Big Tech Cut 950,000 Jobs... And Then Hired Them All Back
All How Money Works transcripts →
Summary
This video discusses the widespread layoffs in big tech, questioning their strategic benefit and highlighting the negative impacts on morale, institutional knowledge, and company reputation. It explores the financial and reputational costs, the role of AI as a justification, and the underlying motivations of executives, suggesting that many layoffs were driven by a desire to reassert control and manage stock-based compensation rather than genuine cost-cutting.
Key points- Meta and other major tech companies have conducted significant layoffs, collectively impacting tens of thousands of employees over recent years.
- Despite widespread layoffs, many tech companies have not substantially reduced their overall headcount, instead cycling through employees.
- Layoffs damage workplace morale and lead to 'turnover contagion,' where remaining employees update their resumes and consider quitting.
- Companies lose valuable institutional knowledge when experienced employees are laid off, which is difficult to replace with AI or new hires.
- The justification of layoffs by AI has often proven flawed, with companies needing to rehire experienced engineers when AI systems fail to deliver expected results.
- Executives may be motivated by a desire to reassert control over empowered tech workers and manage the dilution caused by employee stock compensation (RSUs).
Chapters
Notable quotes (4)
“The most immediate damage is just to workplace morale, which I know might sound a little bit wishy-washy, but in roles that rely on talent in highly collaborative environments like tech development, it can be a big deal.”
“Institutional knowledge like which workarounds actually matter and which cursed legacy code should never be touched and even really basic things like who sits at what desk makes a big difference in how quickly stuff gets done.”
“acknowledging the problem that there was nobody left who could tell the AI why it was being dumb.”
“Layoffs repeatedly failed to cut costs or improve performance once you account for severance, morale damage, and lost productivity.”
0:00In May of this year, Meta announced it would be laying off around 8,000 employees, or around 10% of its entire workforce. Just 2 months before that, it laid off a further 700 people, and 2 months before that, it was another 1,500 people. Again, in 2025, it laid off a further 4,000 people, including a large chunk from its AI division. And this was all on top of the record layoffs from its year of efficiency, which actually lasted 3 years and cut an additional 26,000 jobs. Now, obviously, this hasn't been great for morale, even after the Zuck mandated that people start having fun again, but if you do the math,
0:34that's collectively around 40,000 people that have been laid off over the last 4 and 1/2 years. All from a company that only had around 75,000 employees to begin with. For most of the other major tech companies, the numbers aren't really much better. And in specific cases, they are actually much worse. This also isn't including the people who have just, you know, made their you money and quit. Now, obviously, big tech in particular has been changing rapidly.
0:59Companies are investing hundreds of billions of dollars into AI development, which means they just have less free cash flow to pay their staff and a greater incentive to justify that very same investment by doing supposedly AI enabled layoffs. But if this is really happening so broadly, how many jobs do these companies really have left a cut? Well, the good news or the inconvenient truth is uh basically just as many as they started with. Apart from the outliers like X, most of the major tech companies have just as many people today as they did when layoffs became the hot new trend in the valley. The same is true for more traditional industries as well that
1:33have quietly had to curb their layoff enthusiasm. And this does raise the obvious question of why bother? These companies have effectively passed the same group of technical talent around between one another, killed the vibes that they so clearly covet, undermined the job security that made tech jobs so desirable in the first place, all to generate headlines that didn't really look as good as they thought they would. So, what was all of this for? This week alone, about a dozen major companies announced job cuts, including Amazon and Google's parent, Alphabet. Meta announced Thursday that it's laying off about 8,000 workers as it continues to ramp up
2:08spending on artificial intelligence. In a dramatic twist, Open AI Sam Alman has revealed that Meta is dangling jaw-dropping $100 million signing bonuses to lure his top engineers. 76% of Nvidia employees are millionaires and 37% are worth over $20 million. How crazy is that? Ford has rehired 350 experienced engineers after admitting that artificial intelligence alone did not deliver the quality it expected in vehicle development. Okay, so even if you set aside the very real human cost of doing all of this and just look at it through the cold dead eyes of corporate strategy, these layoffs have still been doing real damage to the companies making them. The most immediate
2:52damage is just to workplace morale, which I know might sound a little bit wishy-washy, but in roles that rely on talent in highly collaborative environments like tech development, it can be a big deal. Workplace researchers have observed what they have dubbed turnover contagion. The idea is that when a round of layoffs goes through, the survivors start updating their resumes and people who are already thinking about quitting naturally become slightly more motivated to make that jump.
