Craig,
Depending on what tools you used to capture the schematic, and what models your flow supports, asking for a "verilog netlist" is a feature of some tools.
The resulting verilog netlist will be at the transistor and wire level, or gate level (which in the hierarchy is made up of gates and wires), and will basically allow you to functionally simulate the schematic, less any timing, or 'analog' behavior (after all, you only have 1's,
0's, 'don't cares', tri-states, and unknowns).
This is commonly done for ASIC/ASSP design: one may synthesize higher level RTL (like verilog and VHDL descriptions that have no specific technology or functional modules like 'multiply this by that') into lower level schematics of transistors (from standard cell library gates).
The resulting circuit netlist (like a spice netlist) will allow one to perform many good (analog) simulations, but will either take too long, or blow up, if there are close to a billion devices in the chip. The next choice is to simulate the verilog netlist of the wires and transistors, which will run much much faster, but will be unable to tell you anything about time, voltage, or current.
Since you posted on c.a.f. I am going to presume your HDL was synthesized for a FPGA, and then it was placed, and routed in the FPGA.
You may also have a schematic of how the HDL blocks are connected together. The resultant netlist is in some format (for Xilinx: XDL) which may then be simulated quickly for the functional behavior.
In an FPGA, the translation all the way down to transistors is not provided, like it would be in the ASIC/ASSP flow....
Make sense?
Austin