Every business that’s seeking to become AI-centric is having to jettison lower-performing employees and projects. (Image: Getty)
Big Tech is in layoff mode. While the latest wave of redundancies at Meta, Microsoft and others reflects a broader post-pandemic correction, artificial intelligence is a major influencing factor.
Big tech leaders keep talking up AI's ability to replace swathes of workers. But the reality is that the technology has not led to a collapse in the software developer workforce, where its effects are being most immediately felt.
Developers are now typically handing off 30-45% of their coding tasks to AI agents, thereby speeding up development cycles.
Roles aren’t being eliminated but are being refactored, with a slowdown in entry-level hiring attributed to general cost-cutting rather than to AI replacing humans.
However, the latest wave of general cost-cutting is a direct result of the colossal spending spree on AI as tech companies invest in AI infrastructure, including massive GPU clusters, data centres, AI talent and research and development.
Slashing on the altar of AI
Every business that’s seeking to become AI-centric is having to jettison lower-performing employees, projects, and even entire divisions to accommodate the pivot to AI.
Shareholders generally like a redundancy round dressed up as an efficiency drive.
But there has been investor backlash against the unprecedented scale of the AI-related spending, which isn’t delivering returns as quickly as expected.
It means boards are now insisting that further AI spending be matched with visible efficiency gains and headcount discipline, essentially an austerity drive in parts of the business.
That's why Microsoft is offering voluntary buy‑outs to around 7% of its US workforce and trimming legacy and lower‑growth parts of the business as it prepares to spend an estimated US$110 billion–$120b (NZ$193.6b to NZ$211.2b) on AI infrastructure in the next year.
To give credit to Asha Sharma, the chief executive of Microsoft’s Xbox gaming division, she didn’t blame this week's 3,200 job cuts across its prestigious gaming studios on AI.
"Our business today is not healthy," she wrote in an email to staff on Monday, noting that Xbox is operating at margins 3 to 10 times lower than comparable businesses.
“We must reset Xbox,” Sharma concluded.
AI’s competition for computer chips, memory and other components is nevertheless squeezing margins on Xbox console sales, and recently acquired games studios that might otherwise have been given more time to turn around their performance are instead being sacrificed on the altar of AI.
It’s the same reason Meta recently announced cuts of around 10% of its workforce (about 8,000 jobs) and is closing roughly 6,000 open roles.
It explicitly frames this as a move to “run the company more efficiently” and “offset the other investments we’re making” as it plans AI capex of between US$115b–$135b this year, roughly double the previous year.
New Zealand companies are not investing billions in AI infrastructure. But many are in the process of reorganising to use AI in everything from contact centres to marketing and business development.
The lesson from the US is to treat AI as a long‑term infrastructure and capability bet that must be balanced against realistic cash flows, disciplined governance and a sober view of productivity gains.
If you are proposing to spend significantly on AI, you need to have a compelling case that there’ll be a return on that investment.
There’s also the risk of simply jettisoning underperforming parts of the business to free up cash for AI investment, rather than doing the harder work of fixing the core problems that are making the business uncompetitive.
Avoiding 'AI‑washing'
It’s clear now that using AI as a catch‑all narrative for cost‑cutting and grandiose future promises can backfire once investors and staff realise the productivity gains are patchy at best.
Our own firms have been beset with genuine non-AI reasons for restructuring – spiking energy prices, sluggish demand, and escalating supply chain costs among them.
But they should make conservative claims about AI‑driven savings to maintain trust with employees and shareholders as they ramp up their AI use – the latest data from Anthropic ranks us 5th globally in an index of use of its Claude AI assistant, which is particularly popular as a vibe coding tool.
Big Tech can afford to make multibillion-dollar AI bets funded by its cash flows – just look at the US$80b Meta has invested in its Reality Labs division and metaverse push, with little to show for it. Our companies don’t have such buffers.
Rather than frontload AI capex and then scramble to impose “austerity” elsewhere, they need to phase AI investments in line with observed revenue or productivity gains, using pilots and staged rollouts to de‑risk larger commitments.
Company boards need to pay attention not just to the ethics of AI but also to its economics, set explicit ROI thresholds, require post‑implementation reviews of AI projects, and be prepared to unwind or pause AI initiatives that underperform rather than doubling down to justify sunk costs.
That’s how successful local adopters of AI I’ve studied operate – they prove the investment case for AI in one part of the company and then scale it business-wide, with an accompanying focus on upskilling staff in using AI.
Companies like Meta and Microsoft are not only cutting staff but also reassigning thousands of employees to AI‑related roles, signalling a shift in skill mix rather than pure downsizing. For NZ employers, that suggests an emphasis on upskilling and redeploying existing staff into data, automation and AI‑enabled roles, using layoffs as a last resort rather than the primary funding mechanism for AI.
The global cuts show that when capital is tight, experimentation in peripheral areas is often the first to go. AI’s rise is requiring every part of the business to justify its existence.
That’s a healthy discipline to embrace. But it could also go very wrong if austerity measures damage brand equity and customer relationships.
We may not be investing huge sums in frontier AI models and infrastructure in NZ, but each funding decision our businesses make has an opportunity cost, and AI needs to wash its own face before being used as an excuse to slash and burn.