Job losses at Big Tech aren’t about efficiency, they’re about finding the money to bet on artificial intelligence, says Niall Kitson
Blogs
Oracle’s decision to let go as many as 30,000 employees across its organisation was communicated not with an all-hands meeting or one-to-one with a HR rep, but a bulk e-mail sent at 6am (across the US, India, Canada and Mexico) notifying staff that they were surplus to requirements and that their corporate network access would be revoked that morning.
According to The Tennessean, workers in Oracle’s Nashville office were greeted with the following e-mail: “We are sharing some difficult news regarding your position. After careful consideration of Oracle’s current business needs, we have made the decision to eliminate your role as part of a broader organisational change.” Oracle had great plans for Nashville, where it expected to employ 8,500 by 2031. Not now.
It was a brutal way to treat employees, some of which had been with the company for almost 20 years and terrible PR for anyone considering joining the company. On social media the mood was a mix of shock, rage and grief. In the last quarter Oracle brought in revenue of $14.93 billion, an increase of 12.2% over the past year. Hardware sales increased 2.3% year-on-year to $670 million and services did even better, bringing in $1.35 billion, an increase of 6.8%. And yet.
The lay-offs have been unofficially attributed to productivity gains powered by AI, however the picture is more complicated than ‘bad robot took my job’. From this remove I can see three justifiable arguments covering the perspectives of business organisation, personal productivity, and market conditions.
The argument from management
In a recent edition of the 20VC podcast rock star venture capitalist Mark Adreessen said AI was a “silver bullet excuse” for reducing headcount. He expanded on this point to say that most companies were overstaffed thanks to additional hiring that took place during the pandemic. The trajectory of software engineering hires has been consistently downwards as global digital transformation projects were completed. According to figures from Indeed US demand for the number of job listings for software engineers has declined by 35% since 2020.
If the teens were defined by the belief that learning to code was the first step on a path to a well-paid job, the 20s has been about recognising that path is actually a rope bridge and safe passage was not guaranteed.
When Big Tech started announcing lay-offs as the Covid pandemic wound down I expected sales and customer service teams to be downsized, offshored or let go en masse. For engineers to be included was a sign that innovation and product had been superceded by restructuring efforts.
Meta was first to the party in 2022 by cutting 11,000 jobs before its ‘year of efficiency’ where a further 10,000 jobs were lost and a hiring freeze put into effect. At time of writing the company has cut 25,000 jobs. The website Tech Layoffs Tracker provides a stark picture of the industry, where three of the past four years saw the loss of an average 245,000 jobs and a further 342,000 to go by the end of this year. 2023 saw the most jobs cut, with 430,000 employees let go. The figures are alarming but perspective is needed. As one former long-term Oracle employee told The Register: “It’s not coming for the entire American economy. It will come for the big guys.”
To mix my metaphors, AI is a boogie man but Covid created the space under your bed that he lives in.
The argument from productivity
Stop me if you’ve heard this before: “Your job is not under threat from AI, it is under threat from someone who knows how to use AI”. I have an agent nearby programmed who can reming you on the hour every hour in a local accent.
From an Irish perspective there are plenty of reports with local data and they’re not encouraging. According to research firm Saros, organisations are spending an average of €770,000 on AI projects that have delivered no value. This is not for lack of trying. Figures from IDC show Ireland is an early AI adopter, with 20% of organisations ranked as ‘optimised’ compared to an international average of 10%.
Recruitment firm Stelfox puts AI usage in Irish companies for 2025 at 91%, almost doubling in a year. Furthermore, some 76% of organisations polled had upskilled their staff in the use of AI over the same period. Make no mistake, Ireland likes AI, but it’s just not paying.
From what I can see we are have people doing more of things that either don’t add value or are flawed to begin with. This is AI exposing weak strategy, while closing the temporal space needed to reflect and iterate.
A side-effect of early AI adoption is a degradation in core software development skills that are being automated away from entry-level positions. As hiring moves away from assessing qualifications to lived experience and examples of where candidates can draw on experience to deliver value. A study from IBM put the lifetime value of a software engineering degree at about 25 months. An university lecturer and proponent of microcredentials I spoke with put it at 18 months. Degrees are coming with a sell-by date and the jobs market is flush with experienced developers. As Oracle, Google, Microsoft and Meta keep downsizing the jobs market is settling into a new pattern of top talent being hoarded by MNCs on one level followed by a fat middle of battle-hardened professionals with experience at multinationals, and diminishing prospects at entry level. The differentiator isn’t AI, they’re raw talent and lived experience.
The argument from the markets
Last month Block (owner of Square) cut 40% of its workforce in a pivot to AI and saw it’s share price go up 20%. Oracle’s went up 6% on foot of its announcement. Investors love lean, efficient organisations. But Oracle’s share price has to be seen in the context of a wider play. Trading today at $146.24, last September it was at $328.15 – that’s a 57% tumble. The downturn can in part be attributed to Oracle’s involvement in the ambitious Stargate data centre joint venture between OpenAI, Softbank and MGX. The $500 billion project has been criticised by Elon Musk who posted on X that the project wasn’t close to being funded, and that Softbank had secured “less than $10 billion”.
We got further signals last week that OpenAI has been struggling – notably the closure of its Sora video creation tool, its decision not to buy massive amounts of RAM (40% of the global supply for 2026, but who is counting) from Samsung and SK Hynix, and its walking away from a data centre construction project in Texas for which Oracle was a partner. And yet as I write this OpenAI has closed a $100 billion funding round, valuing the company at more than $850 billion. Clearly, the market believes in AI, even if no one has made any money out of it yet.
Did AI cost 30,000 Oracle staff their jobs? Yes, but not in the way you’d think. Raising cash for projects by slashing operational expenditure is a risky move, especially when no one has figured out how to make AI pay for itself. But this is the plan and it’s industry wide. That’s some bet.