Ramez Naam warns land-use limits solar-and-battery power for data centers, responding to Conviction founder Sarah Guo
Story Overview
In a quick X exchange, Conviction founder Sarah Guo floated the idea of a public debate pitting solar-plus-batteries against nuclear for data-center power. Ramez Naam jumped in to argue that solar and storage could eventually beat other on-land options on cost in the right places, while flagging land-use limits as the main brake on scaling.
Energy matchups remain wide open
Guo’s post explicitly asks whether anyone has hosted such a comparison yet, and Naam confirms he has not seen one. No pricing data, acreage numbers, or regional case studies appear in the thread, leaving the relative economics still unsettled.
Land availability becomes the real gatekeeper
Naam singles out land requirements as the key obstacle for solar-plus-storage, even while noting the combination could prove cheaper than alternatives in suitable geographies. No further details on how much land or which regions are supplied.
Some users welcomed VC Sarah Guo's proposed debate on solar-battery versus nuclear power for data centers as much-needed and worth attending, while others dismissed it as pointless given known timelines and costs favoring one side.
No Digg Deeper questions have been answered for this story yet.
Most Activity
before someone @ me, yes I understand that there are cloudy days sometimes and they have different generation profiles and solar might need overbuild/diesel backup etc. that’s the point of the debate
has anyone hosted a debate on solar+batteries vs nuclear for datacenter power generation?
(both would be good, but curious)
@saranormous First pass via 100+ AlphaSense / Tegus expert calls
@saranormous Seems like a memo is a better format than debate.
The difference will come down to tolerances, like the number of consecutive cloudy days.
But I want @CJHandmer to steelman the solar side
@saranormous Haven't seen one. I'd be happy to make the case that solar + batteries make a lot of sense in some geographies and will ultimately be cheaper than any other on-land resource. Though land use is an obstacle.
has anyone hosted a debate on solar+batteries vs nuclear for datacenter power generation?
(both would be good, but curious)
@saranormous The debate might depend more on time horizon than technology.
The best long-term solution isn't always the best next-year solution.
Solar + batteries can scale much faster today.
Nuclear may be the better answer if you're planning infrastructure for the next 30–50 years.
@saranormous Progress conference 2024 we did one.
has anyone hosted a debate on solar+batteries vs nuclear for datacenter power generation?
(both would be good, but curious)
@saranormous I think the solar + batteries vs nuclear debate is far less interesting than the grid-connected vs islanded/shadow grid debate
@ShanuMathew93 @saranormous It’ll be a sprint, but I beleive it is possible for Valar to have a GW online by the end of 2028.
@edavidpeterson @saranormous Thanks. I hadn’t heard about it - will look into it.
@saranormous This is not an apples-to-apples debate. Solar and battery power plants can be built in 1-3 yrs. New nuclear is a minimum of 10 yrs to build. If you want a data center now, your choice is solar, wind, batteries, gas, or coal. Maybe geothermal? Or some combination of these.
@CharlieZvible @saranormous How did you create this on AlphaSense?
@rmcentush @saranormous The islanded grid debate is even more interesting than you realize. You might see hyperscalers going off-grid as an admission of failure, but I see it as precisely the way to fix the grid if we enact the right policy reforms
@stavenka @saranormous Really? Ask @elonmusk
Unless you’re talking about 7-10+ years out this isn’t a debate. Nuclear literally wouldn’t come online in time in nearly any major market that’s building out a ton of data centers ex-China. So for the most interesting period for DCs - next 3-5 years - one side literally won’t contribute from a greenfield site contribution basis.
So, depends on what you’re trying to legitimately ask. Nuclear only relevant for this debate if you’re talking about longer-term power needs (and need to entertain government support to backstop and get the projects built) which is a separate debate about power needs (DCs smaller contributor the further you get out). So if it’s the near term power scarcity argument, better off chasing different avenues.
@rmcentush @saranormous Second. Nobody serious thinks you can run a DC off of only one, diversity wins for reliability
@saranormous Much-needed. I’d love to attend if one is organized. Can you organize one?
@saranormous There is not much to debate. Limitations on solar and wind as they relate to utility deployment are well known for about a decade (I know, I used to write about it a decade back). You can ask any decent AI any question you want on the subject and you will get decent answers.
Both sides miss it!
Solar's ~5c, add batteries and you're at ~10c. New nuclear is 14-22c. The only cheap nuclear is the plants we already built.
Cheaper options are deploying at scale. @_panthalassa is going after wave power at 2c, no overseas supply chain. @antoraenergy has a solution too.
@saranormous Solar panels got 10x cheaper. That's not the bottleneck anymore.
Overbuild + batteries are. Most ROI models undercount battery costs.
For grid-scale, that's the math that actually matters.
@saranormous Would imagine cost, speed to power (+ grid connectivity), and reliability are some main points to cover. Would love to see this as well