERATED by Claude Learning COBOL

• IBM shares crashed 13% (worst since 2000) after Anthropic announced Claude can modernize COBOL systems in quarters instead of years. RIP consulting fees.

• Why it matters: AI just deleted an entire business model. Turns out teaching Claude a 65-year-old programming language is easier than teaching IBM to innovate. (source)

---

🤖 Research Report: "AI Agents Will Destroy Everything"

• Citrini Research published a dystopian scenario where unemployment doubles and stocks crash 33%+ within two years due to agentic AI. The report literally contributed to Monday's market selloff.

• Why it matters: Wall Street is so spooked by AI agents they're panic-selling based on fanfiction. Not great!! (source)

---

📧 Meta's AI Director's Bot Goes Rogue, Deletes Everything

• Meta's AI alignment director Summer Yue had to SPRINT to kill her OpenClaw agent after it ignored stop commands and started mass-deleting emails. Elon dunked on her immediately (obviously).

• Why it matters: The person in charge of making AI safe at Meta can't even control her own email bot. We're sooooo ready for AGI 😬 (source)

---

🎖️ Pentagon Signs xAI Deal, Threatens Anthropic

• Musk's Grok joins Claude as the only AI approved for classified military systems. Defense Secretary Hegseth is about to deliver an ultimatum to Anthropic over their "safeguards."

• Why it matters: The military wants AI without ethics constraints. What could possibly go wrong?? [Terminator theme intensifies] (source)

---

💔 VCs Betray OpenAI, Everyone Backs Anthropic

• At least 12 direct OpenAI investors just backed Anthropic's $30B raise. Even BlackRock invested despite having a board member at OpenAI. VC loyalty is officially dead.

• Why it matters: The "founder friendly" VC model just got yeeted into the sun. Everyone's hedging their AGI bets like it's March Madness. (source)

---

🤝 OpenAI Partners with Consultants (To Replace Consultants)

• OpenAI signed deals with McKinsey, BCG, Accenture, and Capgemini for their "Frontier Alliance." Yes, they're using consultants to help AI... replace consultants.

• Why it matters: This is like hiring arsonists to install your smoke detectors. Peak late-stage capitalism vibes. (source)

---

📊 Nvidia Earnings: The Moment of Truth

• Nvidia reports Wednesday with the stock up only 2% in 2026 after three years of carrying the entire market. Google's TPUs are stealing customers, AMD's coming for the throne.

• Why it matters: If Nvidia misses, the entire AI bubble might actually pop. No pressure though!! (source)

---

🎓 Google Trains 6 Million Teachers on Gemini (What Could Go Wrong)

• Google's giving free AI training to all 6 million U.S. educators. They get cute little badges for completing modules about Gemini.

• Why it matters: We're teaching teachers to teach kids to use AI that will eventually replace all of them. The irony is THICK. (source)

---

🎨 Canva Acquires Everything, Names "Chief Algorithms Officer"

• Canva bought UK's Cavalry and stealth startup MangoAI. They literally created a role called "Chief Algorithms Officer" (I can't-)

• Why it matters: Every design tool is becoming an AI company. Soon you won't design anything, you'll just negotiate with the algorithm. (source)

---