Checked my cycling power loss spreadsheet based on data and well-known formulas from Wilson & Schmidt, 2020, Bicycling Science, 4th ed.
By itself, aerodynamic power loss is proportional to drag area (aka CdA) and insensitive to weight and road surface properties. The book's Table 5.2 leads to representative CdA estimates of 0.64, 0.44, and 0.32 m² for upright commuting, touring, and fully tucked racing cases, resp. These figures take typical rider postures and bike aerodynamics into account. The touring and racing cases assume drop bars.
At 28 mph in still air, the corresponding aerodynamic power losses alone come to 777W, 534W, and 388W, resp. Big differences.
Now let's look at the SL 2 with a not-quite-upright posture. If we estimate CdA at 0.54 m² (average of the commuter and touring cases), then the aerodynamic loss at 28 mph is 656W.
A full 320W contribution from the SL 2's motor leaves the rider to put in 336W just to overcome air resistance at 28 mph in still air. At my weight (87 kg) on my SL 5.0 on smooth flat pavement, there'd be another 60W or so of rolling loss to overcome — for a total of 396W just from the rider. That's pro-level riding.
Roughly speaking, max total (rider+motor) torque limits acceleration and climbing rates, while max total mechanical power limits steady top speed. So increasing torque alone won't get you a high cruising speed. To cruise at 28 mph on smooth flat pavement in still air, a mere mortal would need an ebike motor with a lot more mechanical power than the SL 2's.
These estimates back up Stefan's point about realistic top speeds.