It is increasingly recognized that Alzheimer's disease (AD) exists before dementia is present and that shifts in amyloid beta occur long before clinical symptoms can be detected. Early detection of these molecular changes is a key aspect for the success of interventions aimed at slowing down rates of cognitive decline. Recent evidence indicates that of the two established methods for measuring amyloid, a decrease in cerebral spinal fluid (CSF) amyloid β 1−42 (Aβ 1−42 ) may be an earlier indicator of Alzheimer's disease risk than measures of amyloid obtained from Positron Emission Topography (PET). However, CSF collection is highly invasive and expensive. In contrast, blood collection is routinely performed, minimally invasive and cheap.In this work, we develop a blood-based signature that can provide a cheap and minimally invasive estimation of an individual's CSF amyloid status. We show that a Random Forest model derived from plasma analytes can accurately predict subjects as having abnormal (low) CSF Aβ 1−42 levels indicative of AD risk (0.84 AUC, 0.73 sensitivity, and 0.76 specificity). Post-hoc analysis indicates that only six analytes are required to achieve similar performance. Furthermore, we show across an independent validation cohort that individuals with predicted abnormal CSF Aβ 1−42 levels transitioned to an AD diagnosis over 120 months significantly faster than those predicted with normal CSF Aβ 1−42 levels. This is the first study to show that a plasma protein signature, together with age and APOE 4 genotype, is able to predict CSF Aβ 1−42 levels with high accuracy. Biomarkers in plasma have previously been shown to be predictive of PET amyloid levels. This work further highlights the potential for developing a blood-based signature for improved AD screening, critical for drug and intervention trials.