Uranium as mined cannot just be put into a reactor; it must be processed first. Depending on the type of reactor, that can just mean milling and refining it into pellets (which South Africa was capable of doing on its own), or it can mean enriching the uranium (which South Africa was, in the 1950s, not capable of doing). In the case of the South African SAFARI-I reactor in question, it ran on highly-enriched uranium (HEU), which requires extensive enrichment facilities.

Part of the purpose of Atoms for Peace was to help non-nuclear weapons countries get some of the benefits from nuclear technology but not help them become nuclear weapons states. The idea in a case like this was, better to have the South Africans using an HEU reactor with American-supplied HEU fuel than to have them either learn how to use a reactor that doesn't need HEU (which could then later be diverted to military purposes with their existing uranium supply), or to keep them from developing a full uranium enrichment cycle (which could also be put to dual-use purposes — the technology to enrich uranium for a reactor can also be used to enrich it for weapons). So the goal was to give them exactly the kind of technology that would leave them somewhat dependent on the USA, and also prevent their need for them to develop dual-use technology. HEU research reactors fit this bill very nicely.

In the end, South Africa did develop its own uranium enrichment cycle, in secret (using their unusual, natively-produced method of aerodynamic enrichment), and did develop a small nuclear stockpile. Whether Atoms for Peace impeded or helped them and other proliferating states in their nuclear ambitions is something scholars still debate.