Zuckerberg says Meta will beam sunlight from space to power AI data centers

[John Clark]

Zuckerberg says Meta will beam sunlight from space to power AI data centers

John K Clark    See what's on my list at  Extropolis

mjj

[Keith Lofstrom]

On Thu, Apr 30, 2026 at 07:23:49AM -0400, John Clark wrote:
> *Zuckerberg says Meta will beam sunlight from space to power AI data
> centers*
> <https://www.tomshardware.com/tech-industry/artificial-intelligence/meta-will-beam-sunlight-from-space-to-power-ai-data-centers-solar-collecting-satellites-will-orbit-22-000-miles-above-earth-firm-reserves-1-gigawatt-of-orbital-solar-energy-and-100-gigawatt-hours-of-long-duration-storage>

Why not process the power into data before transmitting it?
Why not radiate the waste heat of computing into deep space
rather than the biosphere?

The vague press release mentions beaming the energy as infrared;
but WHICH infrared?

Infrared is defined as frequencies greater than 300 GHz (1mm
wavelength) to 380 THz (780 nm wavelength), a factor of more
than 1000; and all are harder to focus than visible light,
and more difficult to phase-lock than microwaves, difficult to
synthesize electronically, and very difficult to phase-lock
coherently across a large space structure.

I'd write more, but I have a dozen more useful ideas to
research and write about. Zuckerberg is a college dropout,
not a physicist or an electronic engineer.

Could someone with more time drill down into this and figure
out the scam this "news" is a cover story for? My guess is
Zuck's goal is to divert attention from less-impractical
ideas, such as the microwave SSPS we write about.

----

See http://server-sky.com for a better idea.

BTW, a software geek reading this can help me securely
upgrade my server and wiki websites to https. Zuck
was probably a better coder/sysadmin than I am.

Keith L.

--
Keith Lofstrom kei...@keithl.com

[k.a.carroll]

That article is inaccurate. Meta has contracted with Overview Energy to beam the power down to solar farms on Earth. Overview is the company Paul Jaffe works for; with him as their systems engineering guy, I'd say they know their stuff technically. Maybe Paul would be able to answer some questions about this; but I'm sure there are questions he'd not be allowed to answer.

--
You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/20260502010708.GA12666%40gate.kl-ic.com.

[Mark Sonter]

The wavelength is near-IR, and is going to be ca 1000nm. The partner is Overview Energy. The actual size / mass of the orbital facilities is unclear but the impression is these are to be launchable on F9.

In principle, NIR rather than 5 cm microwave overcomes the non-scaleable-down problem with the classical GEO SSPS concepts, viz need for km-size transmitters to get small-enough spot-on-ground.

Also, and critical importance, note that ordinary silicon solar panels do give good capture efficiency out to ca 1100nm. This means that Overview can beam down to presently-extant solar 'farms'...

I think it might have wings, depending on whether or not they can put large-enough assets in F9s, and whether the cost-to-orbit is low enough o close the bizcase....

Cheers,

mark

--
Mark J Sonter, mobile 0447 755 598 (delete '0' & replace with '61' country code if calling from overseas)

Director & Principal Consultant, Radiation Advice & Solutions Pty Ltd, abn 31 891 761 435
Principal, Asteroid Enterprises Pty Ltd, abn 53 008 115 302
Chairman, Off Earth Resources Pty Ltd, abn 31 664 219 372
Adjunct Assoc Prof, Andy Thomas Centre for Space Resources, Univ of Adelaide

116 Pennine Drive, South Maclean, Queensland 4280, Australia

Please use marks...@gmail.com for general or radiation related communications, and mark....@asteroidenterprises.com for space related topics.