3:16Andrea Durler, who leads research at the workplace analytics firm Vizier, describes this whole cycle as a clear failure of workforce planning, which is the polite academic way of saying these companies keep firing people they still need. The second problem is that especially in technical roles, a lot of doing the job well just comes down to having done the job for a long enough time.
3:35Institutional knowledge like which workarounds actually matter and which cursed legacy code should never be touched and even really basic things like who sits at what desk makes a big difference in how quickly stuff gets done. In theory, companies like to imagine that this gets captured in a handover dock. In reality, even if the laid-off workers really did want to put their all into teaching their replacement how to do their job, it still doesn't. I will get into this particular example in detail a little bit later, but Ford recently had to hire back a team of senior quality assurance engineers charmingly referred to as the Greybeards to address certain
4:10failings in the company's production. The company had attempted to replace them with an AI system overseen by a much smaller team of more technically qualified but significantly less experienced operators. And well, credit where credit is due, the company reversed this move, acknowledging the problem that there was nobody left who could tell the AI why it was being dumb. Now, this is a problem because when a company eventually realizes it cut too deep, getting people back usually costs a lot more than keeping them around in the first place.
4:38If a company is rehiring someone, that normally means they were a talented employee and they've usually landed somewhere else. So businesses are paying a premium on top of the recruiter fees and layoff expense to get people back who for obvious reasons don't necessarily want to work for companies who have a reputation for letting people go. Companies like Meta have developed such a bad reputation for morale turnovers and layoffs that they are now paying a measurable premium to hire new staff because if given a choice, people would take almost any other option for the same level of pay. More broadly, according to industry estimates, laid-off roles are now being
5:11refilled at a significant premium from what the original position paid in the first place. Some of the people they want back also just simply aren't coming back at all. Laid-off employees naturally realize that their employer doesn't and didn't really care about them. Of course, that was probably always true, but actually being laid off makes it a bit harder to ignore. Support groups for laid-off Googlers reportedly ballooned after 2023 cuts, full of people who hadn't written resumes in 15 years because they genuinely believed they found the last job they would ever need. Now, I know there's a level of irony in these people losing their extremely highly
5:44compensated jobs, often developing the tech that replaced them, especially since, let's be honest here, a lot of them would happily automate other jobs if it meant a slightly more lucrative RSU package. But the point here is that purely from the business perspective, if layoffs were supposed to cut costs, they haven't been doing a very good job. This is also ignoring things like severance, higher unemployment insurance rates, and recruiter fees to enable the constant churn. The losses aren't random either. Stanford economist Nick Bloom pointed out that attrition driven by return to office mandates loses companies or best people first because the best people are the ones with
6:19outside options. What's left is mostly the people who couldn't afford to leave. Now, obviously, mandatory return to office is slightly different from layoffs, but they were done with the same intention of thinning out the workforce, hoping that people would quit instead of publicly having to reduce headcount. Now, it sounds dumb, but it gets dumber. So far, these have just been the immediate costs. Entry-level hiring has collapsed by somewhere between 55 and 65% at the major tech companies since the layoff era began. IBM's head of HR, Nicol Maro, has been unusually blunt, warning that if you cut the pipeline, there's no secession talent in 3 to 5 years. And IBM is
6:56now tripling its entry-level hiring to patch the hole it dug for itself. And on top of all of that, these companies are burning the thing that made them special employers in the first place. A job at Google used to be the prize. Partly this was because of the pay, which at the time was decent but still inferior to something like finance, but mostly because it was seen as the best place in the world to work. It had cool offices, interesting projects, a decent work life balance, a vibe culture, and most importantly, good job security. If that perception dies, the smartest people can just take their talents to finance instead, which pays comparably, never pretended to
7:31love you, and is somehow now the less risky option for a high performer who doesn't want to find out from a LinkedIn notification that they no longer work there. Now, to play devil's advocate here, despite the beanag chairs and kombucha on tap, these are still for-profit companies, not a jobs program. Even if there are costs associated with doing layoffs, and there are a lot of costs, it could be argued that endlessly accumulating staff who feel overly secure in the role would end up costing a business more, especially if all their competitors are willing to pay a premium to poach their best talent. But well, these companies are still accumulating workers. Alphabet finished
