The End of IB and PE as We Know It
This is the end. Most investment banking and private equity jobs are about to vanish. Compare the first OpenAI model to Gemini 2.5 Pro - AI’s evolution is insane. AI’s getting sharper fast, and it’s already cutting into white-collar work like customer service and parts of IT.
There’s only so many companies to IPO, merge, or acquire - deal flow has limits. AI’s taking over the grunt work: modeling, pitchbooks, due diligence. It’s not some sci-fi fantasy; it’s happening now, just not everywhere yet. Managing Directors will hang on for the client-facing prestige, but the need for young analysts? It’s fading. Firms won’t keep paying for human labor when tech can do it cheaper and quicker. This isn’t decades away - think a few years. The shift’s already started.
I used to think the same until I started the job myself. There will definitely be layoffs that come from increase in efficiency but there's going to be jevons paradox and firms will all use them amd they will compete against each other so there will be a prisoner's dillema game theory situation. I think the top 20% of bankers / PE folks have nothing to worry about. AI will accrue to capital and we are in the capital business.
Also really curious about this. I know people will play this down, but think of were AI already is now and think of where could it be in 5-10 years?
I remember it like yesterday when it was 2010 and how smartphones were used then vs. 2020. So I get that AI is by no means perfectly capable as of now, but how will it be in lets say 2035?
I find what it can do already now wild and saving me tons of hours each week, so I imagine with more integrations (i.e., instead of research analysts granting AI access to firm databases, research report, etc.) it will put tons of people out of jobs.
What people never understand is that yes AI will make bankers more efficient, but will simultaneously make clients more efficient increasing their demands of bankers - so not much will change. The exact same way how computers, excel, mobiles didn't replace bankers - because clients have access to those same things as well and their demands increase.
Exactly. The invention of Excel didn’t lessen work hours. The invention of the internet didn’t lessen work hours. We live in a world in which tech has made our jobs as “enhanced” as ever, and we’re still seeing analysts and associates drop dead on the job. Probably a frequently as ever, if not more frequently.
When technology lessens a societal need, there’s a burden lifted. Farming got easier, we don’t go hungry. Medicine got easier, we live longer.
However, when technology just makes human-on-human tasks (I.e. war, dealmaking, law, investing) more “advanced,” it doesn’t necessarily lessen the labour burden inasmuch as it raises the bar.
Decades ago, deal models did not need to be as sophisticated, research did not need to be as instantaneous, and, quite frankly (at least in the U.S.), no one needed to make as much money, as “good enough” could actually buy a house, such that working 60+ hour became
a bit less “worth it.”
I don’t think AI is “coming for our jobs,” nor do I think it will improve our hours. It’ll just mean more work is expected.
AI isn't just another invention, it is an intelligence that can't be compared to the tractor or even the internet, it is much more fundamental. That said , there will still be a pursuit of "finding an edge" but unclear what that means in relation to human intervention.
That's cope. For one, clients are far slower to adopt new technology so the efficiency gains will be disjointed. For two, the efficiency increase for AI approaches infinity so even if demand for banking services increases 200% AI can service that demand without needing additional workers. For three, you are probably wrongfully correlating demand for financial services with technology improvements when the real driver was a period of widespread consolidation and startup creation juiced by low rates. The economy can only consolidate so much, and demand for tech services is heavily saturated... gone are the days where you can make a 500mn business just automating some random little business process.
I always come back to watching my MD use any new technology… yeesh. The adoption curve at the upper levels is slower and therefore it’s the junior people using the technology and getting more efficient. Is there less need for junior people in the future? Yes. But I don’t see why those currently in the game would get fired unless you’re already in an overstaffed business in which case it would probably happen anyways…
What if there’s a golden scenario where these roles all become paradise because we still get paid and work less because of how efficient we are with AI??? Probably wishful thinking.
Pretty sure OP is not in IB/PE - look if IB/PE were to be replaced with AI in few years, then basically BigLaw, Consulting, and any white collars jobs are to be replaced as well, and middle class will be gone with no one to buy the goods from companies which will be the end of capitalism :)
But on a more serious note, this will be another internet revolution when internet became more widespread or when excel was invented.