“Keep everything as simple as possible, but no simpler” - A. Einstein

-----Original Message-----
From: power-satell...@googlegroups.com <power-satell...@googlegroups.com> On Behalf Of Keith Lofstrom
Sent: Saturday, 2 May 2026 11:07 AM
To: John Clark <johnk...@gmail.com>
Cc: Power Satellite Economics <power-satell...@googlegroups.com>

[John Clark]

On Fri, May 1, 2026 at 9:07 PM Keith Lofstrom <kei...@keithl.com> wrote:

> Why not process the power into data before transmitting it?
Why not radiate the waste heat of computing into deep space
rather than the biosphere?   

Those are very good questions! I think using the power that power satellites produce in space is the only way they make economic sense, it could be the first step towards making a Dyson sphere/swarm. If you have the technical ability to build a solar satellite capable of producing gigawatts of electrical power then you should also have the technical ability to build a space based data center that can use that power. Beaming all that energy back to Earth has always been the weakest part of the idea, and with data-processing there is no need to do so. 

John K Clark    See what's on my list at  Extropolis

ehj

  

 

[Paul Jaffe]

(answering in my personal capacity, not as an employee of Overview Energy)

> Why not process the power into data before transmitting it?

An intriguing possibility, and one that I've been thinking about since you introduced me to http://server-sky.com/, what, about 15 years ago?

But the fact is there will always be energy demand by people, on Earth (hopefully!) or at least for the foreseeable future. Though the energy demands of data centers are growing, and will continue, they are currently a tiny percentage of total energy and electricity demand. https://energy.ec.europa.eu/news/focus-data-centres-energy-hungry-challenge-2025-11-17_en

> Why not radiate the waste heat of computing into deep space rather than the biosphere?   

Certainly, this seems like an opportunity, but of course needs to be balanced in the assessment of a complete system.

> The vague press release mentions beaming the energy as infrared; but WHICH infrared?

Near infrared. https://www.overviewenergy.com/updates/space-to-grid  

> Could someone with more time drill down into this and figure out the scam this "news" is a cover story for?  

I have a conflict of interest here as an Overview Energy employee, so I'll leave this to someone else. Obviously, I don't think it's a scam. 

> My guess is Zuck's goal is to divert attention from less-impractical ideas, such as the microwave SSPS we write about.

I will point out that sending energy to existing solar farms vs. development of rectenna ground receivers overcomes the huge hurdles of ground receiver development, site identification, spectrum allocation, facility purchase/lease/build, and grid interconnection (currently the interconnect queue duration is on the order of five years). https://sustainabilitydialogue.uchicago.edu/news/how-the-interconnection-queue-backlog-is-slowing-energy-growth/

Note I am someone who has had the privilege of being involved with power beaming technology development across the spectrum, microwave through optical. Shameless book plug, currently under $48.65 for the paperback on Amazon: https://amzn.to/4t8iqze 

[John Clark]

On Sat, May 2, 2026 at 7:01 AM 'Paul Jaffe' via Power Satellite Economics <power-satell...@googlegroups.com> wrote:

> Though the energy demands of data centers are growing, and will continue, they are currently a tiny percentage of total energy and electricity demand.

In the USA during 2024 (the most recent figures I could find) data centers consumed 4.4% of the nation's electricity, I wouldn't call that "tiny" and it's certainly higher than that today in 2026.  Some estimate it could grow as high as 17% by 2030, and I see no reason why it will stop there.

John K Clark    See what's on my list at  Extropolis

rxi

[Bryan Zetlen]

You understand a Dyson sphere would be lethal to everything in it?

--
You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/CAJPayv3azN2QsKVMb0cbL-_OszMvy9LFYZuesRYPCWKbmbZZVQ%40mail.gmail.com.

[John Clark]

On Sat, May 2, 2026 at 11:58 AM Bryan Zetlen <br...@virtussolis.space> wrote:

> You understand a Dyson sphere would be lethal to everything in it?

Sure, but if a Dyson Sphere is ever built, and I think it will be, it won't be biological beings like humans who complete it, super intelligent computers will,  and those computers will be outside the sphere not inside it. The era of Humans making all the big decisions is rapidly coming to an end because biology can't compete with electronics. I'm not saying that is a good thing, I'm just saying that's the way things are.  

