Hello, and welcome to TechScape. I’m Blake Montgomery, writing to you from beneath a few inches of snow in New York City. This week in tech: an analysis of the quid pro quo of Donald Trump’s executive order on AI; OpenAI goes on the defensive; and data centers attempt to launch into orbit.

The tech industry got what it paid Trump for

Donald Trump’s executive order, which prohibits states from passing laws regulating AI and threatens punitive action like the withholding of federal broadband funding, is a win for tech industry leaders and lobbyists who have fervently campaigned against AI regulation. It’s also a sign that the tech industry’s embrace of Trump, which has included CEOs presenting him with gifts, attending dinners at Mar-a-Lago and donating to the construction of a new White House ballroom, may be paying dividends.

While tech leaders such as OpenAI’s Sam Altman previously talked about the existential dangers of AI, the industry has shifted messaging during the second Trump administration. Tech companies and their lobbyists now argue that their technology will spur unprecedented economic growth while simultaneously warning that regulation will gift China the lead in the AI arms race. Firms such as OpenAI, Microsoft and Nvidia have spent millions of dollars on lobbying efforts, and executives have advised Trump against letting state laws rein in their companies.

Despite bipartisan backlash to blocking state regulation and widespread public concern over AI’s threats to the power grid and copyright law, Silicon Valley’s efforts appear to have won out. While state officials have questioned the legality of the executive order, and it’s likely to be challenged in court, the message from the White House to AI companies is that growth is more important to the administration than safety.

OpenAI goes on the defensive against Google

Three years since OpenAI released a chatbot (built on Google’s research), Google has caught up. The success of the latest version of Gemini, coupled with the manic popularity of the image editor and animator Nano Banana, has put Google at the forefront of the AI race.

There are signs OpenAI is nervous. First, and most obvious, is that Sam Altman declared a “Code Red” to his employees soon after the November launch of Gemini 3. “We are at a critical time for ChatGPT,” he said.

There are other signs. OpenAI is running ads for ChatGPT during football games to recruit new users. It has purchased billboard space near Google’s large office in the Meatpacking District of New York City and along the highways outside San Francisco. Perhaps Times Square is next.

Altman himself appeared on Jimmy Fallon in an attempt to directly reach mainstream viewers of late night talk shows. He mostly talked about parenting, a far more relatable topic than AI model benchmarks.

Data centers, already blanketing earth, begin to colonize outer space

Tech giants are constructing data centers, the physical engines of artificial intelligence, that measure half a mile (about a kilometer in length) to power their products, but they’re thinking bigger. No expanse is too large for these ambitions, no frontier too far-flung, including the final one.

In early November, the startup Starcloud launched a satellite into orbit that carried a powerful Nvidia H100 chip. This month, the company announced the chip had been used to train a Google large language model, Gemma, during the satellite’s orbit and the model could now respond to queries from space. Other startups are jockeying for similar positions, including one company that says it will create a “galactic brain”.

Why would these companies take on the colossally complicated effort of launching AI’s innards into space? My colleague Robert Booth reports on Google’s outlook on the endeavor. The company revealed a new research initiative called Project Suncatcher in November that involves 2027 plans to launch its own chips into orbit aboard two satellites:

Google’s scientists and engineers believe tightly packed constellations of about 80 solar-powered satellites could be arranged in orbit about 400 miles above the Earth’s surface equipped with the powerful processors required to meet rising demand for AI.

Prices of space launches are falling so quickly that by the middle of the 2030s the running costs of a space-based datacentre could be comparable to one on Earth, according to Google research released in November. Using satellites could also minimise the impact on the land and water resources needed to cool existing datacentres.

One Google executive poured cold water on the sci-fi possibilities of orbital data centers, though. Travis Beals told the Wall Street Journal that replicating a one-gigawatt data center would necessitate deploying 10,000 satellites. There are roughly 11,700 active satellites in orbit today, according to a Harvard astronomer who has been tracking the number since 1989. Most of them – about 8,700 – belong to Elon Musk.

Jeff Bezos’s rocket company, Blue Origin, best known for almost ridding the world of Katy Perry, has been working for the better part of a year to launch data centers in space, which are essentially arrays of semiconductor chips aboard satellites powered by solar panels that face the sun. No need for water to cool the chips, as is the case with terrestrial data centers. Space is quite cold already. AI needs abundant and constant electricity, which earth is having difficulty supplying. Perpetual exposure to the sun poses one solution.

I thought at first Starcloud was owned by Elon Musk, creator of Starlink, but it is not. Musk has said Starlink will likewise “be doing data centers in space”, which is a goofy way to phrase things, but his company is also the single biggest operator of satellites in the world by a significant margin, so take what he says seriously. His company is planning to upgrade its existing satellite models to facilitate the creation of orbital data centers, per Journal. SpaceX is likely to conduct its initial public offering next year, according to reporting from last week, with a valuation somewhere between $800bn and $1tn, potentially the most valuable startup in history. Redirecting the gargantuan amount of money being spent on data centers on Earth could make for a new line of business for Musk’s aerospace company, a prospect that Musk’s lieutenants have cited in their sales pitches, according to the Journal.