🔄 Musk Restarts Supercomputer He Killed

Plus Pinterest calls AI piracy and spending hits $2.5 trillion

January 20, 2026

In partnership with

Reading time: 5 minutes

🗞️In this edition

  • Musk Restarts Tesla Supercomputer He Killed Last Year

  • Global AI Spending Hits $2.5 Trillion Despite Buyer's Remorse

  • Pinterest CEO Says AI Companies Are Pirating Content Like Napster

  • In other AI news –

    • Musk Says Work Will Be Optional in 10-20 Years

    • Banks Bet on AI and Blockchain Working Together

    • IBM Launches AI Service to Build Company Platforms

  • 4 must-try AI tools

Three numbers tell you where AI actually is right now.

$2.5 trillion being spent this year. A supercomputer project killed and restarted within months. And one CEO comparing the entire industry to music piracy.

Each story reveals something about confidence versus reality. About companies spending record amounts while admitting disappointment. About reversing decisions made less than a year ago. About business models that don't work for anyone except the companies scraping content.

The gap between the spending and the results keeps growing. So does the gap between what leaders promise and what they deliver.

This edition looks at what's actually happening behind the forecasts and announcements.

What's happening:

Elon Musk announced Wednesday that Tesla is restarting its Dojo supercomputer project. He killed the program less than a year ago to focus on buying Nvidia chips instead.

Tesla's AI5 chip is now ready and will power Dojo3, the new supercomputer. The company is recruiting engineers to work on 9-month development cycles starting immediately.

Musk said Tesla made a mistake shutting down Dojo. The team spent years building custom AI chips, then abandoned them when Nvidia's H100 looked better. Now they're going back.

The announcement came with new details about space-based data centers. Musk said cooling is the real constraint, and space solves that. He called it "extremely likely" Tesla will eventually use orbital computing.

Tesla stock jumped on the news. Investors see this as Tesla committing to in-house AI infrastructure rather than depending on Nvidia's supply chains and pricing.

Why this is important:

Even Musk makes expensive mistakes. Tesla spent billions building Dojo, then killed it to buy Nvidia chips instead. Now he's admitting that was wrong.

Nvidia chips are expensive and hard to get. Companies that canceled their own chip projects are realizing they lost bargaining power.

Tesla needs this for self-driving cars. The cars create tons of training data. Chips built specifically for that data might work better than Nvidia's general-purpose chips.

The space data center talk is classic Musk. He mixes real news with future dreams. Dojo3 is happening now. Space computers are just talk for now.

Nine months to build new chips is fast. Maybe too fast. Either Tesla found a shortcut, or they're setting themselves up to miss deadlines.

Introducing the first AI-native CRM

Connect your email, and you’ll instantly get a CRM with enriched customer insights and a platform that grows with your business.

With AI at the core, Attio lets you:

  • Prospect and route leads with research agents

  • Get real-time insights during customer calls

  • Build powerful automations for your complex workflows

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

What's happening:

Gartner forecasts global AI spending will reach $2.52 trillion in 2026, a 44% jump from last year. Infrastructure accounts for more than half, with $401 billion going to servers and data centers.

The spending comes as Gartner says AI entered the "Trough of Disillusionment" in 2026. Companies are disappointed but still spending.

Most AI purchases will come from existing software providers like Microsoft and Google, not new moonshot projects. Companies want predictable returns before scaling AI.

Gartner added $500 billion to its forecast since September. The firm expects hundreds of AI companies won't survive past 2027 as consolidation begins.

Why this is important:

Companies are spending despite being disappointed. That's not confidence. That's fear of falling behind competitors.

The "Trough of Disillusionment" label matters. Companies are unhappy with AI results but keep buying anyway. That's not sustainable.

Infrastructure getting half the spending shows the bet is on future capacity, not present results. Companies are building for what might work, not what works today.

Comments from the editor:

Spending $2.5 trillion while calling it a "trough of disillusionment" is wild. That's record investment in something Gartner admits isn't delivering.

Companies tried bold AI bets. They failed. Now they're buying from Microsoft and Google because it's safer, not better.

MIT research shows 95% of companies reported zero return on AI projects in 2025. Yet spending is up 44%. That only works if everyone thinks they'll be the lucky 5%.

The consolidation will be brutal. Hundreds of startups will shut down. Investors will lose billions on failed AI bets.

What's happening:

Pinterest CEO Bill Ready wrote that AI's current approach resembles Napster, where millions downloaded music for free, instead of iTunes or Spotify, where publishers get paid every time their work is used.

Ready says content creators are no longer powerless. New frameworks let them choose whether and how AI companies use their content. Cloudflare launched a pay-per-crawl service that distinguishes AI crawlers from search crawlers.

The difference matters. Search crawlers drive traffic back to the original source. AI crawlers take information without sending users back to creators. That's the Napster model, not the iTunes model.

Ready argues the "Napster phase" needs to end. It's time for a clear exchange of value that benefits content creators, not just AI companies.

Why this is important:

This is Pinterest's CEO saying AI companies are stealing. That's strong language from someone running a platform built on sharing content.

The Napster comparison is deliberate. Napster killed the music industry's business model until iTunes and Spotify created legal alternatives. Ready is saying AI needs the same transformation.

Cloudflare's tool shows there are technical solutions. The question is whether AI companies will adopt them or keep scraping for free until regulation forces change.

For content creators, this matters now. Every website deciding whether to allow AI training affects what data future models have access to. Block AI now or accept getting scraped forever.

  1. MakeShorts - An AI-powered tool that allows users to turn YouTube videos into short clips

  2. Civitai - An online platform that makes it easy for people to share and discover resources for creating AI art

  3. GetMax - An AI-powered content marketing assistant that helps businesses plan, create, and optimize their content strategies

  4. WebWhiz - A tool that allows users to create a chatbot with AI capabilities to answer customer queries on their website

The contradictions are getting harder to ignore.

Record spending during what analysts call a "trough of disillusionment." Killing projects then bringing them back. Calling competitors thieves while building the same scraping tools.

None of these are signs of a healthy industry finding its way. They're signs of an industry spending billions while figuring out what works.

What matters now is whether the spending creates anything sustainable. Or whether it's just momentum carrying everyone forward until someone realizes the numbers don't add up.

If you have any feedback for us, please reply and let us know how we did. We're always looking to improve.

How did we like this version?

Login or Subscribe to participate in polls.

Interested in featuring your services with us? Email us at [email protected]