Robotic milking systems provide the capability of offering different amounts of supplemental feed to cows based on parity, DIM, and milk yield, but settings often do not take milk yield into account until 20 to 70 DIM, past the peak risk period for developing subclinical ketosis (SCK). To determine associations between SCK, milk yield, and supplement intake in robotic milking systems, we monitored 607 cows from 9 robotic herds, testing blood BHB 1x/wk for the first 3 wk of lactation. Positive cases of SCK (incidence = 33%) were defined by BHB ≥ 1.2 mmol/L at ≥ 1 of 3 tests. Milk yield and supplement intake data were obtained. In mixed linear regression models, we analyzed parity and week of lactation (wk 1 = 1–7 DIM, wk 2 = 8–14 DIM, wk 3 = 15–21 DIM) separately due to 3-way interactions with SCK status. The only differences in supplement intake between SCK and healthy cows was for for 3+ lactation cows in wk 3 (supplement consumption was 0.1 kg/d lower with SCK; P = 0.04). Milk production and milk yield per kg supplement were greater for SCK cows, particularly for primiparous cows in wk 1–2, and for 2nd and 3+ lactation cows in wk 1. From 0 to 21 DIM, in linear regression models, milk yield and milk/supplement were positively associated (P < 0.01) with BHB, but supplement intake was not (P = 0.9). The difference in milk yield and milk/supplement between 7-d forward- and backward-moving averages from the day of blood sampling were negatively associated with BHB (P < 0.01), showing that higher BHB values were associated with slower milk inclines. Accounting for parity in logistic regression models, smaller differences between the 2 moving averages for milk yield and milk/supplement were associated with greater risk of having BHB ≥ 1.2 mmol/L (P < 0.01), such that reducing the difference in milk yield by 4 kg and the difference in milk/supplement by 1 unit were associated with 1.35 (95% CI = 1.21–1.50) and 1.44 (CI = 1.29–1.61) times higher odds, respectively. These results highlight the differences in milk production (per day and relative to supplement consumed) associated with SCK and the potential for feed tables in robotic milking systems to reduce negative energy balance by accounting for milk production of fresh cows.