The ITU has identified 5.8 GHz for space-based solar power. Every word of that
\nsentence traces to a real document, and the claim is false.
\nThe document is ITU-R Report SM.2392, "Applications of wireless power transmission via
\nradio frequency beam". A new edition appeared in June 2026, superseding the 2021 text.
\nIts Section 5 addresses spectrum for RF-beam wireless power transfer and names 915 MHz,
\n2.45 GHz, 5.8 GHz and 24 GHz as the bands the field uses. Section 2.9 is titled "Space
\nbased solar power via satellite". If you work on power beaming, this is a genuinely
\nuseful document, and until it existed much of the frequency discussion in this field
\ncited nothing at all.
\nNone of that makes it permission.
\nThe ITU-R publishes several kinds of text, and two of them are easy to confuse.
\nA Recommendation is a technical standard. Other instruments incorporate it by
\nreference. When a national regulator writes a licence condition that points at ITU-R
\nP.618, the propagation model in P.618 acquires legal force in that jurisdiction through
\nthe pointer. Recommendations are what the phrase "the ITU says" usually means when
\nsomebody says it and is right.
\nA Report is a description of a state of affairs. It surveys, it summarises, it
\nrecords what practitioners do. It recommends nothing and it binds nothing. SM.2392
\nbehaves like one throughout. It observes that 2.45 GHz is more desirable than the higher
\nbands because propagation impairments are smaller, which is a statement about physics
\nrather than about spectrum.
\nOne of these you can cite in a licence application. The other you can cite in a
\nliterature review.
\nTake the tempting inference at face value and follow it. The ITU has named 5.8 GHz among
\nthe bands used for RF power transfer. Space-based solar power appears in the same
\ndocument. Therefore 5.8 GHz is a candidate for space-to-Earth power beaming.
\nThe first two steps are accurate. The third does not follow, and the reason is in a
\ndifferent document entirely.
\n5.8 GHz sits in an ISM band: industrial, scientific and medical, shared, unlicensed for
\nmany uses, and busy. What matters for a satellite is not the ISM status
\nbut the Radio Regulations, and specifically Article 21, which sets power flux-density
\nlimits on what a space station may deliver to the Earth's surface in a given band. The
\nRadio Regulations carry no Article 21 space-to-Earth power flux-density row for 5.8 GHz.
\nThere is no limit written for it because the Regulations do not contemplate a space
\nstation transmitting at these power levels into that band at all.
\nSo the honest statement is that **5.8 GHz is not regulatorily available for
\nspace-to-Earth power beaming today.** The band is not restricted or subject to
\ncoordination; there is no framework in which to ask the question at all. A Report naming
\nit as the band the field uses on the ground describes engineering practice. It is not an
\nallocation and not a change to Article 21, and a regulatory assumption built on it rests
\non nothing.
\nThat sentence is uncomfortable if you are raising money on a 5.8 GHz architecture. It is
\nconsiderably more uncomfortable to discover it at the point where a regulator asks.
\nThere is a second reason to be careful with 5.8 GHz, and it is not regulatory at all.
\nThe Radio Regulations designate certain bands for industrial, scientific and medical
\napplications, and the arrangement comes with a condition attached: radiocommunication
\nservices operating in an ISM band must accept harmful interference from ISM equipment.
\nThe band is shared on the understanding that nobody in it is protected.
\nFor a microwave oven that is a reasonable bargain. For a power link it is a strange one.
\nA receiving site is being asked to accept interference from every other user of the band
\nwhile delivering a service somebody is paying for, and the whole commercial case for
\nbeamed power rests on the delivery being predictable enough to contract against. Choosing
\na band precisely because it is lightly regulated means choosing a band in which you have
\nno standing to complain.
\nThat is not an argument against 5.8 GHz on physics. Propagation at 5.8 GHz is workable
\nand the hardware is mature, which is exactly why the field converged on it for ground
\ndemonstrations. It is an argument that the band's convenience on a test range and its
\nsuitability for a contracted service are different questions, and the second one has
\nbarely been asked.
\nWe file the documents we cite in a public registry, with a column recording how far we
\nactually go with each one. The categories are deliberately unflattering to ourselves:
\nwhether a method is computed in our code, whether we parse data in that format, whether
\nwe merely use the vocabulary accurately without executing anything.
\nSM.2392 went in as vocabulary. We cite it for its terms, which are the ITU's own
\ndescription of the application, and we execute nothing from it, because there is nothing
\nin it to execute. Our Article 21 caveat stayed exactly where it was, because nothing in
\nthe new edition touched it.
\nThat is the whole of what a Report can do to a system like ours. It improved the
\ncitation behind a definition. It did not move a limit, enable a band, or alter a
\nconstraint.
\nThere is a second inference in this document worth ruling out, and it is subtler than
\nthe spectrum one.
\nSection 2.9 describes a geostationary system. Most published work on beamed power at
\nscale is geostationary, for the good reason that a satellite parked over one spot solves
\nthe intermittency problem that a low orbit creates. Ground-footprint figures in that
\nliterature follow from aperture and range, and at geostationary range the numbers are
\nlarge.
\nIf you fly in low Earth orbit, your range is two orders of magnitude shorter. Should one
\nof the areas stated in a geostationary study happen to sit near a number in your own
\nmodel, that agreement is a coincidence of scale across a fundamentally different link.
\nIt is not the ITU validating your case. Citing it as though it were is the kind of error
\nthat survives internal review precisely because the citation is real and the number
\nmatches.
\nTake a citation for what it covers, and no further.
\nA Recommendation. If ITU-R published one governing RF power transfer from a space
\nstation to the ground, or if a World Radiocommunication Conference agenda item opened
\nwork on a space-to-Earth allocation for power beaming, the regulatory position in this
\npiece would change materially and I would say so here.
\nA power flux-density row in Article 21 covering these bands for this use would do it
\ntoo, and would be the single clearest signal available that the question has moved from
\ndescription to permission.
\nShort of that, a national regulator granting an experimental licence for a space-to-Earth
\npower beaming demonstration would not change the international position, but it would be
\nthe first real evidence of how a regulator intends to reason about one — and that is
\nworth more than another survey.
\n