No third XB-70 was planned or built. The Valkyrie was diverted from weapons system to research program on December 3, 1959, when the Eisenhower Administration cancelled the B-70 bomber. Two XB-70s were completed as high-speed, high-altitude test planes, in part to support the growing SST project. Thus on April 14, 1967, as Boeing prepared to cut metal in Seattle, SST Program Director General J. Maxwell commented that "this is
the advanced aeronautical program in the United States. This is the
only one."
http://www.labiker.org/xb70.html has lots of great information on the XB-70.
North American, Boeing, and Lockheed all entered the Phase I competition of the US/SST program, which required designs to be submitted in early 1964. North American's was quickly dropped, because the conflicting requirements of a military-spec'ed strategic bomber and an efficient civilian airliner are too different to bridge with one design. Oddly enough, during the Advanced Manned Strategic Aircraft program, the Air Force seriously considered using modified Boeing 2707 airframes as XB-70 replacements (I have a picture of a B2707 bomber refueling in midair from a
KC-5, which also never materialized). The same objections were raised and the idea faded away, although SST purchases as light freighters, troop transports, and medevac aircraft were seen as likely. By the way, the AMSA eventually because the B-1A.
I don't think there's any doubt that Boeing underestimated the engineering challenges of an SST generally (though this was true of all four firms that tried to build one) and a swing-wing design in particular, but the Achilles' heel that initially surfaced on the 2707-100 design was stability, not weight. After moving the engines and support systems from under the 733's wing box to under the 2707's tail, Boeing found that it had to enormously increase elevator and elevon surface areas. This was to compensate for the decreased moment arm caused by the center of gravity's rearward movement. Much of the 2707-100's giant fixed tail comprised control surfaces of one sort or another.
When a final round of wind tunnel tests and computer analyses came in after Boeing won the FAA contract, it was found that extreme flight conditions (e.g. a microburst) requiring maximum upward pitch in a very short time frame could cause structural failure of the fuselage - on such a long aircraft (308 feet) the downward load on the tail induced unacceptable bending of the forward fuselage. Subsequent stiffening and addition of canards to balance pitch loading sent the 2707-200's weight spiraling upward to between 750,000 and 800,000 pounds. For inexplicable reasons, the FAA refused to allow MTOW to exceed 675,000 pounds, so as GDB mentioned, the payload was squeezed to virtually nothing. In a desperate attempt to get the weight down, Boeing had no choice but to drop the swing wing mechanism.
The final fixed double-delta 2707-300 was an extremely poor performer. In order to clear airport noise limits with the decreased lift/drag ratio of a delta vs. a variable-sweep wing at low speed, the -300's main (aft) delta was swept at only 50 degrees and was comparatively much larger than Concorde's or the L-2000's. The large area and low sweep cost it severely in supersonic cruise: the 2707-100's L/D at Mach 2.7 was 8.5; Concorde's at Mach 2 is about 7.5; and the 2707-300's would have been a paltry 7.3. Subsonic L/D also fell substantially. In order to compensate for lower L/Ds, resolve flutter problems on the lightly-loaded high-span delta, and keep the payload around 250 seats, maximum gross weight passed 800,000 pounds in early 1971 and was headed higher - far higher than the -200 but with lower performance. The 2707-300 would not have been a successful aircraft in scheduled service.
2707-100:
2707-300 (not to scale with above; the -300 was 8 feet shorter than the -100):
--B2707SST
Keynes is dead and we are living in his long run.