Supernova 1987A offers a unique opportunity to study an evolving supernova in unprecedented detail over several decades. The X-ray emission is dominated by interactions between the ejecta and the circumstellar medium, primarily the equatorial ring (ER). We analyze 3.3.Ms of NuSTAR data obtained between 2012 and 2020, and two decades of XMM-Newton data. Since similar to 2013, the flux below 2.keV has declined, the 3-8.keV flux has increased but has started to flatten, and the emission above 10.keV has remained nearly constant. The spectra are well described by a model with three thermal shock components. Two components at 0.3 and 0.9.keV are associated with dense clumps in the ER, and a 4.keV component may be a combination of emission from diffuse gas in the ER and the surrounding low-density H II region. We disfavor models that involve nonthermal X-ray emission and place constraints on nonthermal components, but cannot firmly exclude an underlying power law. Radioactive lines show a Ti-44 redshift of 670(+520) (-380) km s(-1), Ti-44 mass of ' 1.73(-0.29)(+0.27) 10(-4) M-circle dot, and Fe-55Y mass of <4.2 10(-4) M-circle dot. The 35-65.keV luminosity limit on the compact object is 2 ' 1034.erg.s-1, and < 15% of the 10-20.keV flux is pulsed. Considering previous limits, we conclude that there are currently no indications of a compact object, aside from a possible hint of dust heated by a neutron star in recent ALMA images.