AI & Layoffs: What if Artificial Intelligence Is Just an Excuse?

Well, here we are—tech layoffs are exploding. According to RationalFX, the total number of departures is expected to reach 273,000 by the end of the year.

And while this figure alone doesn't mean much, know this: it represents roughly 10 times the annual volume of pre-COVID layoffs.

So, can we really say that humans are being progressively replaced by AI as so many claim?

In France, INSEE speaks of a contraction in the job market directly linked to the rise of AI.

But correlation doesn't imply causation, so we're entitled to wonder if there's something else hiding behind all the hype.

So I wanted to dig deeper and explore the root causes to understand this wave. And it turns out AI might not be our biggest concern.

Large waves of departures blamed on AI

If you read the latest news, there's plenty to worry about:

• Oracle just laid off 30,000 people (20% of its workforce)
• Block cut 40% of its headcount

And I could've cited Meta, Amazon, Klarna, ASML, Ericsson, Salesforce—the list goes on.

In most cases, AI is cited as one of the reasons. And this narrative has a major advantage because on paper, these companies say: we're automating, we're gaining productivity, and we're cutting fixed costs. Which tends to reassure shareholders.

Block's stock price, for example, recovered a bit in February following the announcements.

Same with Oracle's stock price (announcement made March 30th).

Now doubts linger, and as one article put it: "Isn't this just layoffs with better marketing—AI washing?"

Block is really compensating for past management mistakes

Block is the new name for Square, a payments company you might know from its little payment terminal that's now fairly ubiquitous:

But Block isn't just a payment terminal—it's also companies in crypto because its founder, Jack Dorsey, is a big believer in cryptocurrencies.

Jack Dorsey is also the former founder of Twitter, which he sold to Elon Musk a few years ago.

And Jack tends to think big. Twitter had 8,000 employees when he sold it—a company that now runs with 2,800 people.

At Block, the company tripled its headcount post-COVID. We're talking about a 12,000-employee company that had just 4,000 pre-2020.

Sure, you can understand it by looking at the COVID effect on Block's stock.

Except the return to reality in 2022 hit hard. The company stagnated, and with an explosion in the payroll, things couldn't end well.

So we started seeing performance improvement plans emerge.

Because if you look at the economic fundamentals, as this article does, you realize that Block is far less profitable than its competitors, with gross margins that are half theirs.

Today, AI is mostly a "pretty" way to hide management mistakes and reassure investors.

Oracle isn't cutting jobs for productivity gains

Oracle's case is a bit different. Officially, it's not about cuts driven by productivity gains, but a reorientation of investments toward infrastructure to support AI.

In their case too, the stock price is rather concerning, but it's not the main driver of changes.

As one article puts it, it's primarily about investment:

The job cuts at Oracle come as it has invested heavily in AI, spending both on its own infrastructure and on partnerships with other companies like OpenAI. It plans to spend at least $50bn on infrastructure this year, and it has also raised $50bn in debt in order to "meet demand" for even more AI infrastructure. Oracle is also part of the Stargate initiative, alongside OpenAI, SoftBank and MGX, an AI investment fund backed by US President Donald Trump.

Here, it's really about reorienting capital from a traditional activity that's flagging to one that's supposed to replace it in a few years.

In reality, I won't criticize it. It's a strategy, a bet. A huge bet, but one that falls into the same category as what Kodak should've done when digital arrived. And Oracle doesn't want to be the next Kodak.

The market is anticipating unrealized gains

And that's really the issue—nobody wants to be the next Kodak.

When a leader (like Block, Google, or Meta) lays off 10% of its workforce and its stock goes up the next day, every other company is tempted to do the same.

Laying people off because you mismanaged your company would be an admission of failure. But laying people off because you're "transforming through AI" is a vision of the future.

And this FOMO—fear of missing out—explains a lot of the current departure plans. Gartner calls it "RIFs before reality", the anticipation of unrealized gains:

The employment deal is being rewritten in real time. CEOs are making bold moves based on AI's promise rather than its proven impact. Layoffs linked to AI dominated headlines last year, but Gartner data shows fewer than 1% were due to actual productivity gains.

This anticipation drives investment reorientations. Oracle's case is representative here. Not everyone is investing in infrastructure, but many are reinvesting in engineering to automate other business functions and, most importantly, to be ready for the future.