I was talking to my MD and even back in the early 2000s (this was when internet still existed) they had to go to the company's physical dataroom, look for a booklet that had last 5 years of financials, and fax them to your office and then hand type them into an excel.
Obviously the internet and advanced Excel plugins like CapIQ and FactSet have made all of this mundane process streamlined, much like AI doing a simple 3-statement model or doing a market research based on your prompt.
Given both the client and IB/PE firms will use AI, they will increase the efficiency and productivity of the workers - which in theory should increase growth of the economy with the same amount of resources (i.e., time) available, hence leading to more M&As and IPOs.
Did the internet and Excel decrease the amount of investment bankers required compared to 20 years ago? Maybe. Has it decimated the industry? Not really. Similar thing will happen as well.
And if you think your fear-mongering scenario of IB analysts and associates being fully replaced by AI is happening soon, then you would know that basically all white collar jobs are replaceable if you have ever interacted with a lawyer or management consultant (or worse, any white collar job ever that are paper pushers)
Lmao come on, this is such a classic slippery slope argument. Just because AI might replace some parts of IB/PE doesn’t mean we suddenly have to throw lawyers and consultants into the mix to make it sound more reasonable. Yeah, maybe everyone's vulnerable in theory, but that doesn’t make it any better for bankers. Saying "hey AI will get us all eventually" doesn’t actually address the specific ways AI could hit banking harder and sooner.
And honestly, you're kind of underestimating how far along AI already is. We're not talking about some vague sci-fi future. This stuff is already insanely good at crunching data, writing summaries, doing basic modeling, scraping market info, you name it. Now give it 5 to 10 more years of development? You think it's just going to plateau? That's wild. It's not Excel v2, it's like Excel on rocket fuel with a brain attached.
And yeah, we've always had tools to make us more efficient, but this isn’t just faster typing or better plugins. This is about AI actually doing the thinking, not just the formatting. That’s a way bigger shift than the internet or Excel ever was.
The whole "well did Excel destroy banking?" thing kind of misses the point. Excel made us faster, sure. AI might actually replace parts of what junior bankers do. That's a different game. Especially if both sides of a deal are using it, that doesn’t necessarily mean more deals or more jobs, it might just mean fewer people are needed per deal.
Anyway, acting like AI equals Excel is comforting but kind of delusional. We’re not getting wiped tomorrow, but pretending everything is fine because lawyers exist too is just dodging the actual conversation.
Never said we are fine - I gave the example for BigLaw and MBB because I have seen their work in LDD and CDD (as you have so in PE) in M&A and it is certainly more prone to automation. I'm sure in few years there will be ChatGPT that can draft a whole SPA and SHA docs of a complex public merger and the PE/IB MDs just need to type few prompts without any lawyers except for in-house counsel to protect themselves. Same with CDD. Just scrape all the available public + prop. data from various sources and output into 50 slides about the latest Healthcare dynamics and its growth trajectory in South East Asia or something. Still, regulation + who is going to take the fall if shit goes wrong + hallucinations in AI are still major issues to solve.
I agree that PE and IB are prone to AI impact - what I am saying is that once PE and IB gets impacted by AI, a large part of white collar workforce will be also impacted hence affecting the middle class thus affecting the overall foundation of capitalism. So there is a bigger social issue to solve then just saying IB analyst classes are smaller now.
Being able to pull up a random company's entire financials on capiq within seconds rather than spending an hour finding it on paper in a dataroom somewhere is arguably more of an efficiency increase than what AI can/will do. And that didn't change the industry much.
I think it’s fair to say that if PE gets automated, any white collar job can be, because imo PE requires a very high level of the specific critical thinking / synthesis / abstract thought / fluent mix of logical+critical thinking+judgement that currently separates humans from LLMs.
The reality is that all of these white collar jobs will be automated.
Everybody won’t be fired but it’s fairly reasonable that you’ll only need half the analyst class.
