Reflections on the White House "Golden Age of Science" interview, and what its comment section tells us that the interview itself does not. See https://www.youtube.com/watch?v=iHv-vNM_XtU
Last month, Peter Diamandis sat down at the White House with Michael Kratsios, the President's science adviser and architect of the US AI Action Plan, the Genesis Mission, and the "Science and the New Golden Age" report. For anyone working in European research policy, it is worth watching twice: once for what was said, and once for how a quarter of a million viewers who viewed it and the 250+ who left a comment.
The American pitch
The vision presented is genuinely ambitious. Humans back on the Moon by 2028. A nuclear reactor in space. A scientifically relevant quantum computer within the presidential term. Five-year research grants funded on day one. Grant reviews in under a month. "Golden tickets" that let a single reviewer fund an unconventional idea against the majority. Autonomous laboratories where AI proposes the hypothesis, robots run the experiment, and the model designs the next one, around the clock, with no human in the loop. And a stated goal of not merely doubling but eventually 10x-ing the productivity of the entire scientific enterprise.
Kratsios also took a swipe at the Europeans, noting that the EU AI Act was finalized before ChatGPT even existed. As is often the case with his political master, this statement does not survice a simple fact check: ChatGPT was launched on 30 November 2022, the EU AI Act was approved by Parliament on 13 March 2024. There is almost 16 months between. The positive take is that the EU produced a substantial bit of legislation in record time.
The comment section is the real briefing document
Here is what struck me most. Once you filter out the noise, the most substantive and most upvoted comments were not about the Moon, quantum computers, or the singularity. They were about data centers. The explosive growth of investment in data centers has not happened (yet) in Europe.
The single most liked comment proposed that data center developers should offer major financial incentives to the communities that host them, "It's all about the money." Others described life next to a facility with hundreds of megawatts of gas and diesel generation, constant noise, all-night floodlighting, and rising electricity bills. One commenter carefully dismantled the claim that golf courses use thirty times more water than data centers, pointing out that golf courses largely use reclaimed water while data centers compete with households for potable supply. Another asked the question the interview never really answered: who actually captures the benefit from these breakthroughs?
The interview diagnosed AI's unpopularity as "a massive PR problem." The comments suggest something different. People in the USA were not asking for better storytelling. They were asking for benefit-sharing, accountability, and honest accounting of local costs. That is not a perception gap. That is a legitimacy gap.
Why this reads differently in Europe
And this is where the European perspective diverges sharply from the American one.
When Kratsios speaks of lunar bases and reactors in space as the missions that will inspire a generation, I suspect most Europeans feel a certain distance. Our anxieties are closer to the ground. After the hike in oil prices after Russia's invastion of the Ukraine in 2022 and after Israel's and the US's bombing campaigns of Iran this year, have lived through energy crises triggered by geopolitical supply shocks. European countries are more dependent on imports than the USA, which is a net exported. Households from Rotterdam to Riga still remember what happened to their bills when pipeline politics turned against them. The idea of adding gigawatt-scale AI demand to grids that are already strained, in a continent that imports much of its energy, lands very differently than it does in Texas.
And while the interview treated climate science mostly by omission (several commenters noted the defunding of NOAA programs and the absence of climate from the national missions), Europe does not have the luxury of omission. Our glaciers are measurably retreating. Floods, droughts, and heat waves are no longer projections in a report; they are line items in national budgets and insurance premiums. For a European audience, a "golden age of science" that does not include climate resilience and affordable clean energy is not a golden age at all. It is someone else's priority list.
This is not an argument against ambition. It is an argument about which moonshots matter. A European Genesis Mission would look different: AI-ready climate and earth-observation data, autonomous labs for better medicines, and better materials for battery chemistry and grid materials, fusion and next-generation geothermal, and compute infrastructure that comes with binding commitments on energy sourcing and community benefit from day one.
What Europe should actually take from this
Three things, in my view.
First, steal the boring parts. Fast-track grants, day-one funding, golden tickets, and institutionalized metascience units inside funding agencies are not ideological. They address problems Europe has in worse form than the US, where researchers lose close to half their time to administration. Much of the intellectual groundwork for these instruments is European anyway. We should implement them before Washington does.
Second, treat research data and lab automation as strategic infrastructure. The most consequential idea in the entire interview is making seventy years of US national laboratory data AI-ready and treating it as a public good. Europe's research data is richer than we admit and more fragmented than we can afford. If autonomous, AI-driven laboratories become the engine of discovery, the question of who owns and operates that infrastructure is a sovereignty question, not a procurement detail.
Third, learn from the comment section, not just the podium. The social license for AI infrastructure is earned with lower bills, quieter neighborhoods, and visible local benefit, not with optimistic narratives. The US administration has proposed a "rate payer protection pledge" requiring data centers to bring their own power. Europe should go further and make community benefit a condition of connection, before resistance hardens the way it visibly has in that comment thread.
The interview ends with a promise to aim for 10x. Fine. But speed without legitimacy is how you end up with 71 percent of your population opposing the very infrastructure your strategy depends on.
Europe's advantage has never been moving fastest. It has been, at our best, moving with consent. In the age of AI-accelerated science, that might turn out to be the scarcer resource.
#ResearchPolicy #AI #ScienceInnovation #Europe #EnergyTransition #ClimateAction #DataCenters #HorizonEurope
