To be clear:
“unenriched natural U238”
should be...
“unenriched uranium metal”
Which means, they did chemical/physical processes to extract the metal from the ore.
They did not boost the amount of U235 (called enrichment).
They just used graphite to moderate the neutrons from the very small amount of neutrons generated from the natural decay of U235 atoms in the metal.
RBMK graphite reactors can run on natural uranium but the Doppler coefficient with natural uranium makes them dangerously unstable as they have not only a positive coolant void coefficient they also have a positive power coefficient as well. This means that as the coolant voids and that positive feed back drives Keff farther past one the increase in fuel temperature effect on the neutron Doppler coefficient is also positive as fuel temp goes up so does the Keff in a never ending feed back loop until the fuel metal melts and attains a noncritical geometry. CANDU heavy water have a slightly positive void coefficient but a negative temperature Doppler coefficient even with natural uranium. The solution for RBMK was to use 1_2% enrichment which lowers the thermal Doppler coefficient to slightly negative but this means that the fuel must be enriched and also means that harvesting weapons grade PU is no longer an option since the burnup needed to justify enrichment would taint the the PU with 240. Put simply the RBMK was the cheapest reactor possible that could use no enrichment not heavy water enrichment while using graphite and light water without a pressure vessel or containment structure and have a burnup low enough to create copious amounts of weapons grade PU