There will still be bullshit iterations but right now, don’t tell me it’ll take you more than a few hours to do comps. ChatGPT can give you 80% of the relevant comps already. Same for market research. Just needs some fine thinking and critical thinking.
Your first paragraph here is exactly what Marx predicted would happen. It’s called a “supply shock.” Basically…workers become “efficiency’d out” of earning what they used to/having requisite purchasing power to afford what they used to, which decreases demand. Hence, what is being sold becomes “oversupply.”
To adjust, in Marx’s view, companies have a few levers, including: 1) increased “emiseration” (longer hours for stagnant pay to keep profits intact…which, as decent-paying jobs become more competitive as a result of the technological advancement … becomes an economic norm), 2) external exploitation (colonization…or a recalibration of trade with the aim of importing cheap raw materials and dumping expensive exports on other nations), and 3) forced labour (such as prison labor).
Right now, we’re seeing 1) in our industry (just look at how much people here complain about what pay is now vs what it used to be for the same work), and the others elsewhere without our economy…with the possibility of intensification.
However, there gets to a point where an economy “supply shocks” itself out of existence, as technology simply eradicates the need for some of its stabilising pillars. It is in this case that upheaval is theorised to follow, but it does not appear that we are close to this at this point.
Bankers and finance more generally will be the LAST to know when their jobs will be automated. Couldn't imagine a cohort more up its own ass and blinded by their own shit to see the writing on the wall
Yea I go back and forth on this. A lot of banks don't even allow you to use AI because they don't want you to upload nonpublic info to a third party platform. I have also heard of (although my bank has not done this, yet) banks developing their own internal tool (no third party risk) and all my friends at these firms say the tools just suck. That will probably change but it will take awhile. I agree that it significantly reduces workload if you can use it (I could in theory break down credit docs much easier and spread them, it could help me build out various scenarios on excel, maybe even help create graphics for new pages). I don’t however think this necessarily means layoffs. Our pipeline is strong (at least for now), so I think it just means we could do these deals faster, and pitch a wider net of businesses. I was really worried about job security (maybe overly so) but I agree with the points above on new technologies not really causing mass layoffs (phones, excel, pptx, etc.). I watched an interview of Jamie Dimon saying how he used to have to go to a physical library to do research. Has JPM’s IB team grown with the advent of the internet or shrunk? I am not that worried. Although I recognize my bias…obviously I want to rationalize the potential impact and minimize / downplay the risk to my job. So I could be wrong.
As an analyst, I really hope it replaces analysts and associates - being grinded into the dirt and being on the brink of death for grunt work makes no fucking sense.
However, it won’t replace analysts - it’ll just increase the efficiency and productivity of analysts and associates - instead of working on 1 model, they can work on 10 at once. Instead of being on 3 pitches, they can be on 20 pitches. My best guess is they’ll probably be a need for ~30% fewer juniors, but the juniors will be like 50% more efficient and productive, and will be able to take a higher level view.
If anything, I hope it kills useless and expensive VPs, many of which are negative value adds and don’t deserve to have jobs.
Lol go back to your intro to accounting course bro, thanks for the market insights
PE is largely a relationship-driven asset class. When your AI can golf, go to conferences, attend MPs and dinners, and get shitfaced with bankers’ AI, you may have a point.
That’s the MD’s job. Analysts don’t go golfing with clients.
If you have well functioning AI tools, I don’t think it’s unreasonable to cut the analysts, or even associates out of the team. An Asso/VP can do the fine tuning.
Do you think MDs grow on trees?
Bro do you think that buyers care about playing golf with some fat douchebag more than they care about saving 50mm in fees that they can then pocket?
There's too much compliance risk, reputational damage, and litigation issues to ask a blackbox for analysis. Unless our corporate law environment changes, this will never happen.
I see someone doesn’t understand the technology they are referring to
Probably used ChatGPT to make this thread.
My take on the “where will the future deal makers come from?” - finance will go back to the way it was originally. The partner’s son and MDs daughter will be the only two analysts in the Goldman class of 2050.
