The conventional beauty of the internet and software was that it had effectively zero marginal cost. That is to say, it might cost you a lot of money to create something initially, but once created, you could scale it very quickly, more or less for free. This has been a great way to make money, and it's the opposite of something like real estate development where everything takes forever and costs too much money.
But the landscape has shifted rapidly. Dror Poleg wrote this week that intelligence, rather than software, is now eating the world. The fundamental difference is that while software had zero marginal cost, AI does not. When we ask AI something, it has to reason it out in real time, and in order to do that, it needs to consume lots of energy and compute.
As a result of the above, we are seeing something we’ve never seen before: Software demand is beginning to bump into physical constraints. The world is struggling to allocate sufficient land to build data centers and to produce and redirect the energy required to meet AI demand. Tech giants like Google, Amazon, Meta, and Microsoft are spending an unprecedented amount of money to build these new data centers, but they are approaching their financial limits. Google has recently partnered with Blackstone, one of the world’s largest landlords, to expand and expedite the construction of new data centers.
All this sounds like great news for real estate developers. Finally, order has been restored in the universe: If you want to grow your business, you need to pay more rent; the natural scarcity of land is asserting itself. Instead of software eating the world, it is now the world that is eating the free cash flow generated by software companies.
However, these specific dynamics may only remain true in the short to medium term. As dystopian as it may seem, there is indeed an organized and real effort to bring data centres into space. Some of the advantages of this include abundant, continuous energy and zero land-use constraints to fetter growth. Now, I don't know enough to comment on the feasibility or timing, but it certainly sounds like great fodder for a Black Mirror episode.
In 1983, the Eiffel Tower underwent a significant structural renovation that included the removal of an original helical staircase used to bring visitors up the tower. New elevators were installed in its place, and the specific section connecting the second and third floors was dismantled and cut into 24 sections.
Four of these sections were saved for French public heritage, and the remaining 20 sections were auctioned off to the public. Since then, these stair sections have traded for staggering numbers, with the record being Section No. 13 selling for €523,800 in 2016.
But now, for the first time since the original 1983 auction, Stair Section No. 1 is about to hit the market through Artcurial. The pre-sale estimate is €120,000-€150,000, but as is customary with auction houses (and auction dynamics in general), I'm sure this is a deliberately low number.
If any of you are in the market for an original Eiffel Tower staircase from 1889, you can register for the auction here, which is scheduled to take place on May 21 at 2pm Paris time. I'll be checking in from afar with curiosity. Because somebody is really going to want this.
Photos from Artcurial
Jet fuel costs have nearly doubled since the US and Israel attacked Iran in February. This is obviously straining the overall economics of air travel, but the most impacted segment is the one that has always been tenuous: short-haul flights.
As I understand it, airlines generally prefer flights that are at least 2 hours long. Takeoff and landing consume the most fuel, and add a lot of wear and tear on a plane's equipment, so you want a long enough flight to amortize these costs. This is why for the 10 years spanning 2016 to 2026, US flights spanning less than 250 miles declined by 11% — the largest drop of any route length.
Now, in some cases, these short-haul flights are simply necessary loss leaders. For example, the flight from Milwaukee to Chicago is comically short. It's only about 70 miles, translating into an actual cruising time of around 20 minutes. But it's an important route for connecting passengers and the overall hub-and-spoke airline model.
This also makes it slightly harder for rail to effectively compete, because you need to solve for two clear passenger demands (again, assuming they're connecting): (1) people leaving Milwaukee will want to check their bags at the point of departure and (2) they don't want to arrive downtown, they want to arrive at the airport for their connecting flight.
That said, both of these wants are solvable. Hong Kong, for instance, allows in-town check-in where passengers drop their bags downtown before boarding the airport train. This is particularly convenient if you have to check out of your hotel and need to rid yourself of your luggage until you arrive at your final destination.
Very cool, so what's my point?
I mention all this because if short-haul flights are the flight segment that airlines don't love to operate, then it only strengthens the opportunity for high-speed rail to fill this gap in the market and become a seamless component of overall global mobility.
Here in Canada, the obvious opportunity is the Toronto-Montreal corridor. This is arguably the single best opportunity in North America when you consider its geography, construction viability (lots of undeveloped land to lay new track), and ability to replace short-haul flights. The broader Windsor-Quebec City corridor is also, as we know, the densest part of Canada with roughly 50% of our entire population.
But the overall opportunity is twofold: it will service origin-destination travel and it will connect Toronto and Montreal as global airport hubs. In fact, this is one of the stated reasons for why Air Canada joined the high-speed Alto project as a core consortium partner:
Connections with other modes of transport, such as rail or bus, are part of the solutions the company is already developing to offer the most relevant mobility option, responding in a sustainable way to the specific needs of each of its customers. In the longer term, the contribution of its expertise to the Cadence team will enable the airline to contribute to the harmonious integration of a future intercity rail network with existing airport hubs in the Quebec-Windsor corridor, for the benefit of all travellers.
Here's a specific example. Montreal largely serves as Canada's direct gateway to France's secondary cities, Francophone Africa, and the Mediterranean. So if you live in Toronto and want to fly to Marseille or Algiers or Mallorca, you are going to connect in Montreal (or connect across the Atlantic somewhere in Europe).
The multi-modal train option would include an in-town baggage check at Union Station in Toronto, a 3-hour train ride to Montreal, a seamless rail connection from Gare Centrale to YUL (with the REM airport train set to open in 2027), and then your flight to Europe or Africa.
The overall travel time should be comparable, except in the high-speed rail option you'd have more uninterrupted time to work, watch a movie, or sleep. And now that Air Canada gets to rid itself of its less profitable (or unprofitable?) short-haul flights, it should have the margin to aggressively market these tickets.
If this customer experience is designed properly — with one booking, competitive fares, clean transfers, and convenient baggage handling — it will quickly dominate the market. We know this because it's already working in Europe.