  John K Clark    See what's on my list at  Extropolis

_)(

[Opener of the Way]

If I had a thousand dollars for every thing Zuckerberg said that wasn't true I would have millions of dollars. 

--
You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/CAJPayv2_dRzBf5jnmL277WHc45pLx4cOdPeno2Zqom6TUa-s%3Dg%40mail.gmail.com.

[Jay Lewis]

Hi jeff, 

On Saturday, May 2, 2026 at 4:01:49 AM UTC-7 Paul Jaffe wrote:

> The vague press release mentions beaming the energy as infrared; but WHICH infrared?

Near infrared. https://www.overviewenergy.com/updates/space-to-grid  

> My guess is Zuck's goal is to divert attention from less-impractical ideas, such as the microwave SSPS we write about.

I will point out that sending energy to existing solar farms vs. development of rectenna ground receivers overcomes the huge hurdles of ground receiver development, site identification, spectrum allocation, facility purchase/lease/build, and grid interconnection (currently the interconnect queue duration is on the order of five years). https://sustainabilitydialogue.uchicago.edu/news/how-the-interconnection-queue-backlog-is-slowing-energy-growth/

Great solution!  Many solar plants are already in sunnier places, less clouds day and night.  What will it do with all that power at noon when the panels on earth are already near max?  

Jay 

[Keith Henson]

Not very many people have looked into the cost of a 5 GW rectenna. I did, and came up with a billion dollars. That's around one part in 12 of building a power satellite that makes competitive power. It turned out that you can't bury the cables; they get too hot, but overhead wiring does not interfere with microwaves.  

For what it is worth, the cost to GEO needs to be $200/kg. If you can get the cost to LEO down to  $100/kg, electric self-propulsion could get you to the target lift cost. It is relatively speculative, but a 50 MW engine exists (though it is not considered an engine). https://en.wikipedia.org/wiki/High-Enthalpy_Arc_Heated_Facility Talking to the people who run it (ten years ago) 70% of the energy put in comes out as Ke of the exhaust. This will get a 5 GW power satellite to GEO on a mass budget of 50,000 tons, including reaction mass.

Keith

Best wishes,

Keith

--
You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/52fae9f1-d243-49f0-98ef-0c58ac7b4e16n%40googlegroups.com.

[pja...@yahoo.com]

Send it to a solar farm where it isn’t noon. From geosynchronous orbit, there will always be solar farms visible that are at times of the day other than noon.

Sent from my iPhone

On Jun 8, 2026, at 7:57 PM, Jay Lewis <jle...@marginalx.com> wrote:



[Keith Henson]

Maybe ten years ago, I considered big lasers for propulsion. Even had some artwork done. 

https://htyp.org/design_to_cost#Space_laser

This one was powered from the ground. The big circle is a rectenna receiving 20 GW from the ground. The laser cooling in the illustration was before the design change to tapered tubes.

A multi-GW laser is remarkably useful. It will make short work of cleaning up all the space junk by vaporizing it. It is also a weapon. Focused to a few meters across, it will burn a squad of soldiers down to their belt buckles in a second and sink a tank into a lava pool in half a minute (on a clear day, of course).

That was a long time ago, but I don't think laser efficiency has improved a lot. The energy path is electrical to big laser diodes, and the light feeds fiber optic lasers. The laser light bounces off mirrors and heads for the ground.

Keith

[Jay Lewis]

That makes sense but requires aiming. RF can be aimed with phased array so can you do that with IR lasers too or will it be mechnical?  When its noon the GEOsat will be a cosine angle to the panals AND max distance.  

[pja...@yahoo.com]

The beam can be steered but I’ll hold off discussing technical details of the approach. The statement re: cosine angle is not necessarily true since many solar farms have single axis tracking.  The distance difference is largely negligible, though the path through the atmosphere is longer the farther you point off nadir.