AI is no longer just a growth story; it's a cost-reduction tool, and firms are restructuring accordingly. What we're witnessing is a shift from headcount-driven expansion to automation-led productivity, a transition that will define the tech sector in the coming years. —Alan Cohen, analyst at RationalFX

Now, this isn't new, and I notice a certain hypocrisy among some developers who are discovering today that their profession has always been about automating others' jobs. It's a shame to discover it when it touches us personally.

Anyway, what's certain is that companies are anticipating cuts without yet having proof of the gains to come. It's not just a few layoffs—we're seeing signs, notably raised by levels.fyi's founder: we're witnessing a simplification of career paths.

A layoff plan is temporary. But when you start eliminating rungs in career ladders, it signals you're anticipating a durable, global reduction in headcount.

And yet, once again, the gains aren't that obvious so far.

Productivity gains aren't widely demonstrated yet

We all have our opinion on this.

I consider myself more productive with AI. But not everyone agrees.

But in any case, these are just opinions.

There are studies on the topic of productivity, but there's no consensus.

You can find studies showing we're less productive, but you can also find others saying the opposite.

The causes are multiple. The first is what's called the productivity paradox:

You can see the computer age everywhere but in the productivity statistics

Yes, back then we wondered if computers really made us more productive. It was far from certain.

This paradox is explained two ways.

First, companies spend more time configuring tools, training people, and reorganizing workflows than actually producing more.

Second, a new technology requires a learning period that can be quite long to master. And that's what we're seeing today—AI usage is totally new. Many are just faster at doing what they did badly before.

And it's not like we know how to measure developer productivity anyway. I'll remind you that this question still hasn't found a universal answer since we started asking it.

Now, I've also heard plenty of CTOs and IT directors privately say they have the means to prove it. But they don't want to. Because proving it would mean making decisions they don't want to make.

And I can tell you that in this period, I'm glad I'm no longer a CTO.

Deeper economic reasons

Still, as we've seen, productivity gains or not, can we really say all current layoffs are AI-related? Probably not.

A recently cited study shows that 59% of HR leaders admit that AI was used as a "cover" to justify budget cuts that were actually driven by:

• Over-hiring post-COVID
• Investor pressure to increase margins
• Internal strategy mistakes

But I think that would be overly reductive.

It's mostly the nth demonstration that we've entered a new era post-COVID. Between rising inflation, ongoing trade wars, endless debates about tariffs, skyrocketing energy costs, various conflicts that paralyze parts of international commerce—we're really in a recession.

AI is a facade to hide the rest. When Trump gets excited about his Stargate project (building datacenters), it's storytelling to hide the mess, even if it's true that AI is probably one of the drivers of the military sector in coming years and the US losing ground on it is probably making them nervous.

Yes, because the worst part is that even on AI, it's not certain the people leading the dance will be American.

Recent Chinese models like Ernie, DeepSeek, Qwen, and Kimi are largely on par with Gemini or ChatGPT, without necessarily costing the same. Kimi and DeepSeek reportedly cost 10% of their American counterparts during training phases.

Which, incidentally, is encouraging but mostly logical—technology improves, and we've never seen tech stay this inefficient over time. The computer that sent a rocket to the moon was less powerful than our smartphone despite consuming far more energy.

And for all these reasons, US companies are in full downsizing mode. Players in AI need to become more competitive. They're investing heavily while cutting payroll at the same time.

Other tech companies are following suit, further constrained by hyper-unfavorable economic conditions and in a context where saying you're laying off to increase productivity is more sellable than admitting reality.

A new era?

And us in the middle of all this? Well...

I'll be honest—I really wondered how to conclude this piece. I always try to end on a positive note, but the exercise is difficult here. I'll try anyway.

Is this the end of an era? Probably the era of unreasonable hyper-growth, which isn't so bad. This forced downsizing might help us get back to basics instead of just chasing vanity metrics (like headcount).

It's also a global economic shift and a US bloc that seems to be faltering. I want to see some positivity in thinking that Europe has cards to play. We're less affected than the US by the recent massive waves of layoffs. Probably because we have less insane payrolls than the US and more solid social models.

While American giants painfully refocus, it's our moment in Europe to catch up. These new technologies, more accessible and efficient, let us move faster with fewer resources. Maybe it's finally time to create real European tech alternatives—more sober and pragmatic.

On that note, you can go back to normal activities.