Minds, Bodies, and Terawatts

Google's 2026 environmental report reveals its annual electricity consumption rose 37% in 2025 — the biggest increase in company history — driven by AI data center construction. Its data centers now burn more power than entire countries, and Google concedes its buildout is outpacing grid decarbonization even as it buys clean energy. Amazon reported record emissions the same week.

Continue the story on unscarcity.ai:

• The Electron Gap: Why Energy Is the Real Bottleneck in the AI Era
• Who Pays for AI's Electricity? The Grid Cost Fight

Full episode notes on unscarcity.ai

In a Forbes exclusive on July 2, Boston Dynamics revealed its fifth-generation Atlas humanoid is 'almost an order of magnitude' simpler to build, with far fewer and less unique parts, making it cheaper and more reliable to manufacture. Director of Robot Behavior Alberto Rodriguez says performance is equal or better and that software, not hardware, is now the limiting factor. Parent company Hyundai plans to build at least 30,000 units a year.

Continue the story on unscarcity.ai:

• The Humanoid Robot Revolution: 50,000 Units Shipping in 2026
• The Substitution Threshold: When You Stop Being the Cheapest Option

Full episode notes on unscarcity.ai

OpenAI has held early talks to hand the US government a 5% equity stake, worth about $42.6 billion, and Sam Altman proposed that every leading US AI lab contribute the same share into a public wealth fund modeled on Alaska's Permanent Fund, which pays annual dividends to residents. Altman raised it with Trump, Commerce Secretary Lutnick, and Treasury Secretary Bessent; the FT called the talks conceptual and early-stage.

Continue the story on unscarcity.ai:

• When AI Goes Public: Shareholders vs. Abundance
• The Bootstrap Paradox: How Do You Fund the Death of Money Using Money?

Full episode notes on unscarcity.ai

The Wall Street Journal reported that SpaceX showed investors a slim, iPhone-like AI device built on xAI's Grok technology ahead of its record June IPO. Elon Musk denied the report on X as 'utterly false,' and SpaceX's newly listed stock dropped about 7.3%, wiping more than $50 billion off Musk's net worth. The move would pit SpaceX against OpenAI's rival AI-device effort with Jony Ive.

Continue the story on unscarcity.ai:

• Network Effects: Why Facebook Won and Why AI Monopolies Loom
• When AI Goes Public: Shareholders vs. Abundance

Full episode notes on unscarcity.ai

Multiple outlets reported July 1 that Meta is building a cloud unit to sell access to its excess AI computing power and models, directly challenging Amazon, Microsoft and Google. The pivot mirrors SpaceX/xAI, which now leases out data-center capacity, and sent Meta shares up 8-10 percent in a single day.

Continue the story on unscarcity.ai:

• AI Compute Clusters 2026: The $800B GPU Arms Race

Full episode notes on unscarcity.ai

The first complete 30-day cycle under GitHub Copilot's new usage-based billing ended June 30, with agentic developers reporting bills 10 to 50 times higher than their old flat-rate plans. Bills jumped from $29 to $750 and from $50 to $3,000 for power users running multi-step autonomous coding sessions.

Continue the story on unscarcity.ai:

• Software Developer Unemployment: AI Ate the Best Job
• The Substitution Threshold: When You Stop Being the Cheapest Option

Full episode notes on unscarcity.ai

Elon Musk announced June 28 that xAI's Grok 4.5, a 1.5-trillion-parameter model built on the V9 architecture, has entered private beta testing exclusively at SpaceX and Tesla. Musk claims performance may exceed Anthropic's Claude Opus, but no independent benchmarks exist. xAI plans monthly model releases through end of 2026 while posting a $2.47 billion operating loss in Q1.

Continue the story on unscarcity.ai:

• AI Compute Clusters 2026: The $800B GPU Arms Race
• 3 AI Futures: Elite Capture, Peaceful Transition, or Chaos?

Full episode notes on unscarcity.ai

The US government restricted both OpenAI's GPT-5.6 Sol and Anthropic's Mythos 5 to handpicked circles of government-approved companies this week, creating an unprecedented two-tier AI access system. Commerce Secretary Lutnick signed clearance letters after two weeks of security talks, citing cybersecurity concerns after Anthropic's Mythos found vulnerabilities in classified US systems within hours and OpenAI's Sol emerged as its most capable cyber model.

Continue the story on unscarcity.ai:

• Anthropic Pulled the AI Alarm. Who Can Act on It?
• 3 AI Futures: Elite Capture, Peaceful Transition, or Chaos?

Full episode notes on unscarcity.ai

Governor Newsom partnered with the Employment Development Department and UC California Policy Lab to launch the AI-Unemployment Tracker, the first government tool in the nation to correlate AI exposure metrics with real-time unemployment insurance claims data. Initial findings through May 2026 show no statewide surge but localized elevated claims among college-educated workers in high-exposure Bay Area roles.

Continue the story on unscarcity.ai:

• The Substitution Threshold: When You Stop Being the Cheapest Option
• AI Layoffs 2025-2030: AI Now #1 Reason for US Job Cuts

Full episode notes on unscarcity.ai

OpenAI and Broadcom on June 24 unveiled Jalapeño, OpenAI's first custom-built AI inference chip, designed end-to-end in nine months with assistance from OpenAI's own AI models. The chip cuts inference costs roughly 50% versus current GPUs, is fabricated on TSMC's 3nm process, and will deploy at Microsoft data centers by year end.

Continue the story on unscarcity.ai:

• AI Compute Clusters 2026: The $800B GPU Arms Race
• Altman's UBC: Why Give Computing Power Instead of Cash?

Full episode notes on unscarcity.ai