Sent from my iPhone

On Jun 9, 2026, at 12:39 PM, Jay Lewis <jle...@marginalx.com> wrote:

That makes sense but requires aiming. RF can be aimed with phased array so can you do that with IR lasers too or will it be mechnical?  When its noon the GEOsat will be a cosine angle to the panals AND max distance.  

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/9166c45b-29ca-4101-9270-65cffcb2bbd3n%40googlegroups.com.

[Jay Lewis]

On Tuesday, June 9, 2026 at 3:51:41 PM UTC-7 pja...@yahoo.com wrote:

The beam can be steered but I’ll hold off discussing technical details of the approach. The statement re: cosine angle is not necessarily true since many solar farms have single axis tracking.  The distance difference is largely negligible, though the path through the atmosphere is longer the farther you point off nadir.

When I looked into tracking decades ago the extra costs to buy and maintain didn't pay back. Now almost half are tracking and over 80% of new plants track.  I guess the cost and performance of every aspect of the system has improved.  

With aiming you can also cheery pick whoever is sunny at a given moment, and also match demand and send it when needed.  One pattern for Phoenix is people come home from work near sunset and crank down the AC, so just as solar is fading demand starts spiking.  

 

[Christian Claudel]

As a near IR system, what is the max power density that can be safely sent (not blinding someone that would be looking up) though? Is it something around 10 W/m2 or less?

--

You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/e907ef45-caeb-47e2-9d44-5f5031936708n%40googlegroups.com.

[Keith Henson]

On Tue, Jun 9, 2026 at 7:40 PM Christian Claudel <christia...@gmail.com> wrote:

As a near IR system, what is the max power density that can be safely sent (not blinding someone that would be looking up) though? Is it something around 10 W/m2 or less?

Peak sunlight is a kW/m2. The original Dartmouth studies on birds set 230 W/m2 as a saft level or microwaves for  birds.

One problem we may find with microwave transmission is birds roosting on the rectenna to stay warm on cold nights.

On Tue, Jun 9, 2026 at 6:57 PM Jay Lewis <jle...@marginalx.com> wrote:

On Tuesday, June 9, 2026 at 3:51:41 PM UTC-7 pja...@yahoo.com wrote:

The beam can be steered but I’ll hold off discussing technical details of the approach. The statement re: cosine angle is not necessarily true since many solar farms have single axis tracking.  The distance difference is largely negligible, though the path through the atmosphere is longer the farther you point off nadir.

When I looked into tracking decades ago the extra costs to buy and maintain didn't pay back. Now almost half are tracking and over 80% of new plants track.  I guess the cost and performance of every aspect of the system has improved.  

With aiming you can also cheery pick whoever is sunny at a given moment, and also match demand and send it when needed.  One pattern for Phoenix is people come home from work near sunset and crank down the AC, so just as solar is fading demand starts spiking.  

 

--
You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/e907ef45-caeb-47e2-9d44-5f5031936708n%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/CAN980cmhh4BXowXsyNWffQO_A6hiwECthYOPL5T6Aro%2BTSVGsw%40mail.gmail.com.

[Paul Jaffe]

For microwave, IEEE C95.1 sets 10 W/m^2 as the limit for unrestricted areas and 50 W/m^2 for restricted areas. Both are a lot lower than 230 W/m^2.

See for yourself: https://standards.ieee.org/ieee/C95.1/4940/

Overview's system is "~350 W/m²" well within the IEC 62471 RG1 limit. IEC 62471 is unfortunately not free, but a summary is here: https://www.zgsm-china.com/blog/photobiological-safety-of-led-light-as-per-iec-62471.html

More info here:

https://www.overviewenergy.com/updates/about-the-beam

To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/CAPiwVB6-j_PVc6kHMMcRd%2BQSpBRzZb9XJF73%3Djo9VytV%2Bgztrw%40mail.gmail.com.

