Meta just signed a 20-year deal to secure nuclear power, as AI pushes tech giants to find new energy sources.
Key Points:
- Meta will boost output at an Illinois nuclear plant by 30 megawatts starting in 2027.
- The deal helps power Meta’s growing AI operations while cutting fossil fuel use.
- Big Tech is turning to nuclear to meet energy demands and meet climate goals
Details:
Meta is locking in nuclear energy to fuel its future AI systems. The company announced a long-term agreement with Constellation Energy to tap into clean power from Illinois’ Clinton Clean Energy Center. This comes just as the plant’s state support ends in 2027. The added power will support Meta’s AI growth, create over 1,000 jobs, and generate $13.5 million in local taxes. Meta joins Microsoft, Amazon, and Google, all of which made nuclear investments in the past year, as part of a broader shift by tech firms to secure stable, low-emission power for energy-hungry data centers and AI workloads.
Why It Matters:
AI systems are pushing energy demand to new limits, and reliable power is becoming the bottleneck. Amazon, Google, Microsoft, and Meta are racing to secure long-term nuclear deals, while China is already far ahead in building out the energy supply for its AI goals. The real competition now is not just about models, but about who can power them.
The $1.5B AI startup that wasn’t AI at all
Builder.ai, once worth $1.5 billion, is bankrupt. The company claimed to use AI to build apps, but it was actually humans pretending to be bots.
Key Points:
- The company claimed to use AI to build apps.
- It actually used developers pretending to be bots.
- It raised $450M from big names like Microsoft and QIA.
Details:
Builder.ai raised $450 million by claiming it used AI to help companies build apps. But it didn’t. The “AI” was actually a team of developers in India pretending to be bots. Big names like Microsoft and the Qatar Investment Authority backed it. QIA alone invested $250 million. For years, the company kept up the act, until things started to fall apart. The company admitted it had inflated sales numbers and brought in auditors to review its books. Not long after, a lender pulled $37 million from its accounts, leaving just $5 million behind. That wasn’t enough to keep things running. Most of the team was laid off, and the company began bankruptcy proceedings across five countries.
Why It Matters:
This wasn’t just a failed startup.. it was a company that tricked investors for years. QIA alone lost $250 million. It shows how easy it is to slap “AI” on a product and raise huge sums without delivering anything real. As AI hype continues, this is a warning: not everything labeled AI actually is, and not every pitch deserves a $1.5 billion valuation.
Apple’s AI strategy still playing catch-up
Apple is preparing to show off its AI progress at WWDC, but the real breakthroughs are still stuck behind closed doors.
Key Points:
- Apple will let developers use its small on-device AI models for simple features like summarizing text.
- Its biggest AI model performs on par with ChatGPT but won’t be released to the public.
- Major upgrades like a new Siri and health tools are delayed to 2026 or later.
Details:
At WWDC, Apple will open up its own AI models to developers. These models are lightweight, with around 3 billion parameters, and can power basic features inside apps. But Apple’s most advanced work is happening in private. Internally, the company is testing much larger models, including one with 150 billion parameters that reportedly matches ChatGPT in performance. Other upcoming features include small updates like an AI translation app, a rebranded Safari, and a new battery-saving mode. But many of the big AI projects, including a more conversational Siri and a smarter health assistant, are still far from launch.
Why It Matters:
Apple missed the first wave of generative AI and hasn’t done much to catch up. While others launched new chatbots, AI search, and voice tools, Apple stayed quiet. Even now, its strongest model is kept internal and real features are either minor or delayed. The past year hasn’t inspired confidence that Apple can become a serious AI player anytime soon.
