I saw this map on reddit:
And noticed that most of these Combined Statistical Areas are richly connected with infrastructure / high-capacity transportation links, or soon will be - most notably: Acela in the northeast, Brightline / Virgin Trains USA in Florida, Texas Central in Texas, and California High Speed Rail in California.
As a lover of cities, and a believer in agglomeration effects, I set out to analyze the CSAs on a connected, aggregated basis, as meta-cities. For reference, because I extend beyond the top 15, here's a more complete but less clear map of the top 30:
As you'll see, there are limited and vastly different prospects for each region, and I expect this to inform an indication of future development.
Note I refer to each CSA by its population rank, in an intentional attempt to separate these narratives from your existing ideas about these cities. I address each aggregate roughly by combined population, and the population numbers below are necessarily imprecise as they engage at the level of the CSA, so do not consider other populations not captured by the CSAs. Finally, when I list a combined population together with a rank, ala #1, the rank refers to the rank of that aggregation relative to the existing CSAs. That is, if two cities combined, how would they compare to the largest existing US cities.
Northeast
Unfortunately very little improvement has occurred in the Northeast corridor since electrification was completed in the year 2000, but they have an ongoing plan for improved service up to an average speed of 91 mph by 2040.[1] The government has arguably seriously under-invested in this corridor, given its import, returns, and potential for growth in speed and passengers, but there are promising signs that Amtrak may be taking this route more seriously going forward.
California
If/when the California High-Speed Rail project is completed, it would connect #2, 5 and 45 (combined population 30 million, #1) with rail speeds averaging ~143 mph (2 hours 40 minutes). The project is currently in question, with no certain completion date for the crucial LA - Bay Area connection, which will drive the vast majority of the ridership.
-LA + Vegas
The 2018 acquisition announcement indicated construction would begin in 2019 with initial service in 2022. The progress since then is unclear, and according to Wikipedia: "Although Los Angeles County finished an environmental assessment for the project in 2016, the exact date the project will start is unclear; however, San Bernardino County is still moving through the process."
-Capitol Corridor
Great Lakes
The midwest, centered around Chicago has a strange arrangement relative to the other possibilities. Chicago's population dwarfs the others, which perhaps explains why they have not seen cause to invest in high-speed rail improvements to connect the others.
#3 and 33 (combined population 12 million) are linked by Amtrak's Hiawatha at an average of 57 mph (90 minutes).
#3, 16, and 33 (combined population 16 million) are linked by Amtrak's Empire builder at an average speed of 50 mph (8 hours).
There's a proposal to increase frequency but not speed on this route, but for this to qualify as a practical megalopolis, you'd have to run the trains much faster.
#3 and 12 (combined population 15 million) are linked by Amtrak's Wolverine at an average of 48 mph (5 hours 20 minutes)
#3 and 20 (combined population 13 million) are linked by Amtrak's Lincoln Service at an average of 52 mph (5 hours 25 minutes)
Speed improvements seem to be in the works, if stuck in a bureaucratic quagmire. Local leaders would like the Governor to take the lead on this.
Virgin's Hyperloop One has identified #3, 24 and 26 (combined population 15 million) as a winning route for Hyperloop development.
Texas
Texas Central is an effort to build a world-class high-speed rail link at 186 mph (1 hour 30 minutes). Construction is planned to be completed by 2024. Future expansion to cover the so-called Texas Triangle is also planned.
-The Triangle
Virgin's Hyperloop One has identified #7, 9, and 25 (combined population 18 million) as winning route for Hyperloop development. They are engaged in a feasibility study of the route and have initial engagement with the Texas government.
I think it's remarkable that as a combined megalopolis, Texas would compete directly with Los Angeles and New York City on the level of economic network effects.
Florida
Virgin's Hyperloop One has identified this pair as a winning route for Hyperloop development.
Cascadia
#14 and 19 (combined population 8 million) are served by the Amtrak Cascades route, at an average speed of 50 mph (3 hours 30 minutes)
These also naturally associate with nearby Vancouver, Canada (combined population to 10.5 million, or #3). But the Amtrak Cascades route that serves all 3 does it at an average speed of 40 mph.
