←News detail

Cailian PressUpdated Nov 2 18:51

Google Energy Executive Says Space Data Centers Could Cost Up to 100 Times More Than Ground-Based Ones

AI summary

Amanda Peterson Corio, Google's head of data center energy, said at the WTF summit that the launch cost of sending computing hardware into orbit is extremely high, potentially up to 100 times the cost of ground-based construction, and noted that the real bottleneck lies in the approval process and lengthy evaluation studies required to connect projects to the grid. She mentioned that if launch costs fall to $200 per kilogram or below, the computing cost per unit of power in space could become roughly comparable to that on the ground, a threshold that may be reached around 2035. This month, Google's research team still announced a partnership with PlanetLabs to explore large-scale space data center clusters, and has already launched a small space server prototype carrying its self-developed TPU under the Suncatcher project.

Why read this

AI score 55

Google's energy chief gave a specific assessment that space data centers could cost up to 100 times more than ground-based ones, and noted that grid-connection approval is the real bottleneck.

Topics

Tags

Body · AI translation

Cailian Press Financial News Agency

Open the APP

Space computing power meets a dose of reality: Google energy chief says costs can reach 100 times those of ground data centers

Cailian Press, Zhou Ziyi; Editor: Lin Qi, 2026-11-02 17:16

① Google's head of data center energy said the launch cost of sending computing hardware into orbit is "extremely high, up to 100 times the cost of ground construction"; ② The real bottleneck lies in the approval process required to connect projects to the grid and the lengthy assessment studies; ③ If launch costs fall to $200 per kilogram or below, the computing cost per unit of power in space is expected to be roughly comparable to ground levels.

Cailian Press, November 2 (Editor: Zhou Ziyi) — Amanda Peterson Corio, Google's head of data center energy, said in the latest remarks that the cost of space data centers could be up to 100 times that of ground data centers. These remarks poured a dose of realistic cold water on the currently hot concept of space computing power.

Speaking at the WTF summit, she said, "Before rushing into space, we should first thoroughly explore all feasible options on the ground."

Her remarks came as SpaceX is planning its orbital data centers, one of the key rationales supporting its market value of about $1.8 trillion; Google itself is also testing AI server chips in orbit.

High costs

Peterson Corio mentioned that opposition to ground data centers is gradually rising in American society, and a large part of the responsibility lies with a group of newly established companies with limited track records, which should be subject to higher regulatory standards; by contrast, before this round of AI computing power boom, companies that had been deeply engaged in data center construction for many years "operated in a relatively standard manner," and their related projects rarely sparked controversy.

She pointed out that the launch cost of sending computing hardware into orbit is "extremely high, up to 100 times the cost of ground construction." In addition to cost, operators also need to solve the difficulties of maintaining and upgrading in-orbit equipment. On the ground, data center staff can replace servers and rewire and debug systems at any time.

Despite her cautious attitude toward space computing power, Google's research team still officially announced this month that it will cooperate with satellite imagery company PlanetLabs to explore the feasibility of large-scale space data center clusters using satellites.

Under a project framework called "Suncatcher," Google has already launched a small space server prototype into space, equipped with an AI tensor processing unit (TPU) custom-built jointly with Broadcom.

Google researchers involved in the "Suncatcher" program, together with the company's chief researcher James Manyika, previously published a paper stating that "if launch costs fall to $200 per kilogram or below, the computing cost per unit of power in space is expected to be roughly comparable to ground levels." This threshold is expected to be reached around 2035, though the timing depends on the development progress of SpaceX's Starship and the pace of increasing launch frequency.

Compared with Google, other companies are more optimistic about space data centers.

For example, SpaceX Chief Executive Officer Elon Musk said the company could build a space data center as soon as next year; Jeff Bezos, founder of Amazon and head of SpaceX competitor Blue Origin, predicted that within a decade there will be space data centers with power consumption reaching 1 gigawatt.

It should be noted that Peterson Corio did not completely reject the idea of space computing power. She called the direction "very interesting," and said the reason Google continues to lay out such high-risk frontier projects is that the company is accustomed to "thinking ahead about what will happen 10, 20, 30 years from now." But she also said that bringing this technology to fruition is "by no means imminent."

The real bottleneck

Regarding space data centers, many people point out that U.S. power supply is insufficient to support the expansion of AI computing power. Peterson Corio offered a different view.

