Abstract The Transiting Exoplanet Survey Satellite (TESS) has discovered ∼5000 planets and planet candidates after 3.5 yr. With a planned second Extended Mission (EM2) spanning Years 5–7 on the horizon, now is the time to revise predictions of the TESS exoplanet yield. We present simulations of the number of detectable planets around 9.4 million AFGKM stars in the TESS Candidate Target List v8.01 through 7 yr of observations. Our simulations take advantage of improved models for the photometric performance, temporal window functions, and transit detection probability. We estimate that 4719 ± 334 planets should be detectable with the Prime Mission alone (Years 1–2), and another 3707 ± 209 should be detectable across the current Extended Mission (Years 3–4). Based on a proposed pointing scenario for EM2, we predict that TESS should find another 4093 ± 180 planets, bringing the total TESS yield to 12,519 ± 678. We provide our predicted yields as functions of host star spectral type, planet radius, orbital period, follow-up feasibility, and location relative to the habitable zone. As TESS continues, new planets will be progressively smaller, with longer orbital periods, and will orbit fainter stars. Half of the planets found in EM2 will be smaller than 4 R ⊕ , and over 1200 will have orbital periods longer than 20 days, effectively doubling the TESS yields of both kinds of planets. The number of small (<2 R ⊕ ) habitable-zone planets will also double, bringing the total TESS yield to 18 ± 5. We also compare our predictions to the actual Prime Mission yield, finding good agreement.