The foundations of large skyscrapers are not just simple rebar reinforced concrete floating on soil or atop flotation footings like your house or commercial office/school campus buildings. Because the building has both some significant dynamic wind loads to resist as well as a high center of pressure (c.p.) and center of gravity (c.g.), the foundation must be capable of resisting those loads internally and then transmitting them to the bedrock (all true skyscrapers mount to some kind of bedrock base). The foundation basically consists of a reinforced metal truss structure that looks somewhat like a multi-level truss bridge that is attached to pilings driving down to bedrock. The foundation itself is typically sealed from water intrusion specifically to prevent the kind of phenomena questioned by the o.p., for if a large mass of water were to freeze inside of or below the foundation it could very easily deform and possibly even lift the foundation.
If you were to intentionally pump and freeze water under the foundation, you’d probably topple the building as one side would start to freeze and push the liquid water to the other side. However, if you could control the freezing rate to keep it homogeneous and get it significantly below 0 deg C as described by Crafter_Man and Chief Pedant, you could certainly lift a building. This is, after all the same phenomena that during ice ages lifts and carries large slabs of rock significant distances.
I once came up with [POST=9514841]a concept[/POST] for a large, easily constructed, expandable, and relatively fail-safe rotating habitat using short and long fiber-reinforced ice as the hull material. The inherent energy in the latent heat of fusion makes the structure both stable (if protected against radiation incidence by a thin film of foil) and resistant to puncture by impact.