She said, "Is there enough electricity? Of course there is enough. The real bottleneck lies in the approval process required to connect projects to the grid and the lengthy assessment studies."

Alphabet's laboratory launched the Tapestry project and is cooperating with PJM, the largest grid operator on the U.S. East Coast, to use AI to accelerate the relevant assessment processes.

In Texas, state power regulators have slowed approval of some data center projects and strengthened supervision to avoid new computing facilities adding burden to the grid and pushing up electricity prices.

Peterson Corio said she supports the industry introducing regulatory rules and setting higher standards. She also gave examples of how Google is increasing investment in local communities where projects are located, while setting up contingency plans to avoid the risks brought by unfinished projects.

Peterson Corio's remarks came against the backdrop of investors currently re-examining the valuations and viability of a group of emerging AI cloud vendors: most of these companies have received heavy backing from Nvidia and are rapidly launching data centers through heavy borrowing and without long-term operating records. The companies Peterson Corio referred to as needing tighter standards most likely include such vendors.

Global Market Intelligence

U.S. Stock Market Dynamics

Cailian Press statement: The content of this article is for reference only and does not constitute investment advice. Investors who act on it do so at their own risk.

Business Cooperation

Popular Unlocks

[Trend Research Report · Company] This cross-border leading brand's third-quarter revenue and profit hit a record high for the same period in history, with category expansion + channel upgrades, and revenue in multiple regions growing explosively; its net interest margin was the first among peers to stabilize, and two insurance funds entered together. This rural commercial bank's "one foundation, four arrows" opens up growth space

① This cross-border leading brand's third-quarter revenue and profit hit a record high for the same period in history, with category expansion + channel upgrades, and revenue in multiple regions growing explosively; ② Its net interest margin was the first among peers to stabilize, and two insurance funds entered together. This rural commercial bank's "one foundation, four arrows" opens up growth space.

[Intraday Treasure] As winter approaches, institutions say sales of related categories in this industry may improve. This company cooperates with relevant institutes and has a rich product line

As winter approaches, institutions say sales of related categories in this industry may improve. This company cooperates with relevant institutes and has a rich product line that can fully meet the needs of segmented fields.

[Telegram Interpretation] Sales of the 3A game "Phantom Blade Zero" exceeded 4 million within 72 hours after its release. Institutions say domestic 3A games are expected to replicate the successful path of mobile games going overseas. This company's invested enterprise provides motion capture technology for "Phantom Blade Zero"

Sales of the 3A game "Phantom Blade Zero" exceeded 4 million within 72 hours after its release. Institutions say domestic 3A games are expected to replicate the successful path of mobile games going overseas. This company's invested enterprise provides motion capture technology for "Phantom Blade Zero", and another company customizes an exclusive graphics mode for "Phantom Blade Zero".

Related Reading

A "spectator" timetable for the U.S. midterm elections: When is the final result expected to be released?

Cailian Press Liu Rui 2026-11-02 17:45

① November 3, U.S. Eastern Time, is the official voting day for the U.S. midterm elections, when all 435 seats in the U.S. House of Representatives and 35 seats in the Senate will be up for re-election; ② The control of the U.S. House of Representatives can likely be determined relatively early, but given several close Senate races, the Senate election results may be delayed by several days before they can be confirmed.

Beware of rising U.S.-Iran tensions after the midterm elections! Goldman Sachs sharply raises oil price forecasts: supply faces major risks

Cailian Press Bian Chun 2026-11-02 16:05

① Goldman Sachs Group recently raised its oil price forecasts for the end of this year and for 2027, citing elevated crude oil supply risks; ② The bank's analysts said the market is "increasingly pricing in the risk of escalation after the U.S. midterm elections on November 3, and a subsequent possible weakening of crude oil supply in the Middle East."

Is the "space AI era" coming soon? Musk makes a bold prediction: orbital computing power will crush ground-based computing power by 2029!

Cailian Press Huang Junzhi 2026-11-02 15:38

① Musk predicts that by 2029, the artificial intelligence and cloud computing processing power deployed in low Earth orbit will far exceed ground-based computing power; ② "Compared with orbital computing, ground-based computing power is simply negligible," he wrote.

Source

Cailian Press