Done by 10 AM. Here’s how.

The detail that matters in this week’s Georgia Power filing is not the megawatts. It is the financing model.

Alphabet’s Google will finance so-called uprates at the Vogtle and Hatch nuclear plants in Georgia via a new tariff. The agreement would allow the facilities to add approximately 96 MW of new capacity to the electric grid to serve Georgia Power customers, according to Georgia Power. That figure is modest by any grid-scale standard. What is not modest is what the structure signals: a hyperscaler is now willing to fund the physical modification of operating reactors to secure carbon-free baseload, and it is doing so in a way that shifts the cost away from other customers.

The filings include a new Nuclear Uprate (NU-1) tariff structure as well as a request to approve a new extended power uprate for Plant Hatch Units 1 and 2. As part of this transaction, Google will subscribe to the new NU-1 tariff and receive the Zero-Emission Credits, which represent the carbon-free attributes of nuclear energy, associated with the power generated by the nuclear unit uprates. Georgia Power says Google’s subscription is projected to enable about $900 million in benefits for customers over the life of the units. Regulators still have to approve the arrangement, but the basic architecture is clear: the technology sector is absorbing capital costs that would otherwise be socialized across the broader customer base.

The deal comes amid growing scrutiny over who should pay for additional power infrastructure needed to accommodate soaring energy demand, and the agreement with Google will help shield residential and other industrial customers from the cost of the upgrades, Georgia Power said. That framing is politically durable. It gives the Georgia PSC a reason to approve quickly, and it gives other utilities a template to replicate.

The template already has precedent. Constellation Energy’s 20-year, 835 MW power purchase agreement with Microsoft for the Crane Clean Energy Center restart and the 20-year agreement with Meta for the Clinton Clean Energy Center’s 1,121 MW output demonstrated that Big Tech would commit to nuclear at scale. The Georgia Power deal is a structural step further: rather than simply buying output from an existing or restarted plant, Google is funding the capacity increase itself. Hyperscalers ultimately want to become offtakers of nuclear, but some are willing to act as investors in the short term to kick-start the industry. This filing is evidence of that shift moving from theory to contract language.

For precious metals investors, the relevance is uranium. Every incremental megawatt committed to nuclear power by a creditworthy offtaker tightens an already strained fuel market. The long-term uranium price indicator reached $97.00 per pound in mid-2026, the highest level in more than 18 years, according to TradeTech data cited in recent company filings. More uranium mining and production are needed to meet growing demand, and the decades-long lead times required to bring new uranium projects into production mean supply deficits are expected to last well into the 2030s. Deals like the Georgia Power arrangement do not add meaningful near-term uranium demand on their own. What they do is lengthen the visible demand horizon, which is what drives long-term contracting prices and the capital decisions of producers.

Risks to Monitor

The Georgia PSC has not yet approved the NU-1 tariff or the Hatch uprate request, so execution risk is real. If regulators push back on the subscription structure or question how ZECs are allocated, the timeline stretches. Southern Company (SO), which operates the plants through its Southern Nuclear subsidiary, carries the reputational exposure if either uprate runs over budget, a risk Vogtle Units 3 and 4 made painfully familiar. On uranium, any meaningful reversal in AI capital spending or a faster-than-expected build-out of grid-scale renewables could soften demand expectations.

Bottom Line

Ninety-six megawatts is not the point. The point is that Google just agreed to fund reactor modifications, take the associated credits, and keep other customers whole, because reliable carbon-free power is worth more to a hyperscaler than the cost of the uprate work. Each deal of this kind shortens the distance between AI infrastructure demand and uranium mine supply. That is the floor the market keeps building under uranium prices, one corporate filing at a time.