You've set up an either/or dichotomy, when an event like setting off a nuclear weapon can easily do many things at once. Consider North Korea's nuclear and missile tests today: they are, on one level, technical tests meant to give their scientists and engineers information; on another level, they are also motivated by very high-level international and domestic politics; at another level, they are also deployed and used as a form of posturing. There's no reason to try and meticulously figure out whether they fit into one category purely or another — the real world is often many things at once. (Scholars of things like nuclear weapons often like to use words like "technopolitics" to indicate that there is no clean boundary between the political and the technical.)

In the American case, nuclear testing did a lot of things at once. Individual tests and test series varied quite a lot along these lines. The Trinity test of 1945, for example, had no international display of military power, because it was kept a tight secret, and it was strictly about evaluating the feasibility of the new weapon. But the Crossroads tests of 1946 were quite different: the weapon feasibility was taken for granted (they didn't test anything new); from a technical standpoint, it was merely an "effects test" ("what happens if you nuke boats?"), but, at the same time, it was also a form of international posturing (UN observers were even invited, including Russians!), and a form of domestic interservice rivalry (the Navy was trying to prove its continued relevance in the atomic age).

In most cases, the tests had to have strong technical justifications. Nuclear weapons tests are expensive, often come with significant hazards (esp. in the era of atmospheric testing), required direct presidential approval, often required significant coordination with military forces, and so on. They were not something you could just do for the fun of it, and scientists are clever in the sense that if you give them the opportunity, they'll eagerly find a way to "advance the field of knowledge" one way or the other (even if it is just incrementally, as opposed to the experiments that test some new principle or theory, which are relatively rare, as there are only a handful of "new principles or theories" involved with nuclear weapon design). That being said, the act of testing nuclear weapons has, since the Cold War begun, also been seen as a very political one, one that sends signals both domestically and internationally about a state's strength and power, and so the approval of test series, and the setting off of weapons of very high yield (for example), was often undertaken with input from high-level political operators (e.g. Congressmen) or political agencies like the Department of State.

Even though the USSR was a very different sort of state than the USA, one sees similar (though not identical) dynamics in looking at their testing history. Again, the scientists' questions mostly drove the immediate planning of the tests, but political factors were always part of the discussion and the heads of state and the Politburo directly approved of test series. There are a few examples of the head of state actually dictating that a test would occur independent of scientific considerations, the most famous example being that of the famous "Tsar Bomba" test of 1961, which was announced by Khrushchev over a month before it happened, in an act of pure political posturing, given enough time for the UN to pass a resolution denouncing the testing. Was the Tsar Bomba, then, a purely political test? No — the Soviets did get a lot of data on "very high yield" nuclear testing, and did consider fielding such weapons on actual missiles. The US government's public response, that the bomb was "too big to use," was not exactly true, and the US government knew it. Private assessments by its own scientists and intelligence officers made clear that if one wanted to, one could find uses for 50-100 Mt weapons, maybe even higher, and that the Soviets did, as a result of such testing, have an "edge" in that department. Through the late 1960s, if the Soviets had violated the Partial Test Ban Treaty, the US had a program in place to rapidly test scaled-down versions of very-high-yield concepts in outer space, with an idea of "bridging that gap" but also, in a very political fashion, demonstrating the consequences of Soviet testing. Which is to say, once again, the technical and the political are inextricably wrapped up in these questions — it is not a matter of separating them out, but recognizing that they are complementary and feed one another.

I don't know the French and British situations as well but you can use this lens to analyze practically any nuclear testing. It is a common approach in the field of Science and Technology Studies (some might say that the technical facts and the political order are "co-produced," created at the same time, by such tests and their planning). It is a more useful way of thinking about it, in my mind, than trying to ask whether politics drove the science or whether science drove the politics — it's always a messy, hand-in-hand situation, though sometimes one hand is more obvious and prominent than the other. But when your "technical" act is setting off weapons with the power to kill cities, how can it not be "political" on many levels?

(Could a test be "purely" political, in the sense that it gives no technical information? Well, in a sense, maybe — I mean, you could always just not look for technical information, and in that sense, it is absent. But even a test of a totally routine, stockpile nuclear weapon — e.g., just picking a bomb and setting it off — gives some information, if only about reliability. This is not to suggest that some experimental setups do not yield more or better technical results than others, or that motivations cannot come from "the top", e.g. politics, as opposed to "the bottom", e.g. scientists, or vice versa. It is only to say that there are always many ways of viewing and using these kinds of events.)

Testing also served other ends not indicated here. In the US, for example, nuclear testing was very much part of the culture of the employment at nuclear weapons laboratories during the Cold War. You could not rise in the ranks of the lab unless your name was associated with several tests, and being responsible for a particularly successful test was the best intraorganizational currency you could have (John Nuckolls, at the age of 30 or so, secured his future laboratory career by having several audacious nuclear weapons designs successfully tested in the early 1960s; as he put it, one of the tests was of such a "radical" design that "my colleagues thought it was an April Fool's day joke"). So that is an entirely separate level from the purely technical or the high-level political.

On nuclear weapons testing culture at labs, see Gusterson, Nuclear Rites. On the interplay of technopolitical forces, MacKenzie, Inventing Accuracy, is great. On "co-production" in general, see Jasanoff, States of Knowledge. My comments on the "Tsar Bomba" come from my own research into the topic (much of the published literature on it buys the "public" argument that the US government put forward; my research has been on the "behind the scenes" views of it by the AEC, CIA, DOD, etc.). On the issues relating to "does technical information exist if you don't look for it?" see esp. Latour, Science in Action, and maybe also Hecht, Being Nuclear.