<p>Intracranial aneurysms (IAs) pose a significant risk of rupture, necessitating accurate hemodynamic assessment for effective clinical decision-making. This study explores the feasibility of using physics-informed neural networks (PINNs) to predict blood flow velocity, pressure distribution, and wall shear stress (WSS) in a simplified 3D sidewall saccular aneurysm model, without relying on conventional computational fluid dynamics (CFD) simulations. A PINN was developed to approximate velocity and pressure fields by enforcing the Navier–Stokes equations for steady-state conditions and boundary conditions within a computational domain. The model was trained using a collocation-based approach with 20,000 domain points and 2000 boundary points. A fully connected neural network with nine hidden layers was implemented, and training was conducted using the Adam optimizer with a learning rate of 0.001. Velocity, flow vectors, pressure, and WSS were predicted from both 3D geometry and 2D profile. Results demonstrated that the PINN model effectively captured physiologically relevant hemodynamic patterns, including velocity reduction within the aneurysm sac and elevated WSS at the aneurysm neck. Performance metrics showed an MSE of 5.57 for velocity and 3.32 for pressure, with an R-squared value of 0.66 and 0.60, respectively. This study presents an initial demonstration of PINNs as a CFD-free modeling framework for hemodynamic analysis in idealized aneurysm geometries. While not benchmarked against CFD, the results underscore the potential of PINNs for future patient-specific, image-based modeling pipelines, offering a physics-informed approach for cerebrovascular risk assessment and treatment planning.</p>