Rent Growth & AI job displacement
Hi all!
One thing that has lingered in my mind for the past 6ish months has been the prospect of stagnant multifamily rent growth in the future due to white collar automation. Previously, I was an AI skeptic who would sometimes spiral, but largely I remained on the "it's just a tool" end of things. In my former job, it was insanely unreliable and in many cases slowed us down. Now, with new tools hitting the market like Opus and codex, I'm a bit more concerned. I spend (too much) time in IB forums, big law forums, and talking to my friends in high finance.. they all are kinda wowed by these new tools being deployed -- chiefly PowerPoint and excel plugins that are saving them hours. A few are at the VP level and have expressed much lesser need for analyst grunt work already.
I really don't believe developers can be displaced - not in my lifetime, anyway. It's a big part of why I chose to pivot into the field. However - I feel kinda flanked from the side. Let's say by 2035 we reach a place of sticky unemployment due to automation, which I am beginning to think is a legitimate possibility. How does that affect rent growth? Will capital dry up on the predication that tools advance so quickly, it's tough to commit to 3-5 year timelines? At first I thought I was overworrying, but this has now become a legitimate point of conversation in my program - including among our real estate professors who yes, have worked or are working in the field. Curious to hear folks opinions here. For the record: I hate Elon, Sam altman, Dario amodei, and all the other bullshit artists who make fantastical claims detached from reality. We're not approaching AGI. I am however, no longer convinced that these tools aren't transformative and disruptive to the workforce.
The new Claude excel feature is "better" than any current analyst / associate in the United States right now. The sticky unemployment from automation will happen well well well before 2035.
I work for a developer who usually needs an analyst once every 2-3 years. There is quite literally no need to hire someone for that role anymore. Developers wont be displaced- the amount of seats available in development will shrink dramatically.
Thanks for the reply. However, I'm going to need a bit more from you as it seems like every time I see someone opine anonymously with hardcore certainty and sweeping statements, while mentioning a singular tool, they don't answer follow ups and seem to either be bots or paid boosters.
I'd have to imagine that, if you are working in development, you likely realize that associates often times do way more than grind in excel all day in end to end development. Curious on what you are actually seeing Claude do to support your claim that it's "better than any associate in the United States". I'm not confident you are real and/or a developer.
Edit: and for the record, if you are real, I apologize. You wouldn't believe the level of astroturfing and paid Marketing across the internet on these discussions. I'll also add: I actually can see the need for what we consider to be analyst level tasks reducing dramatically in the coming years. That to me is fundamentally different from the claim that there will be significantly less developers
That's why i used "better." The only aspect of the job where AI is not "better" than the average associate would be the human side- interfacing with vendors, handling monthly bank draws with lenders, reviewing pay apps, vetting SDs DDs and CDs, working with local gov officials, etc.
Everything else, I would rather have AI do. I'm in the midst of a 9% LIHTC application which would have taken me a solid 3-4 months to complete. On track to do it in less than a month with Claude.
C'mon man. I've played around with the Claude excel tool plenty (I wrote the review on WSO). Calling it better is just not true. You have to watch it like a hawk. Might it get to that point, possibly, but it's not there now and you do AI bulls a disservice lying about its capabilities.
It doesn't work well with creating models from scratch, rather modifying and expanding on current templates.
For this specific use case in development, Claude has saved me countless hours in the last week handling our bank draws and model updates.
Sure, it sucks now when you need to to create things from scratch. I don't use it for that though.
Just read your post on the Off topic forum.
My development model has 60+ tabs, 15+ different sources in the capital stack, multiple condos selling at different times.
Was able to pop out perfect sensitivities and update the draw schedules with ease. And I have 5 separate GMPs and budgets. TBH man it sounds like user error on your end.
60+ tabs sounds like a skill issue on your part. What are you trying to do model out how many times your residents flush the toilet.
You’re a muppet. I’m in IB (distressed debt advisory) and it doesn’t touch the vast majority of my analytical work. Calm down lol.
If the end game of AI is substantial white collar workforce unemployment, it probably means less homeownership, excluding those that inherited largely paid off real estate from boomer parents, perhaps more rent growth in the near to intermediate term because less people can afford homeownership, and in the long run (barring a major political and some extent social upheaval) less desire to immigrate to the US, because what else does the US have to offer over many other developed nations if a) job market is terrible for those outside the elite/executive positions and b) less socialist programs, which on one hand could stymie longer term growth. All of this ignores what may be happening in other developed countries in the future along the lines of AI, employment, etc.
