HIV test-and-treat strategies depend on structural conditions, including proximity to testing. We evaluated associations between access measures with new HIV diagnoses and late presentations in Tennessee (TN). We used statewide surveillance data among adults diagnosed with HIV between 2018–2023 across TN’s 95 counties. Outcomes were annual county counts of new diagnoses and late presentations (AIDS-defining event or CD4<200 cells/µL ≤90 days after diagnosis). Exposures were population-weighted mean drive time to nearest test site and site density (effective sites per 100,000 residents down-weighted by distance-decay; γ=10 minutes), using sites from the CDC’s National Prevention Information Network. We constructed census tract population-weighted centroids, estimated drive times using an origin-destination matrix, and aggregated to county means. Spatial GAMs included year and a county adjacency smoother: log-link Poisson with population offsets for new diagnoses (IRR) and binomial-logit for late presentations (OR). County z-scores for model-based outcomes were averaged separately for drive-time and effective sites. In 2018–2023, 4,715 people were diagnosed; 18.69% (n=881) presented late. There were 234 sites, with a mean drive time of 14.75 minutes (range 3.15–31.23). Per 10-minute decrease in drive time, the rate of new diagnoses was 51.06% higher (IRR=1.51, 95%CI: 1.10–2.07) and the odds of late presentation were 26.56% lower (OR=0.734, 95%CI: 0.498–1.082). Each 5-site increase per 100,000 residents was associated with 5.14% higher rate of new diagnoses (IRR=1.05, 95%CI: 1.03–1.07) and 1.75% lower odds of late presentation (OR=0.982, 95%CI: 0.968–0.997). Priority scores identified high-priority clusters in northeast and south-central TN (Figure). Geographic access to testing sites is an important determinant of disease progression, and expanding access in high-priority areas could improve timely diagnosis. Findings may be affected by unmeasured service differences.