AI is going to push us into a new gilded age where these middle / upper middle class white color jobs will funnel down to just the connections of the very small tip of the corporate pyramid and minorities and poors will need to find a new ladder to climb up.
God I hate these doomposts. Even interns in FDD do better than AI there.
I still have yet to see AI build a good simple deck or excel.
Another way to look at it is on the supply side. AI has effectively democratized the development of software products, at least at the high level and front end. So more people are building businesses than ever, which means (hopefully) more will need financing, deal management, and other value adds that the financial sector provides. You can't look at this topic in a vacuum. There are many other factors at stake.
I also really don't think that the boomers running the high finance game are all that interested in quick adoption of this type of technology - they're simply too stuck in their own ways. Besides, most funds I've talked to don't allow their associates to use ChatGPT due to privacy and security reasons. There's a big opportunity out there to create a model or layer/wrapper that allows the use of LLMs without feeding it back data.
Lastly, I'm personally super fucking bearish on Agents. I've been in or around the VC space for a minute, and I really don't see the use case in the majority of startups building agentic AI.
Saw a post the other day about a plumber coding a schematic app or something using LLM’s, no prior programming experience. In my college python class, most students were reliant on gpt. I remember showing my engineering buddy gpt-3 in 2022 and he was absolutely flabbergasted along with the rest of the public, now we’re desensitized.
In 2023 the top models were about at middle school level of competence, now we’re approaching the level of a graduate student. The critical point is when these tools can be assigned a task and work for long periods of time. Imagine an employee that will focus 24 hours a day, does better work, costs less, and won’t cause problems. No benefits, bonus, office space required.
We could argue that people will always want to talk to people not robots, and firing all your employees for ai slop to save a few bucks will hurt not help your business, ie Klarna. No one wants to be in a wreck caused by an autonomous vehicle or be wrongly convicted under the defense of a bot lawyer. However as these models analytical aptitude improves, so does their emotional reasoning. On paper, autonomous vehicles have lower crash rates than human drivers, mental health counseling outcomes are equal or above traditional methods in some studies, and some diseases are more successfully diagnosed.
A lot of the public figures have been sounding their own alarms with Daniel Amodei, Sam Altman, Chamath Palihapitiya predicting layoffs and a mass deflationary event. Marc Benioff going so far as saying, “I’m here to witness the death of capitalism.” I might not go that far but it’s not unreasonable to assume that as software becomes commoditized, the premiums we’ve ascribed to these growth tech companies lose a lot of credibility.
We’ll probably see a mix of people using the tools to improve at their jobs and people getting fired or assigned to monitoring roles. Maybe those who are displaced adapt and apply their skills elsewhere and I’m sure we’ll come up with a solution, but the effects won’t be negligible.
I don't see it because MDs don't have the patience nor the interest to write out a sufficiently good prompt for AI to be able to do its work. MDs want to be able to type "pls fix thx" from an uber and have the analyst just know what to do.
Seems kinda facetious but still gonna bite - an integrated, highly competent AI model will have the same critical reasoning and (in)ability to interpret vague, lazy instructions as any analyst. If there's some MSFT suite equivalent except AI all things business productivity, it would just fix the deliverable (or even suggest 2 versions that the MD could pick between) in seconds regardless of time of day/night
ib/pe skills at the junior level is so replaceable, all you need to do is know basic to advanced modeling + stick though the grind and pain of the train of pls fixes, and if AI can align numbers and catch errors summarize things then why would banks need to pay analysts such high salary if half of the grind is gone? there won't be less junior finance jobs but the pay should definitely come down.
also agree with above, super bearish on agents, but bullish on the rest
Was just talking to a good friend who works in AI over the weekend. Things are going to change massively in 5-10 years. If the majority of your job is sitting behind a computer with low to mild critical thinking, you’re likely to be considered redundant.
Will there be fewer analysts and associates in IB and PE in 5 years due to AI... probably.. but I wager it's a 20-30% reduction and not a 70-80% reduction. The reason is that IB and PE are zero-sum, ultra-competitive industries - firms are constantly trying to find an edge. If AI gets good enough to produce great pitch decks, every firm will adopt it and then seek out the next competitive edge which requires humans to push to find that edge.
