Tech Layoffs Hit 94K in 2025 as Microsoft, Intel Lead Cuts

The generative‑AI gold rush is still minting billion‑dollar valuations—but it is also printing redundancy letters. Fresh tallies from Layoffs.fyi and TrueUp put tech‑sector job losses at about 94,000 in the first half of 2025 and in fact, Nigeria’s recently shut down open banking startup Okra features prominently, already surpassing the full‑year total for 2024. That works out to 507 jobs per day—the fastest contraction the industry has ever recorded. 

The Top 5 Job Cuts Hall of fame 

Across 392 separate companies, the common refrain is: cut operating costs, fund AI infrastructure. Gartner estimates that enterprise spend on AI hardware and cloud services will reach US$350 billion this year, up 41 % YoY.

“We’re flattening management layers and reallocating capital to GPU clusters,” Microsoft CEO Satya Nadella said in a memo to staff on 7 July.(microsoft.com)

Microsoft is not alone as AI which has proved an asset to organisations is also expensive to set up and maintain. OpenAI’s computing costs is said to be around $7b and the story is similar across board especially for the big players like Anthropic and Google via its Gemini. AI solves problems but it is expensive, so you can understand if corporations would rather spend on training models that can replace most of their key staff. Here’s a quick overview of some factors driving the AI investments higher and higher. 

1. Hardware sticker shock. Nvidia’s H200 GPU pods now retail for US$4–5 million each; training GPT‑class models can cost upward of US$200 million.
2. Model‑ops over maintenance. CFOs favour lean, specialised “tiger teams” for AI deployment rather than large legacy app squads.
3. Shareholder pressure. After 2024’s rebound, investors want margins > 35 % in cloud and advertising divisions—labour is the easiest lever.

Its not all gloom for tech jobs though. Some jobs are still in demand and will continue to be for the foreseeable future even though I imagine that a few of them will soon be taken by AI by the end of this decade. 

• Prompt engineering & RLHF (Reinforcement Learning from Human Feedback)
• MLOps / model governance
• Cyber‑security for AI pipelines
• Edge‑AI hardware design
• Quantum‑safe cryptography

Recruiters at Andela say demand for mid‑senior ML engineers is “up 23 %” QoQ even as generalist software roles shrink. Might I add to the Andela list that I think Cloud jobs especially at professional levels will still be in demand. AI applications have to be hosted somewhere and while DevOps roles can be performed by AI, actual deployment and configuration management functions may stay human for a while. 

That said, there is need for government intervention across board. Economists warn that without intervention, AI‑driven displacement could widen unemployment gaps—especially in emerging markets where tech hires are a growing share of the formal workforce. Like every other disruptive tech, governments must step in to secure jobs so that we don’t go into a full machine driven world. 

• Nigeria has earmarked ₦300 billion (≈US$200 million) for a Digital Skills & Jobs Fund in its 2025 Supplementary Budget, focusing on cloud, data and AI training vouchers.
• Kenya is piloting tax credits for firms that upskill at least 10 % of staff in AI literacy.
• EU & Singapore models show that public‑private ‘skills passages’—paid sabbaticals linked to accredited courses—reduce time‑to‑reemployment by up to 40 %.

“Governments cannot leave re‑skilling to chance; the pace of AI adoption is outrunning traditional labour‑market programmes,” says UNESCO labour‑economist Aïscha Doucouré.(unesco.org)

What tech guys must do now

1. Audit your skill stack. Map tasks vulnerable to LLM automation and identify complements: data‑quality, safety, domain context.
2. Invest 10–15 hrs/month in structured learning. Top picks: Stanford CS‑329s (ML Safety), Coursera’s Prompt Engineering Specialisation, Azure AI Fundamentals.
3. Build a public portfolio. Ship small but visible projects on GitHub, Hugging Face or Kaggle to demonstrate hands‑on AI chops.
4. Network laterally. Join Slack groups like Weights & Biases Africa or PyLadies Nairobi—referral hires now account for ~40 % of ML roles.
5. Consider sector pivots. Health‑tech, climate‑tech and defence continue to raise capital even as consumer apps cool off.

The AI boom is not a rising tide that lifts all careers. Automation is eliminating routine dev work faster than the industry can create novel roles. Continuous re‑skilling is therefore insurance, not optional enrichment. Tech workers who layer AI‑native skills onto their existing domain knowledge will remain valuable; those who ignore the shift risk becoming ultimate victims of the boom.

At the same time, governments must step up: save‑as‑you‑learn tax breaks, portable training credits and stronger unemployment cushions can make the difference between a short detour and a lost generation of talent.