8:092025 with around 190,000 employees, essentially back to its all-time peak. Microsoft ended its last fiscal year with 228,000 people, identical to the year before, despite cutting roughly 15,000 across two heavily publicized rounds. And Amazon is within 2% of its 2021 high even after 4 years of going allin on headcount efficiency. So, the workforce is right back to where it started. And since most entry-level hiring has been so slow, this pool of workers has mostly been maintained with people just moving to the next company in line. So they did all of this, accounted for all of these costs, accepted the human toll, took the morale and reputational hits, all to effectively
8:48just play past the parcel of employees. Which raises the obvious question of why? Well, Stanford business professor Jeffrey Feffer has spent decades on this exact question, and his conclusion has been depressingly consistent. Layoffs repeatedly failed to cut costs or improve performance once you account for severance, morale damage, and lost productivity. The same was true for headcount trimming efforts like return to office mandates. Researchers at the University of Pittsburgh's Catz Business School went looking for financial improvement at S&P 500 companies that imposed strict return to office mandates and found no significant improvement at all.
9:21Employee satisfaction, on the other hand, dropped measurably. Now, say what you want about the senior leadership of these companies. Most of them aren't stupid. They are capable of reading these studies and understanding the basic cost benefit of what they are doing. But across corporate America and particularly in tech, they are doing it anyway for three simple reasons. The first is quite simply human vindictiveness. For the last decade in particular, tech workers had a little bit too much power and security, safe in the fact that nobody else could do what they do, which bred a culture very different from a standard corporate hierarchy. Staff pushed social agendas,
9:54spoke out against their own bosses, worked their own schedules, and demanded recognition for their achievements. You know, a true nightmare scenario. A lot of these companies leaned into this idea when it was convenient for talent acquisition. But it has become pretty clear that the actual executives at the top of these businesses, not so secretly, despised this culture. People like Zuck may have been the CEO of one of the most powerful companies on Earth. But his entitled employees didn't kiss the ground around his feet like they ought to in a more traditional company. The Wall Street Journal ran what might as well be the era's mission statement as a headline. The bosses
10:25are back in charge. Executives saw the layoff wave as a chance to put entitled workers back in their place after the leverage workers built up during the great resignation and the remote work boom. The commentator Ed Zitron went further in a piece bluntly titled tech elite hates labor, arguing that the industry resents having created a pampered class of worker that it now has to cater to and overpay. Even several prominent people within tech have been happy to say the quiet part out loud. Venture capitalist Keith Rabo boy said Meta and Google hired thousands of people to do fake work out of vanity called those workers extraneous and held up Elon Musk's
11:00Twitter purge where roughly 80% of the company was cut as the model the rest of the industry should copy. Venture capitalists and real life Baron from Dune. Mark Andre mocked the remote laptop class. David Saxs called Twitter bloated and overstaffed. All of this meant that when mass layoffs became more commonplace across the industry and these executives saw that their stock price wouldn't take a hit for doing it and in many cases would actually trend upwards, a lot of them were more than happy to dive in, even if the long-term benefits would be questionable. Sure, they might have just ended up passing the same people along to the next company, but it did still
11:34send the message that nobody is irreplaceable. The HR software company Bamboo HR surveyed more than 1500 managers and found that roughly a quarter of sea suite executives openly admitted that they hoped return to office mandates would get people to quit voluntarily. About 20% of HR leaders said their inoff rules were specifically designed to drive attrition. And around 40% of managers said layoffs followed because not enough people left on their own. And again, this was just what they were publicly admitting to. I don't get it. Why are they confessing? They're not confessing. They're bragging. The sad part is it kind of worked. In one survey, the share of workers who said
12:12they would quit over a mandatory return to office collapsed from 51% to just 7% in a single year. Again, this is just one variable. But ultimately, people have now been trained to hold on to their jobs at all cost. It's not great, but even the most vindictive CEO still needs to justify their smiting to shareholders who don't have endless patience for workplace power politics.
