AI Data Centers Need Firm Power; Google's 50 MW Deal Is a Hedge

By
CTOL Staff Reporter
1 min read

As grid pressure grows, Google's 50 MW agreement secures an early position in firm geothermal power, while Quaise still has to prove it can deliver.

A 50 MW geothermal contract is small beside Google's clean energy portfolio, but it arrives as power hungry AI data center developments strain the electrical grid. OPB reported on 17 November 2026 that those developments had increasingly strained the grid, while federal power planners estimated the Northwest energy gap at about 11 gigawatts by 2032. In Oregon, data center companies need renewable energy to comply with state environmental regulations, increasing the urgency of new supply.

The power from Quaise's Project Obsidian is slated to support Google's existing data center operations in The Dalles, Oregon. Google pays for 100% of the power its data centers use, and its three Oregon campuses have achieved 83% carbon free energy. Lucia Tian, Google's head of advanced energy technologies, calls geothermal a “clean, firm” resource. Access to firm power gives a data center a planning advantage over one waiting for grid capacity.

Our position is constructive on superhot geothermal and cautious on Quaise. We would buy the category through an operator with operating megawatts, but pass on Quaise until Obsidian demonstrates output approaching its contract. For Google, the agreement is an early position in a possible source of firm electricity, not a supply line already delivering power.

From 2010 to 2025, Google signed more than 240 agreements to purchase nearly 35 GW of new clean energy, according to Quaise's announcement. At a rounded 35,000 MW base, 50 MW divided by 35,000 MW is 0.14%. For Google, you are looking at a strategic option on a new resource, not a substitute for its existing portfolio.

Quaise's first flow test is expected to demonstrate 10 to 15 MW of electricity production potential, against Google's 50 MW agreement. The Contract Capacity Gap is the distance between megawatts a buyer contracts and megawatts a project has demonstrated. A signed agreement secures a possible source; a successful flow test and connected plant turn it into supply. Purchases are due to begin as early as 2030.

Quaise says the first 50 MW will use conventional drilling; millimeter wave drilling is reserved for phase two. The Department of Energy selected Quaise for up to $25 million on 22 September 2026, with the full award subject to negotiations. Its three well configuration is designed to demonstrate more than 25 MW of gross electric power capacity, while phase one targets more than 50 MW by 2030. Gross design capacity, flow test potential and contracted power are different measures.

The strongest case for Quaise is that conventional drilling reduces technology risk in phase one, a buyer has signed before delivery, and federal support backs a three well design above 25 MW gross. Carlos Araque of Quaise says, “A data center just happens to be the first mover, the fastest mover that is able to take that risk.” A hyperscaler needs firm power, has a reason to learn and can absorb development risk. Yet the flow-test target covers only part of the 50 MW contract. The early position gives Google a chance to learn before the resource is mature, a benefit you should value separately from delivered electricity.

Fervo shows what successful execution can look like, and what capital it can require. Its 15 year, grid connected, take or pay Google agreement supplies 396 MW, with first power expected mid 2028. That is 7.9 times Obsidian's 50 MW commitment: 396 divided by 50 is 7.92. Fervo also committed to offering Google an option to expand by approximately 600 MW, taking the total to nearly 1 GW. Its contract is a large customer commitment, not proof that every contracted megawatt was operating.

The capital behind Fervo is a separate part of the precedent. Fervo had 36 MW in operation on 30 September 2026 and reported $1.9 billion in cash in Q3 2026 after its May 2026 public listing. The customer agreement creates a long term revenue opportunity; the listing and cash balance supplied financing capacity for construction. Quaise has $280 million in total funding as of 27 August 2026, a very different financial base.

Fervo's cost figures frame the financing question, but neither is a Quaise estimate. Its expected cost for GeoBlocks 1 through 3 is approximately $7,000 per kilowatt; its $5,500 per kilowatt figure for Cape Google 400 is an installed cost target. In a 50 MW scenario, 50,000 kW at $7,000 per kilowatt is $350 million, or $70 million above Quaise's $280 million in cumulative funding. At Fervo's $5,500 target, 50,000 kW is $275 million, or $5 million below that total. These scenarios describe benchmark sensitivity, not a demonstrated funding gap.

Alphabet's Q3 2026 property and equipment purchases were $51,355 million, and free cash flow was negative $9,065 million. The $350 million Fervo benchmark is 0.68% of quarterly capex. For Google, 50 MW is a modest portfolio hedge; for Quaise, capital is still required before it can sell electricity.

The grid connection is another execution problem. Quaise needs the 10 mile transmission line from Newberry Volcano to the nearest substation in La Pine. The federal entity owning about 75% of regional transmission built only one mile between 2020 and 2025. A successful reservoir has little value to a data center without a grid connection. Nearby Mazama Energy's Project Ceres has potential to generate 15 MW in 2027 after drilling more than 10,000 feet and reaching temperatures as high as 629 degrees Fahrenheit. Newberry's earlier EGS demonstration began in 2010 and achieved a three to four fold increase in injectivity, but did not become a power plant.

OPB reported that Google's agreement is Quaise's first commercial geothermal deal. When you value the agreement, distinguish future capacity from electricity available to run a data center today. The Contract Capacity Gap is why the agreement should be valued as an option on future firm power rather than supply already secured. Google's 50 MW is a hedge and learning position beside nearly 35 GW of clean energy agreements; Fervo's 396 MW contract and expansion option show how much larger a proven commercial commitment can become.

A public Project Obsidian update by the end of December 2029 reporting at least 50 MW of demonstrated electric output would prove us wrong. Until then, the strategic advantage belongs to the buyer securing an early position, while the company building the plant still has to deliver.

You May Also Like

This article is submitted by our user under the News Submission Rules and Guidelines. The cover photo is computer generated art for illustrative purposes only; not indicative of factual content. If you believe this article infringes upon copyright rights, please do not hesitate to report it by sending an email to us. Your vigilance and cooperation are invaluable in helping us maintain a respectful and legally compliant community.

Subscribe to our Newsletter

Get the latest in enterprise business and tech with exclusive peeks at our new offerings

We use cookies on our website to enable certain functions, to provide more relevant information to you and to optimize your experience on our website. Further information can be found in our Privacy Policy and our Terms of Service . Mandatory information can be found in the legal notice