What I'm trying to get across is there are so many large variables that seem very volatile and it's anybody's guess how it will impact long term rent trends, with a lean towards something more negative, but actually in the shorter horizon I would take the opposite stance that rent growth will be bolstered by the broader economic challenges ahead not just related to any sort of unemployment rise from AI but also the decrease in development in recent years. I think if treasuries and mortgage rates continue to fall (therefore counter to my premise here, homeownership affordability increases), if they fall very significantly, it would likely be pretty short lived.
Depressing but noteworthy potential tailwind re: home ownership. Yeah, ultimately you're right. No one has the answers. It's almost as if I should stop catastrophizing and focusing on things out of my control or something? Weird
I failed to more clearly state my ultimate point: if things go along the lines of what I presented, I would wager the majority of us have more to worry about than rent growth as it relates to the assets we monitor or technically partially own a small sliver of. Whole world would be upside down.
To be honest man, you need to be worrying about what's in front of your face in the apartment market. All of your primary markets are sucking wind, supply hasn't come down to anywhere near it needs to be, we probably won't see meaningful rent growth for another 18-24 months, and cap rates run a real risk of expanding. The market is shit, people have been lying to themselves about how shit it is, and it will continue to be shit through 2027.
I think AI could be disruptive to white to collar work for a short period but I think technological unemployment of the sorts the doomers worry about is very unlikely. Might we need to shift where people work - sure. But people are still going to work and probably have (much) higher real incomes. After all, we will have access to these AI tools which should increase our marginal productivity of our labor. I find the argument that we can have 10%+ unemployment with 10% GDP growth to be highly implausible.
People learn about comparative advantage in Econ 101 and then throw it out the window when it comes to AI. I am better than my analyst at literally every task required of him. But you know what, I still give him stuff to do because we are all better off if I specialize in the thing I'm best at and he specializes in the thing he's best at. The same thing will happen (in the long run) with AI. Comparative advantage not absolute advantage is what ultimately determines who does what.
Stagnating rent growth in the 2030s is a real risk but it's not necessarily because of AI's impact on demand IMO.
The first reason is demographics. Beginning in the early 2030s, the total number of people in their 20s will plateau and then eventually decline in during the middle and second half of the decade. Unless this reverses due to loose immigration policies (which I do not support FWIW), you could see apartments underperform in the 2030s. And it only gets worse from there. Now maybe immigration and or continued home ownership challenges (people renting longer) bail you out. But I wouldn't bet on it.
The second reason is that if we get the sort of powerful AI Dario, Sam Altman, and Elon talk about, you will likely see a revolution in robotics and manufacturing which should bring down construction costs. The construction market has been the biggest thorn in the side of multifamily development over the past 10 years. When I started my career you could build a nice 3-story walk up for $150K/unit all-in. Today, that costs $250K/unit, an annualized increase of 5%-6%. The main driver has been construction costs. If we get significant construction cost relief we could actually build more units that are actually cost competitive. And there are RE folks who say this sounds like crazy sci-fi. But if you handed someone Claude Opus 4.6 in 2021, they also would have thought it was magic sci-fi tech. Sure the world of atoms is harder than the world of bits, but don't underestimate the exponential and the amount of capital flowing into AI and tech.
Fair point on construction cost relief, though I'd think that some of those savings would go into the land owners pocket (land becomes more of the limiting factor than construction costs and owners of land capitalize on it). Still, I've had a similar thesis though I think we have some time til then and so many other variables can change in that time that I don't think it's worth building a thesis around yet
Yes, certainly the land owners will have the ability to eat some of those savings. Presuming @larry david is in the MF business, I am curious why he's generally against immigration (or is it just looser policy, and not immigration itself?), as he points out the bigger challenges are going to be on the demand side of the equation with population trends in the next couple decades, to the point where development may really slow down for an extended period beyond what we've seen yet.
On one hand, I would hope product isn't going to be rendered obsolete or much less attractive as quickly with the quality of a lot of the larger amenity rich developments these days, but on the other surely tech advances are going to accelerate some pretty cool new features in apartments. But in the event of significantly reduced overall MF demand, it's going to be tricky to navigate that. Too many variables to have a meaningful discussion.
Predictions on what will happen a decade in the future is for academics and economists - not real estate people who have to make and do things in the here and now. The only prediction we need to make is that in year 2037 (or 2047 or 2077), every head will need a bed (true).
The relative predictability/stability of real estate is both its downside and its upside. During tech boom of the past quarter century, there was way more upside elsewhere. But if there's a deflating/unwinding in software/tech (recession/obsolescence in age of AI), then the boring stability of real estate will also prove to be its upside. My multifamily building or retail center will not be made obsolete in the blink of an eye like Turbotax (or fintech or whatever is cutting edge) can be.