12:33So, it's time to learn How many works to find out how they have actually justified this so far and why it's kind of backfiring on them. This video is sponsored by ODO. ODU is the all-in-one business management platform that pulls sales, accounting, inventory, manufacturing, e-commerce, and CRM into a single connected system instead of forcing you to juggle five different tools that do not talk to each other. What makes ODU work is how seamlessly it ties everything together. You only install the apps you actually need and they share data automatically. A sale in your storefront updates inventory and accounting in real time, so you do not have to manually update your different
13:07systems, and you get to stay more organized. We use ODO ourselves. Our works library runs entirely on ODO, and having everything in one dashboard has made the day-to-day so much smoother. It is also intuitive and highly customizable, which makes it a good fit for small to mediumsiz businesses that have limited technical expertise and want software that adapts to how they work instead of the other way around. And because ODU automates the routine stuff, invoicing, restock alerts, follow-ups, you spend less time doing admin or stitching tools together, and more time actually running your business. If you want to see how ODU can streamline your own business, you can
13:42book a free meeting with an ODO expert or try the full platform free for 15 days with no credit card required or even keep a single app free for life. Check it out using my link in the description. So big tech bosses with bruised egos finally getting one up on the people who actually built their technology in the first place is nice and all, but outside of their emotional support podcasts, this kind of justification isn't going to fly. Expensive and financially damaging layoffs still need to hold up to the scrutiny from shareholders. But well, if anything, they have been right on board, largely because of the equity math. A huge chunk of tech compensation is paid in
14:17restricted stock units or RSUs. Grants of company shares that drip out to an employee over about four years depending on the exact agreement, but only as long as they stick around long enough to collect them. This has been a great deal for these companies because it allows them to pay worldclass engineers worldclass compensation with shares it can print rather than money it has to earn. And it also means that staff are compelled to stay with their company to actually receive shares they have technically already earned. But printing shares creates its own problem. Every RSU that eventually vests is a brand new share, and every brand new share makes everyone else's slice of the company
14:53a little thinner. Shareholders are uh not fans of getting thinner. Now, over the last decade, the solution to this has been stock buyback programs largely just to mop this up. The company takes real cash and buys its own shares off the open market to cancel out the new ones being minted for staff. Analysts have estimated that over the last several years, effectively all of Meta's roughly $96 billion of buybacks and around 90% of Google's 156 billion went to sterilizing the dilution from employee stock compensation. So yeah, effectively they were just paying their staff out of their retained earnings in a very roundabout way. Eventually shareholders realized that this
15:30would all work a lot better, at least in the short term, if companies kept doing buybacks while doing less stock creation for their employees. When someone gets cut before their RSUs finish vesting, the unvested shares just disappear. They were never created. So, the dilution never happens, and the company even gets to reverse the expense it had already booked. In normal times, this would mean that more of the buyback money could go to bidding up the price of stock in regular markets.
15:54But well, we are not really in normal times right now. All of the money that was going to buybacks to offset these stock bonuses is now going to data centers instead. Now, I have already spoken endlessly about this spending, so I am not going to waste your time again. But the point is that they can either spend money on stock buybacks, spend money on neutralizing RSUs, or spend money on data centers. They don't have enough to do all three. Google and Meta have already started scaling back their buyback programs. And some of these companies are now issuing debt to fund data center construction or going even further and actually issuing new shares onto public markets.
