Retrofit and the AI revolution
Maybe we should stop knocking down old buildings and replacing them with bigger ones that we don't need.
I packed a lot of slides into my 20-minute lecture at the 27th International Passive House conference last Saturday, including a few of 270 Park Avenue in New York City, the new headquarters being built for JP Morgan Chase. The conference focused on retrofit, which is what should have happened here, instead replacing the largest building ever demolished on purpose, and the largest building ever designed by a woman (Natalie de Blois).
I called it a pointless exercise that would result in 65,000 tonnes of unnecessary carbon emissions, questioning whether, between the effects of the pandemic and the new world of artificial intelligence, the new building was even needed. I have followed the saga of this building for years, so will expand on this here.
Jamie Dimon, CEO of JP Morgan Chase, justified the project by saying that he wanted to have all of his employees under one roof instead of scattered around town, saying in a 2020 letter to shareholders: “We will, of course, consolidate even more employees into this building, which will house between 12,000 to 14,000 employees.”
Unlike many companies that have allowed employees to work from home part of the time post-pandemic, which has already created a huge glut of office space, Dimon dragged many of his employees back to the office kicking and screaming while admitting that it needed less space:
“Remote work will change how we manage our real estate. We will quickly move to a more “open seating” arrangement, in which digital tools will help manage seating arrangements, as well as needed amenities, such as conference room space. As a result, for every 100 employees, we may need seats for only 60 on average. This will significantly reduce our need for real estate.”
I wrote in 2021 that, with the working from home (WFH) and hybrid work trends, we would likely need much less office space in the near future.
With less demand, other wonderful buildings like Union Carbide can be restored, and brought back to their former glory, with more than enough square footage to accommodate everyone. Jamie Dimon says "we never prepared for a global pandemic," but we all can learn from it: things change. We don't have to knock everything down and replace it, we can fix it instead. And JP Morgan Chase and everyone else needs a new definition of sustainability that recognizes this.
Alas, it came down anyway, and a few years later, years later we have AI banging on the doors of the real estate market. Dimon writes in his latest letter to shareholders that the “consequences [of AI] will be extraordinary and possibly as transformational as some of the major technological inventions of the past several hundred years: Think the printing press, the steam engine, electricity, computing and the Internet, among others.”
“Over time, we anticipate that our use of AI has the potential to augment virtually every job, as well as impact our workforce composition. It may reduce certain job categories or roles, but it may create others as well. As we have in the past, we will aggressively retrain and redeploy our talent to make sure we are taking care of our employees if they are affected by this trend.”
Many are sceptical about the usefulness of AI and whether it is as important as Dimon suggests, but big business and critics like Brian Merchant see its usefulness in eliminating jobs and cutting wages. Banks are planning major cuts as AI replaces “armies of analysts” banging away in Excel spreadsheets. According to the Globe and Mail,
“Top executives at Goldman Sachs, Morgan Stanley and other banks are debating how deep they can cut their incoming analyst classes, according to several people involved in the continuing discussions. Some inside those banks and others have suggested they could cut back on their hiring of junior investment banking analysts by as much as two-thirds, and slash the pay of those they do hire, on the grounds that the jobs won’t be as taxing as before.”
Real estate consultants suggest that the impact of AI on real estate will be far more significant than the pandemic. Don Catalano notes: “AI stands to be one of the most substantial disrupters to the labor market since the Industrial Revolution. And its ability to leverage technology will make the effects that the pandemic had on work and the office market seem weak by comparison.” He suggests that it has “potential catastrophic ramifications” for the office space market, with “higher vacancy rates and a subsequent devaluation of office properties across the board.”
Many of the prominent buildings that are under threat are making way for office space. The Marks and Spencer store on Oxford Street in London was going to be replaced with a new building containing only two floors of retail space; the rest is proposed as office space, something I cannot imagine London needs more of in a trifecta of WFH, AI, and Brexit.
I wrote earlier about Marks and Spencer, noting that everyone is “studiously ignoring the truth about the importance and scale of upfront carbon emissions because everyone is having such a good time knocking buildings down and building bigger ones.”
But I suspect that the good times and the era of knocking down buildings and building bigger ones are over. Simon Kuper of the Financial Times writes:
“Vacancy rates for office space in the UK and US remain around record highs, reports the research company CoStar. Certain collapses in value have been spectacular: St Louis’s largest office building, the former AT&T Tower, which sold for $205mn in 2006, will soon be auctioned with bids starting at $2.5mn.”
And that is before the impact of AI. The organizers of the International Passive House Conference were brilliant in their focus on retrofit; we have 220 billion square meters of buildings to fix, and a lot of it is going to be empty office buildings.
And maybe all those unemployed armies of analysts can learn PHPP or how to hang windows.
See my earlier post: Let's get creative with our office-to-residential conversions
"...and the new world of artificial intelligence..."
I DO assume that you know the current estimate is for $2 Trillion USD to be spent on new data centers to house all those server farm hardware, right? You DO know that most old buildings are, even with retrofit, totally unsuitable for retrofit for this purpose. Thus, a LOT of new building is going to happen as all those AI instances are going to need a whole lot of space.
Heh! I can't WAIT to see you do a post on the electrical grid demands this will take. SMRs, everybody?
AND!
"...as AI replaces “armies of analysts” banging away in Excel spreadsheets..."
Except for the smallest firms (or even departments), this shows how far behind you are on real business intelligence ("BI") software tools. Sure, ad hoc stuff would use a spreadsheet, but as a famous person in the US is wont to say: "C'mon, man!".
Some buildings are going to be more difficult to re-use than others, Hospitals & Schools come to mind. Often the "best" the owners may offer is to "save" are the facades of these kinds of structures.
Having said that, I think your point is valid in that the re-use of some older office buildings can be reimagined as mini server farms. Take for example the 100+ year old Whitney Block in Toronto, part of the Provincial Government office buildings in and around Queens Park. As a heritage building, it can't be demolished. Due to many constrains imposed by modern building codes, the Whitney Block's use as either office or residential uses are almost impossible without impacting the heritage elements.
If this building was rather used for government servers, building code restraints would likely not apply, & the building could not only be preserved, but find a new life. I'm sure there are many similar examples around the world.