Olivine is the most abundant and among the weakest phases in Earth's upper mantle, and thus, its rheological properties play a critical role in governing thermal structure and convective flow in the upper mantle. A persistent obstacle to constraining the in situ flow properties of olivine by laboratory experiment has been the difficulty in resolving the effect of pressure, which is weak within the 0‐ to ~2‐GPa pressure range of conventional laboratory deformation instruments but potentially strong over the 1‐ to ~14‐GPa range of the upper mantle. Using a deformation‐DIA, one of a new generation of bonafide deformation devices designed for operation to ≥10 GPa, we have deformed dry, polycrystalline San Carlos olivine in high‐temperature creep with the singular intent of providing the best achievable measurement of activation volume V* and a comprehensive statement of uncertainty. Under strictly dry conditions, at constant temperature (1,373 K) and strain rate (1 × 10−5 s−1), varying only pressure (1.8 to 8.8 GPa), we measure V* = 15 ± 5 cm3/mol. We have reproduced the well‐known mechanism change from [100]‐slip to [001]‐slip near 5 GPa and determined that, whatever the change in V* associated with the change in slip system, the effective value of 15 ± 5 cm3/mol is still accurate for modeling purposes in the 2‐ to 9‐GPa pressure range. This is a substantial pressure effect, which in the absence of a temperature gradient would represent a viscosity increase from the top to bottom of the upper mantle of 5 ± 2 orders of magnitude.