[Keith Henson]

POn Wed, Jun 10, 2026 at 8:56 PM Paul Jaffe <pja...@yahoo.com> wrote:

For microwave, IEEE C95.1 sets 10 W/m^2 as the limit for unrestricted areas and 50 W/m^2 for restricted areas. Both are a lot lower than 230 W/m^2.

See for yourself: https://standards.ieee.org/ieee/C95.1/4940/

That's true. I don't remember what it was rated in the long ago studies outside the rectenna fence. Not likely to matter since power satellites have so many other problems.

There is a funny story associated with the Dartmouth study. One very cold night someone noticed the birds (starling because nobody likes them) were missing. The birds were in a long term exposure test and them escaping would ruin the test. On entering the cage, looking for the birds, they found the lot of them had stuffed themselves in the microwave horn to keep warm. Which gives you another hazard for microwave power transmission, millions of birds roosting on the rectenna in the dead of winter soaking up a GW or so of microwave to keep warm. I only now thought about it, but you might have the same problem with birds on solar farms soaking up your IR beams.

Overview's system is "~350 W/m²" well within the IEC 62471 RG1 limit. IEC 62471 is unfortunately not free, but a summary is here: https://www.zgsm-china.com/blog/photobiological-safety-of-led-light-as-per-iec-62471.html

That is about 1/3 of noontime sunlight. Probably safe, though it would be uncomfortable on a hot day. You need a parasol, which is less  weird than an aluminum foil hat for microwaves.

Keith

[Geoffrey Landis]

Mark Sonter wrote:

>The wavelength is near-IR, and is going to be ca 1000nm. 

I expect the wavelength will be 1064 to 1070 nm, because that's what you get out of diode pumped fiber lasers (the last digit depends on matrix effects on the Y, I believe). Since they say they're beaming to existing solar arrays, that's about as long a wavelength as you can reasonably use; silicon's absorption constant gets pretty weak above that. (Even 1064 nm is somewhat weakly absorbed; you won't get peak efficiency at that wavelength. Would be nice to drop down another 100 nm or so, but 900 nm lasers with good efficiency have lower beam quality)

Existing solar arrays farms don't cover 100% of the ground area, so you're also going to take a loss from light not hitting the arrays.

christia...@gmail.com wrote:

>As a near IR system, what is the max power density that can be safely sent (not blinding someone that would be looking up) though? 

>Is it something around 10 W/m2 or less?

At 1064 nm, the eye-safe exposure limit I believe is 5 mW/cm2.  The wavelength has the one advantage that you can't see it, so people on the ground won't catch a glimpse of a glint in the sky and look up to see what it is.  Nevertheless, you're going to have to make sure that there is an exclusion zone around the receivers, and have a safety interlock to make sure the beam is cut (or defocused) if it's off target. That's doable. 

jle...@marginalx.com wrote:
>When I looked into tracking [for a terrestrial solar assay] decades ago the extra costs to buy and maintain didn't pay back. Now almost half 

>are tracking and over 80% of new plants track.  I guess the cost and performance of every aspect of the system has improved.  

Tracking's big advantage is that it flattens the power generation curve, so it doesn't have a big peak at noon and fall off in morning and evening. Still not uniform power, but closer.

--Geoff

[Jay Lewis]

jle...@marginalx.com wrote:
>When I looked into tracking [for a terrestrial solar assay] decades ago the extra costs to buy and maintain didn't pay back. Now almost half 

>are tracking and over 80% of new plants track.  I guess the cost and performance of every aspect of the system has improved.  

Tracking's big advantage is that it flattens the power generation curve, so it doesn't have a big peak at noon and fall off in morning and evening. Still not uniform power, but closer.

--Geoff

Its making the curve "flat" but not by sacrificing the center, its adding morning and late afternoon energy.  That extra energy has the cost of the tracking system power and repair consumed. Also more investment up front. Similarly, there is also a seasonal adjustment for everyone off the equator that would improve it even more, but thats a smaller benifit for another investment.  If the cost trend continues perhaps 2 axis tracking becomes the norm some day

Jay