16:29For existing shareholders, this made them much more amendable to a level of employee churn to avoid employees actually collecting their stock to sell on public markets. Most of the big firms vest RSUs at a flat rate of about 25% per year. So, if you earn $200,000 in stock compensation in a given year, you get $50,000 after the first 12 months, $50,000 12 months after that, and so on. But if you keep working there, those future bonuses keep stacking up. So, even assuming your bonuses stay totally flat, if you quit, you are still walking away from as much as $500,000 in compensation that you will have theoretically already earned. And again,
17:05this is in companies that spread things evenly. Amazon schedule is infamously backloaded at 5, 15, 40, and 40%. So, an Amazon employee cut 2 years in has collected just 20% of the equity they were promised. People have started reading these schedules very carefully for reasons that are probably obvious by now. In a culture of permanent churn, a lot of these RSUs simply never live long enough to vest. Now, if you hypothetically wanted to be a bit cynical, according to company financials, there are currently over a quarter of a trillion dollars in outstanding stock waiting to be accessed by employees in just the top seven tech companies alone. That's also only as of
17:41the date those bonuses were actually granted. If you get awarded 100 shares worth $50,000 and then the price doubles next year before it's actually handed over to you to liquidate, you still get 100 shares. they are just now worth $100,000. Now, of course, in theory, this is a nice way to motivate staff to align them behind a stock price, but it also means that the true value of outstanding stock incoming RSUs is likely to be significantly higher than what has been expensed. Because over the last 3 years, when a majority of that stock was vested, the average value of a mag 7 company has roughly quadrupled. Realistically, that means there is something like half a trillion dollars
18:17worth of stock that will be hitting the market over the next 2 years coming from the employees at these companies. And that's also ignoring the stock coming from this year's slate of mega IPOs. That's a lot of stock hitting a market on top of a potent combination of reduced buybacks, increased capital raises, stretched valuations, and further bonuses going towards newly hired staff in the AI space. Even more simply, more money coming out of the market than going into it. Now, to be fair, these pools of unreleased stockbased compensation are still at record highs thanks to record high prices and huge AI based pay packages, but they would be significantly higher again if they
18:53weren't continuously scraped down over the last 4 years. I know this sounds a little bit tinfoilty, but it's worth keeping in mind that locked up expiration and insiders looking to liquidate their winning positions was one of the key catalysts for the unwinding of the dot bubble. At the very least, it's not unreasonable to think that executives are considering ways to manage this ballooning obligation. Now, if all of that was a lot of boring numbers, well, that's kind of the point. Equity engineering is boring, which is why most of the headlines have been focused on the third major justification, which is the AI tools that all of these shenanigans were going
19:26to finance in the first place. AI was supposed to make the whole layoff era make sense. There are naturally a lot of headlines when a major company cuts tens of thousands of workers, but usually a lot less attention when they hire back just as many to cover over shortcomings, but people are now starting to notice. And well, it's kind of getting a little bit embarrassing. I mentioned Ford earlier because they leaned hard on AIdriven quality inspection. Now, on paper, this is actually a perfect use case for pattern recognition software because inspection work is repetitive, datari, and hopefully consistent. But well, to nobody's surprise, according to their
20:01chief operating officer, Kumar Ghotra, the systems were not getting the desired results while quality failures were costing the company billions. So Ford went out and rehired roughly 350 of their veteran Graveybeard engineers it had let go and put them back on the line. The CEO of Coinbase, Brian Armstrong, apparently went full founder mode and reportedly fired engineers who didn't on board the company's AI coding tools within a week. According to their own claims, somewhere between a third and 40% of Coinbase's code is now written by AI. Unfortunately, they announced this a few days before their entire platform went offline for several hours due to technical issues. And then just last week, literally as I was writing this video,
20:40Coinbase's AI powered prediction market alerts pushed breaking news to millions of users, announcing that Norway had beaten Brazil 3 to2 in the World Cup before the match had even kicked off. Now, putting aside the horrors of having a prediction market on top of a supposedly respectable investment platform, this is obviously not a great look. Companies mess things up all the time, but customers and even investors are a lot less forgiving of those mistakes when leadership has made a big deal about getting rid of people that were there to avoid them in the first place. More broadly, a report by Robert Half found that roughly one in three managers who
21:12cut a role citing AI has already rehired for the same or similar position. Org view found that 55% of leaders who made AIdriven cuts now admit it was a mistake and Gardner projects that by 2027 half of all AI blamed layoffs will have rehired the role under a new job title. Forester found roughly the same with one extra twist. A lot of the rehires are coming from offshore at lower pay. So the original workers still lost. And then of course there is the part we already knew. According to a rumé.org or survey, 59% of companies that cited AI is the reason for layoffs admitted they emphasized AI's role because it plays better with stakeholders than admitting
21:50to financial constraints. It sounds bad, but it gets worse. This whole game has created almost a cult-like culture around a really strange idea, particularly in the tech industry. So, go and watch this video next to find out why you are apparently going to spend the rest of your days in the permanent underclass. And don't forget to like and subscribe to keep on learning how money works.
Works on any YouTube video.
Swap youtube.com for 2outube.com in the address bar — or paste a link.
Read another video