I don't think the cost of a new highspeed line from London to Bristol/Plymouth/Cardiff can be justified. Realistically you're not going to get the demand from Wales and the South West to fill three or four 400 metre trains an hour.
Off peak there is already demand to fill 2 trains an hour from Cardiff, 2 trains an hour from Bristol and at least 1 train per hour From Plymouth (all currently about 200m trains, but mostly soon to be 234m trains). That's at least 5 trains an hour now, with the potential growth in passengers between now and 2041 (remembering this likely to be about train services post 2033) and you'll be looking at needing double the capacity (assming 2.5% passenger growth).
It is therefore not unreasonable to run:
1tph CC (2*250m) - London, Reading, Exeter (split) and then onto Cornwall calling at most stations including Plymouth (Newton Abbot and Paignton)
1tph CC (250m) - London, Exeter, Plymouth
2tph CC (2*250m) - London, Heathrow, Bristol Parkway (split), Taunton, Exeter, Newton Abbot, Plymouth (Newport, Cardiff, Swansea)
2tph CC (2*250m) - London, Heathrow, Reading, Swindon(split), Gloucester, Cheltenham (Bristol)
2tph CC (250m) - London, Bristol
1tph CC (2*250m) - London, Reading, Bristol Parkway (split), Newport, Cardiff (Swansea)
Which is:
2x250m trains to Gloucester (four time the number of direct services) and Cheltenham (providing an all day direct service)
4x250m trains to Bristol Parkway and Wales (3 to Cardiff and 3 to Swansea) - a little over double now
4x250m trains to Bristol TM (Unless 400m platforms could be built at TM then 2 of the services could be 400m long) - a little over double now or a little over 2.5 the current provission
4x250m trains to Plymouth (which proberbly would be a little much, but given that is where HS3 would benefit the most it proberbly would see the largest amount of growth post opeing, but not all of the services would need to run from day one)