POLICY
OpenAI Offers Trump Administration a 5 Percent Ownership Stake
Here's a number that stops you cold: based on OpenAI's most recent valuation, a 5 percent government stake would be worth roughly $42.6 billion. That's not a rounding error. That's the size of the offer Sam Altman reportedly floated to the Trump administration as a way to keep regulators friendly and the public less hostile.
The pitch, first reported by the Financial Times, has apparently been in the works for a while. Altman is said to have first brought the idea to Trump early last year, framing it as a way for ordinary Americans to share in the financial upside of AI development. Whether that framing lands as genuinely democratic or as savvy political cover depends a lot on your level of cynicism.
The proposal isn't just about OpenAI. According to the FT, the idea would involve other major US AI companies offering the government similar ownership stakes. That's a significant ask, and it's far from clear that competitors would line up to hand Washington a slice of their businesses voluntarily.
What makes this worth paying attention to is the context. The Trump administration has been unusually interventionist when it comes to the AI industry, and not always in ways that benefit the leading players. Anthropic, one of OpenAI's closest rivals, has been designated a supply chain risk by the Pentagon. The administration also hit Anthropic's newest models with surprise export controls last month, yanking them from international markets and rattling confidence in the broader US AI sector.
OpenAI, by contrast, has largely stayed out of the crosshairs — and this proposal might help explain why. Giving the federal government a financial interest in your success is one of the more elegant ways to align incentives. It's harder to regulate something into the ground when you own part of it.
The idea also lands in the middle of a much larger debate about who actually benefits from the AI boom. Senator Bernie Sanders has pushed for a one-time 50 percent tax on AI company stock values to seed a sovereign wealth fund. The Trump administration has already taken a 10 percent stake in Intel and reportedly demanded revenue cuts from Nvidia and AMD on their China chip sales. The government-as-shareholder model is gaining traction across the political spectrum, just for very different reasons.
Whether this specific deal ever closes is genuinely uncertain. The talks are described as early-stage, and the logistics of getting multiple AI companies to agree on a uniform framework would be complicated at best. But the fact that the conversation is happening at all signals something important: the age of AI companies operating as if Washington is a distant observer is over. The government wants in, one way or another, and OpenAI has apparently decided it's better to hand them a seat at the table than wait to be dragged there.
Source: The Verge
AI
Google AI Expansion Drove a 37 Percent Surge in Electricity Use
Google's data centers now consume more electricity than the entire nation of Nigeria. That sentence deserves a moment to sink in. According to the company's latest sustainability report, its data centers pulled more than 42 million megawatt-hours of electricity in 2025, up from 30.6 million the year before — a 37 percent jump that represents the single largest annual increase in the company's history.
To put the trajectory in even sharper relief: Google's total electricity consumption has grown by more than 250 percent since 2019. The AI buildout is the primary driver, layered on top of existing growth from Google Cloud and YouTube. The company isn't apologizing for any of it, but it is at least being unusually candid about the tension between its climate commitments and its infrastructure ambitions.
The sustainability report includes a notably honest line acknowledging that the AI infrastructure expansion is currently moving faster than the electrical grid is decarbonizing. That's not a minor footnote — it's a direct admission that the clean energy math isn't keeping up with the compute math, and that the gap may widen before it closes.
The headline carbon number looks better than you might expect, though. Google says it actually reduced its operational emissions by 2 percent in 2025, even as electricity use spiked. The company has maintained a practice of matching 100 percent of its electricity consumption with renewable energy purchases for nine consecutive years, and in 2025 it signed agreements for 12 gigawatts of new clean energy — the largest such commitment it has ever made.
But the full picture is murkier. Supply chain emissions — the carbon generated by Google's contracted manufacturers and suppliers — grew by 25 percent last year. The culprit is an Asia-Pacific supply chain running on grids that haven't made meaningful progress toward clean energy. Factor that in, and Google's total emissions actually rose 18 percent between 2024 and 2025, landing at roughly 14.5 million metric tons of carbon dioxide equivalent.
For context, that puts Google's annual carbon footprint somewhere between the nations of Ivory Coast and Panama, around 100th place globally. It's a strange benchmark, but it captures the scale of what a single technology company's infrastructure appetite now represents.
The broader AI industry faces the same reckoning. Microsoft, Amazon, and Meta are all racing to build out data center capacity at a pace that grid infrastructure simply wasn't designed to handle. The clean energy purchases help, but they're increasingly a financial instrument as much as an environmental one — companies are buying credits and capacity that may not translate to actual grid decarbonization in the near term.
Google's willingness to publish these numbers openly is worth crediting. The harder question is whether transparency alone changes anything, or whether it's just well-branded accountability theater while the power bills keep climbing.
Source: Ars Technica
Enjoyed this?
Get stories like this delivered every Tuesday — free.