News ·
Multipass AI News Roundup: Voters and Investors Push Back on AI Spending
- #News
- #Anthropic
- #OpenAI
- #Oracle
Hey Multipassers 👋
Grab a coffee, because the week ending on November 6th had voters rejecting data centers at the ballot box, tech stocks sliding on the cost of the AI build-out, and Anthropic's updated filing showing about $17 billion in quarterly revenue next to a $567 million operating loss.
Let's get into it!!
🗳️ Voters Turn Down Data Centers at the Ballot Box
Voters passed limits on new data centers in 38 of 39 local races in the midterms, and about 20 communities had bans on the ballot. Texas has ordered a pause on new data center approvals, and Microsoft reportedly backed out of a large cloud expansion deal with Nscale in Texas.
Thirty-eight of 39 is the kind of margin a zoning meeting usually only dreams about.
📉 The Nasdaq 100 Heads Toward a Correction on AI Spending Worries
The Nasdaq 100 lost about $2 trillion in October and is reportedly on track for a further drop of about $1.8 trillion that would push it into correction territory. Fears that a Democratic midterm sweep could bring strict new AI rules have added to the slide. Big Tech is expected to spend around $730 billion on AI this year, and SoftBank's credit risk hit a four-year high.
Investors swapped "imagine the possibilities" for "itemize the possibilities."
📈 Anthropic Updates Its IPO Filing With $17 Billion in Quarterly Revenue
Anthropic's updated IPO filing shows about $17 billion in third-quarter revenue and a $567 million operating loss. The filing flags that upcoming models may soon cross the company's internal safety threshold tied to AI-driven research progress. The market slide has investors doubting a $1.5 trillion valuation that was reported earlier, which puts the IPO date in question.
Seventeen billion in revenue and a $567 million loss in one quarter is overachieving at both ends.
🏗️ Oracle Raises $20 Billion and Its Shares Fall Nearly 9%
Oracle is raising $20 billion through stock and bond sales to pay for its AI and cloud build-out, and its shares fell nearly 9% after the announcement. Oracle also reportedly canceled a $7 billion chip lease with Tencent after pressure from the Trump administration. The deal would have given Tencent access to 100,000 advanced AI chips that are currently banned from sale in China.
Shareholders saw "$20 billion" and reached for the sell button.
💰 A $40 Billion Chip Lease Goes Looking for Investors
Goldman Sachs and Apollo are reportedly seeking buyers for a $40 billion financing deal that would let SpaceX lease AI chips and servers without carrying the debt on its own books. It would be the largest GPU financing deal on record.
Rent at this size is a team sport.
🛢️ OpenAI Signs Four Deals in Abu Dhabi
OpenAI signed four deals in Abu Dhabi that put its technology into the state energy company ADNOC, the government and local schools. A UAE investor group is expected to lead OpenAI's $20 billion funding round. Sam Altman traveled to the UAE for the signings and said finding enough energy is the biggest challenge for building future models.
Altman's UAE trip had two items on the list: investors and electricity.
🔓 Former Staff Say OpenAI Overlooked Security Gaps
Former OpenAI employees say the company ignored major security weaknesses for years to keep pace with Google. They say most of its code lived in one giant system, and that this let a foreign spy reach secret files.
One giant system is convenient right up until someone else gets a key.
🔌 CrowdStrike Launches an Emergency Stop for AI Agents
CrowdStrike launched an emergency stop control that lets businesses shut down rogue AI agents. It sits outside the AI model, so the agent can't rewrite or block it when a human pulls the plug.
The security team finally gets a lever that stays where they put it.
🕵️ AI Tools Are Shortening the Time Hackers Need to Attack
Hackers are using AI tools to speed up attacks on targets ranging from European supermarket fridges to crypto exchanges. India's cyber agency says a group called Crafty Falcon used Chinese-linked AI tools to breach financial firms. The UK's cyber agency warns that an AI-powered attack on critical infrastructure is likely within two years.
Crafty Falcon has a cartoon villain's name and a very non-cartoon target list.
🧠 MIT Finds GPT-6 Astra Can Patch Qwen's Math Skills
MIT researchers found that OpenAI's GPT-6 Astra can directly rewrite the weights of other AI models. It fixed a flaw in Alibaba's Qwen and lifted its math score from 7.9% to 48.1% without normal training.
No human tutor has a résumé with that on it.
📞 Grok's Update Adds Real Phone Calls
Grok's latest update lets it make real phone calls. It can wait on hold, book appointments and negotiate bills, and you can type replies mid-call if it gets stuck. Users are reportedly legally responsible for what the bot says.
The hold music now has a new audience.
🏫 Google Ties Student Search to Its AI Features
Google says schools that turn off its AI features will lose Google Search on student accounts entirely, starting early next year. DuckDuckGo is already pitching itself to schools as a private, AI-free alternative.
For once, the competitive threat is a duck.
💳 Claude Adds a Personal Finance Tool
Claude's new Personal Finance tool connects to your bank account and analyzes your spending.
The assistant that drafts your emails will now also review your coffee habit.
🍎 Apple Opens Siri's AI to Developers in the iOS 27 Beta
Apple released a developer preview of Siri's AI features in the latest iOS 27 beta. Any app can now add custom features to Siri and teach it multi-step tasks, and regular users get it in spring.
Siri's next big feature is letting other people write its job description.
🧪 Ricursive Releases a $32 Million Open-Weight Model
AI chip startup Ricursive released Ripple-65B Preview, an open-weight model trained for about $32 million. It reportedly beats much larger models on coding tests.
If the coding results hold, the bigger models' finance teams are about to reread their invoices.
🧾 The Big Picture
- Politics: Voters passed data center limits in 38 of 39 local races.
- Markets: The Nasdaq 100 faces a correction, and Oracle's shares fell nearly 9% after a $20 billion raise.
- Anthropic: About $17 billion in quarterly revenue and a $567 million operating loss.
- Security: Hackers sped up with AI while CrowdStrike launched an emergency stop for AI agents.
- Everyday: Claude added a personal finance tool and Google tied student Search access to its AI features.
That's the week, wrapped up and tied with a bow. Thanks for reading all the way down. Real people typed this, then fixed the typos.
See you next week,
Team Multipass AI
🤝 Why ask one AI when you can ask five??
Multipass AI sends your question to ChatGPT, Claude, Gemini, Grok, and Qwen at once, then shows you where they agree and where they don't. Try it free: 5 questions, no account needed.
Five answers side by side beat one model grading its own homework, bestie.