Improvement has been incremental and slow, in part due to set-backs like the 2017 derailment due to "gross complacency" that significantly delayed adoption of an improved route pending positive train control.
Deep South
Apparently Virgin Trains has identified this route as interesting, but it is not expected to be acted on in the near term.
Missouri
#20 and 27 (combined population 5 million). They are linked by Amtrak's Missouri River Runner at 50 mph (5 hours, 40 minutes).
Virgin's Hyperloop One has identified has completed a feasibility study (video) of the route.
Analysis
A few things jump out at me here:
Few Potential Megalopoli
Many urban areas are relatively isolated which means travel time and/or infrastructure cost would weigh against their incorporation with nearby neighbors. Examples: Salt Lake City, Denver, Phoenix. After Cascadia, the benefits of aggregation drop of substantially, and according to my own intuition, tend toward depending on broader integration to justify themselves. Similarly, many urban areas are relatively low-population which weighs against their incorporation with nearby neighbors.
That leaves us with only 6 potential near-term megalopoli, which in my mind elevates the significance and urgency of each.
Unbalanced pairs may not support strong connections
Take Chicago, for example. It has many nearby neighbors, but because it's the largest by far, it has relatively little to gain from enhancing rail, presumably because a relatively small satellite primarily draws from the smaller to the larger city. In contrast, Dallas / Houston, Cascadia, have mutually obvious benefit due to relatively balanced travel interest between the pair. I suspect in unbalanced cases, the pride of the larger city may weigh against investing in integration.
Wildly Different Prospects for and Rates of Progress
I think it's quite remarkable that Texas, California, Florida, and the Great Lakes have such dramatically different prospects for progress. California's bold efforts are mired in a failure of management, Chicago is not setting its sights so high, and having difficulty at that, while Texas and Florida are having substantial new high speed rail development thanks to functional, incentives-aligned approaches to this development.
Leaves me glad we have so many different experiments running in this country, as it seems there are many ways to get this grossly wrong.
Conclusions
These assume the expected successful timely completion of the Texas Central and Virgin Trains USA Orlando link.
Short California, Long Texas & Florida
The long-term prospects for connectivity and relative commercial significance of these areas is substantially connected to their ability to create an arrangement of infrastructure and services conducive to supporting business and quality of life in their areas, which leads to a virtuous cycle of population growth and further agglomeration benefits, such as ability to justify and accrue benefits from large infrastructure projects.
The difficulties apparently experienced most clearly by California, but also by New York and Chicago, augur a future in which other regions gain relative prominence over them, by way of successful infrastructure projects and increased connectivity / agglomeration. Consistent with these signs, population growth in the top CSA of each region is significantly higher in Texas (17%) and Florida (14%) than California (6%), the Northeast (3.7%), or the Great Lakes (1.6%).
Design your Infrastructure incentives well
Texas and Florida are taking substantially different approaches to their high-speed rail than those which have foundered or are otherwise improving quite slowly in other regions, and I expect them to benefit from these differences. In particular, they are following the examples of private rail transit in Asia, and using a combination of fares and real estate development around the station to fund the rail service. This is a variant of land-value capture that aligns incentives, because it creates a material monetary reward for those for fast and efficient construction efforts, accruing to those with the ability to manage construction costs. In the Northeast and California, lacking such incentives alignment has meant that construction projects are seen as a windfall to be exploited by those doling out and receiving construction funds, which then leads to project cost blow-up. This is devastating because there is no upper limit to the cost of construction inefficiency - that is the cost inefficiencies can dwarf the project construction costs.
An alternative to fully-private infrastructure ownership, is to choose a project delivery method which creates a natural incentive to control construction costs, such as through Design-Build-Operate-Maintain contracts.
Expect new technology to further change the game
Nascent technologies such as Hyperloop could significantly improve prospects for agglomeration by making private construction feasible in more cases - smaller city pairs, longer distances.
[1] Dividing the distance (457 miles) by the expected future journey time (5 hours).
[2] Dividing the distance (168 miles) by the average journey time (3.25 hours) listed on Wikipedia.