@xyzdonttell and @monkeyvjit curious to hear your takes on why you're so bearish on AI agents?
There will be fewer analysts if you keep that grammar up
thanks bud, typed that up on my phone.. that's what we need AI for!
Went from paper and pencil to calculator to excel. Seems like a lot of new jobs sprung up in light of the tech adoption. Can you imagine having to do financial modeling using nothing but graphing paper and a pencil?! 😵💫
I don't agree at an extreme that the Analyst will be completely removed because at the end of the day, AI is just a tool. But I do agree that the required skillset of a typical IB / PE Analyst will evolve. The current recruiting criteria is a joke; as someone who worked at a PE firm and now at an IB firm, the base knowledge set required to do this job can be learned in a 4 week finance 101 crash course; everything after that comes down to knowledge accumulation and work ethic and repitition that's it.
And as someone whose been working for about 3-4 years now, it's a joke how this process entirely hinges upon a software that's been running since the 90s. I can't explain how many things I do on the daily which can be automated, filling in numbers in an Excel manually from a PDF, even the valuations and models we run; there is nothing overly complicated about the mechanics of valuing an asset at the end of the day - it's all a series of steps and rudimentary calculations that a model can be easily trained on. The actual value you assign is where the expertise comes in yes. Finance is one of the most antiquated industries out there, even Law has seen a proliferation of a lot of legaltech startups that have taken over the document management process.
Fintech is getting there slowly but OP is right, there is going to a lot of disruption in this space. We're already seeing fintechs come in to automate the diligence and dataroom aspect of PE underwriting deals. The image of a typical sell-side advisor and the expertise required to be an MD will drastically change - knowing how to manipulate data effectively and having some sort of coding knowledge, layered on top of it an understanding of how the backend of AI functions, how it can be integrated into different business operations to increase efficiency, all of this will become super important for the typical Analyst/Associate. A finance degree is a joke, there wasn't even such a thing as it before, the real theoretical and quant part of finance sits within econ which is an entirely different type of work that still requires a lot of quant / data science knowledge. The actual mechanical part of finance (accounting, modelling, etc) - I could teach this to a high school student, even a middle grade student in a couple of months. And because everyone knows this, people try to solve for it by basing hiring upon things like the prestige of the school, nepo connections, GPA, etc... which is a joke, gone are the days when an Art History major from Harvard has the necessary knowhow to advise an AI company looking to raise capital. The only reason PE firms and strategics use sell-side advisors is because that's the best thing available and because of easy access to network - once you see a sell-side advisor that's AI-centric, tech-first in the way it manages the entire sell-side process, hires people with data science and tech backgrounds layered on top of finance, the whole game is going to change very fast very quickly.
Data manipulation and data science knowledge is such an understated skillset and is going to become increasingly important in finance. All we do is manipulate different datapoints on a day-to-day basis: financials from companies, qual info from research reports, commercial due diligence documents... you're telling me the most efficient way to do that is the current way it's done, where one measly Analyst inputs it manually into an Excel file, pastes it into a wonky looking CIM, manually adjusting logos and diagrams, and inputting the most salesy text ever, that looks like it was made by a high school student, and if you went in and checked every single number in it 99% you'll find at least ONE error in them. Please this whole process rests on prestige and the old guard being unwilling to innovate and transform and the fact that there's nothing better than this currently. Once that changes, things will shift very quickly
I think it will consolidate work to the top groups and those groups may even expand.
Bottom feeders, survivors and low margin groups will probably struggle or get chopped
I now run two AI companies, coming from PE. You need a human in the loop. We will just need less analysts for the same amount of deal flow until the data availability increases to have a higher demand for the increased productivity.
People like working with people at the end of the day.
AI won't die on the job or complain about work hours.
