The fact of London 2026 — three men inside Kelvin Kiptum's old world record on the same start line, two of them under two hours — was processed in real time on Sunday morning. The shape of it, the section-by-section data that explains how it actually came to pass, took rather longer to settle. Twenty-four hours on, with full intermediate splits now public from World Athletics and the AIMS-certified course markers, what looks at first like an outlier appears under analysis to be the predictable end-state of a six-year trajectory in which the marathon has been training its athletes, its courses and its shoe technology in roughly the same direction.
The splits themselves tell a clean story. The lead group of seven, paced by Mathew Kimeli and a rotating crew of three, came through 5 km in 14:18, 10 km in 28:37, 15 km in 42:46, 20 km in 56:55 and halfway in 1:00:29. None of those numbers are extraordinary in isolation; the world's best 5 km in any pacing-eligible block over the last decade is a 13:54 split run by Kiptum at Chicago 2023. The 1:00:29 first half, however, was the deepest sub-60-pace lead group ever assembled — six runners through halfway with all six on sub-2:01 trajectory. The decisive move came in the third 5 km of the second half, between 30 and 35 km, where Sawe and Kejelcha covered 13:54 to drop Kiplimo by exactly 21 seconds.
That 30-to-35 km block is where the race ended. Sawe's recorded splits for the second half — 14:14 (15-20), 14:09 (20-25), 14:06 (25-30), 13:54 (30-35), 13:58 (35-40) — describe an athlete running progressively but never panickingly faster, and the closing 2.195 km in 6:09 (4:31 per mile, slowing only as the last 200 m turned hard onto The Mall) is the single most impressive marathon closing-block in the published record. Kejelcha shadowed those splits within a second through 40 km, then ran a closing 2.195 km in 6:21 to lose 11 seconds and the win, but break two hours nonetheless. Kiplimo's 35-to-40 km of 14:06 was, in absolute terms, faster than anything in his half-marathon career; it simply could not match what had happened in the previous 5 km block.
The shoe story, which the trade press rushed to publish on Sunday afternoon, deserves a more cautious read. Sawe ran the new Adidas Adizero Adios Pro Evo 2; Kejelcha was in a final non-commercial Nike prototype that Nike's Oregon team have flagged as the basis of a 2027 release; Kiplimo wore the Puma Fast-R Nitro Elite 3 that Puma launched at Boston ten days earlier. All three shoes use a carbon-plate-plus-supercritical-foam construction within the World Athletics-mandated 40 mm stack height, and all three were on sale or available to wear under regulations on Sunday. The performance gap between Sawe and Kiplimo — 58 seconds — is not plausibly explained by any current footwear differential; it sits inside the noise of a fast late-race move on a flat course.
What the numbers actually tell us is that the marathon's training, racing and equipment ecosystems have converged on a set of capabilities that make sub-two consistent rather than unique. London produced three sub-2:01 runners on a single start line because three athletes arrived in London capable of running sub-2:01 on a controlled day. The implication for Berlin in September, where Sawe is now expected, is the obvious one: the 1:59:30 was a ceiling for one runner on one course on one cool morning, but it is a floor for what the sport will now expect of its top tier. Kejelcha will run a second marathon by year's end, Kiplimo's coach has hinted at a three-month-out Berlin or Chicago target, and at least four runners not in London on Sunday — Tola, Cheptegei, Korir Jr., Mantz — sit within reach of the same band on the right course.