And if the AI'poclypse is so bad that they ban private ownership of real estate, then everyone (RE or otherwise) is good and F'ed anyway, so why catastrophize?
Read a lease or a loan or other legal doc from 50 years ago and compare it to today. They're basically identical. That's because our business is a steady/stable. Humans spend 90% of their lives indoors and that ain't changing.
Long live real estate.
there's got to be some WSO policy again people coming on here to push/pitch their product, no?
F that I will shill for Anthropic all day if they give me a free Claude max subscription
AI isn't displacing many jobs in general, let alone in real estate.
How many people do you know who got replaced because of AI? The answer is probably somewhere between zero and zero. This is why AI boosters are constantly talking about how all the magical fantasy outcomes from AI are "around the corner" or will be here "by 2030".
We've had AI operating at more or less it's current capability for nearly four years. It's proven to have nearly no transformative commercial application. Not sure why anyone thinks that will change - but aren't we glad we've collectively spent trillions of dollars so we can have Shrimp Jesus, instead of solving lots of real world problems?
While it's been a little more than three years since the release of ChatGPT, I would disagree that we've had AI operating at it's current capability over that period. Current SOTA models are much more capable than they were at this time last year (and more capable in 2025 than they were in 2024). My understanding is that they've changed the world of coding forever. However, I do share your frustration sometimes with leaders of the big AI labs - everything does always seem 1-2 years off. Although this time I think it might be.......
Well, agree to disagree. Functionally ChatGPT has not evolved in a significant way since 2022. It is slightly less inaccurate than it was. Maybe for certain specialized tasks it has gotten meaningfully better. However, it was being hyped for all the same reasons then, and all of those claims were damp squibs. Why in the world should it be any different now? What is it that is "right around the corner" which is going to meaningfully differentiate current models from future ones?
ChatGPT and Anthropic and the big LLMs have received trillions, literally trillions, of dollars of support and investment and the biggest free media campaign in business history. Every possible advantage a new business could possibly hope for has been poured into this industry, to a degree that is orders of magnitude greater than anything that has ever come before.
And for what? Again, who do you know who genuinely lost their job to ChatGPT? We were promised a technology has fundamentally transformative as the railroads, and we got... Shrimp Jesus. A five trillion dollar meme generator.
I have heard of REIT's bringing in AI-forward guys to replace large property accounting teams (and PM teams - but I think that has potential to be disaster).
I have no doubt the REITs will try this but if neighborhooding is any precedent I am skeptical it will work well
I developed apartment buildings in Silicon Valley for 6 years from 2010-2016. One of my biggest talking points was a tech job could be created overnight, but the housing unit for that person with the job could take 5 years to build. At some point there was - I forgot 5X or 20X- new jobs to every new unit under development. It was compelling from an investment standpoint.
The AI impact of jobs, I think has been more impactful in tech, especially big tech where cost cutting is emphasized. Layoffs are the low hanging fruit, while new services, products and industries sprout and grow and mature (including services beyond LLM’s but became economically feasible with the help of AI). I think we are in that period. I believe AI will have a profound impact to demand and rent growth in the coming years.
It took automation around 40 years to really displace factory jobs. Its not apples to apples but these things do not happen as quickly as we think. Manufacturing was 40% of the workforce in the 40s before automation. I was at an MF before my current gig and they have been developing and testing ai for years, and not a single employee has been layed off because of ai. It's a hot take but, it's more likely ai creates more startup RE companies than creates mass unemployment.
Cum voluptas quisquam nostrum ut. Saepe minus qui veniam molestiae expedita illum quae voluptatem. Ipsam sed ducimus non repellat ut.
Sit rerum excepturi vel cumque culpa. Aut provident omnis et non aut tenetur. Unde laboriosam molestiae vitae iusto nobis excepturi. Officia voluptas quos voluptas consequatur. Quo ipsum adipisci aut provident quae illo ducimus odit.
Iusto dolore ea autem quo. Harum sed sit reiciendis est ad incidunt eos. Dignissimos voluptate est consequatur aut et quia nihil maiores.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Eos corporis harum velit non perferendis. Similique deleniti aut non natus facere dolor quia. Et at modi quia vel.
Repellat ipsam est et ut. A est consequatur rerum perspiciatis hic et aut. Sequi voluptate doloremque itaque accusamus.
Ipsam voluptas recusandae dolores aut. Eum perferendis eos quia distinctio ea eligendi blanditiis. Veniam cum porro deleniti veritatis vel et autem. Laudantium eligendi est ut nesciunt. Minima quod dolorum fugit pariatur voluptates voluptatem. At quidem praesentium officia.