Put the fries in the bag bro
I think people are overestimating the cost of analysts / associates. I run an analyst / associate intensive group at a boutique and pay top of the street. My total cost of analysts and associates is 5% of group revenues. I personally make 2x the total cost of analysts and associates. This is just not an area where cost savings matter rather than the risk of losing out on any incremental productivity gains. Junior bankers aren’t going anywhere.
Yes, people are going to say we should pay our analysts more but that’s capitalism.
Yes, people are going to ask why we don’t hire more juniors and that’s because I don’t want to dilute real deal experience which I think is the bedrock of learning.
Yes, people are going to ask about junior layoffs. I think it’s ridiculous but not every team has the flow we do, but we certainly only have ever let go for performance which needs to be a constant process.
And note: VPs, Da and unproductive MDs are a whole different story on cost / usefulness.
I would also add two things.
I would love to have AI automate a lot of the nonsense my juniors have to do, which are currently necessary evils. Believe it or not, if there was less crap in their day (that comes from client requests or more often, poor systems), I could assign higher level work to my analysts and it would help them and help the business.
The other piece of this is that analysts and associates are 5% of my revenues. My allocated overhead costs (IT, internal legal, compliance, HR) is 15%. AI could take a knife to that and have a huge impact relative to the false savings of reducing analysts and associates.
I embrace the role AI will play in the business - think it will be good for our clients, good for me and good for my As / As
No offense, but bankers will be the last to know when they are automated out. Also, your revenues are going way down going forward. Rising rates, lower growth, demographic shifts, geopolitical instability etc.
D
This post was written by AI. The hypens are a dead giveaway. And the new ai models just say whatever u want them to say to you, so pretty easy to craft this 1 up
The only thing it's going to do in practice is provide more ammo for the top half (MD/VP/etc) to keep the bottom half (Analyst/Associate) busy with different types of non-value adding work.
Pitch books are going to even prettier.
In the next decade, AI will fundamentally reshape how we work in investment banking — but it won’t replace it.
Tasks like basic DCFs, comparables, LBO models, market research, teaser drafting, and preliminary data room reviews are becoming increasingly automated through smart scraping, instant summarization, and clause/KPI extraction.
But the real value will remain where AI falls short:
AI will elevate the role of investment bankers — but it won’t eliminate it.
MDs do not grow from trees. If Analysts or even Associates disappear who's gonna carry the torch into the future. AI is for sure gonna pose challenges but I am much more worried for other kind of jobs where replacing you is not a long-term cost for the company.
I think the biggest shift in the industry will be the death of small, generalist shops (both within banking, legal, and DD providers). Size matters more in the future because AI is dependent on large datasets. If you are closing 10 deals a year, your dataset will simply be too small to create meaningful value, and I think the volume-driven MM banks will put the smallest out of business.
I also think DD processes will take half as much time, as the software for financial reporting becomes better. The numbers are what the numbers are, and as Targets utilize better software, a lot of the mundane "tick the boxes" questions will disappear; You'll see small companies being able to produce highly granular reports, similar to what you might find in a PE-owned MM company today. Similar story with LDD reports, as lawyers can sift through the dataroom in minutes. rather than days. The freed up time will be spent on commercial items and risk mitigation, making it more important for lawyers to be commercial rather than pure legal experts.
The bankers will not just oversee the other human advisors but also manage the Targets AI-advisors/agents.
Pariatur illo deserunt dolor. Sapiente voluptas aliquam optio mollitia officiis eligendi beatae. Voluptas aut debitis est incidunt quia. Ea sed quo accusantium quia blanditiis suscipit est. Omnis et necessitatibus qui non. Quod optio adipisci illum et necessitatibus asperiores rerum.
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...
Delectus rerum tenetur non corporis et earum. Tenetur molestiae quas dolores. Sed quo ducimus consequatur et voluptatem itaque. Voluptatem aut dolorem dignissimos sunt est eos. Quos corrupti adipisci provident sint est deserunt.
Earum nisi et quisquam ipsa quos est. Perferendis aut provident perspiciatis dignissimos est dicta architecto. Exercitationem pariatur et asperiores rerum occaecati eligendi.