AI Capex Hits $725B as Companies Slash 100K Jobs
Major corporations have cut 100,000 jobs while pouring $725 billion into AI infrastructure and tooling, signaling a dramatic capital reallocation that will reshape software, media, and creative industries.
The AI economy has entered a paradoxical phase: while corporations announce record layoffs, they simultaneously commit unprecedented sums to artificial intelligence infrastructure. According to recent reporting, companies have collectively cut more than 100,000 jobs in 2025 while pouring an estimated $725 billion into AI capital expenditure. The scale of this reallocation is reshaping not just labor markets but the underlying compute, model, and media infrastructure that powers everything from enterprise automation to generative video.
The $725 Billion AI Capex Wave
The $725 billion figure aggregates spending from hyperscalers — Microsoft, Google, Meta, Amazon — alongside enterprise AI deployments, GPU procurement, and infrastructure buildouts. Microsoft alone has guided to roughly $80 billion in AI-related capex this fiscal year, with Meta projecting north of $60 billion. Google and Amazon have each signaled spending in the $75–100 billion range. The bulk of this capital flows toward Nvidia GPUs, custom silicon (TPUs, Trainium, MTIA), data centers, and the power infrastructure needed to operate them.
This isn't speculative R&D spending. It's hard infrastructure that must generate returns. And to fund it, companies are restructuring their cost bases — which means headcount.
Where the Job Cuts Are Landing
The 100,000+ layoffs span Microsoft (roughly 15,000), Meta (thousands across performance-based cuts), Amazon, Google, Intel, Salesforce, and a long tail of mid-cap tech firms. Salesforce's Marc Benioff has been explicit: AI agents are absorbing customer support and routine engineering work. Microsoft has openly tied workforce reductions to AI productivity gains. The pattern is consistent — cuts concentrate in middle management, customer support, content moderation, and junior software engineering, the exact roles where LLMs and agentic systems are demonstrating viable substitution.
Implications for Synthetic Media and Creative AI
For the AI video, deepfake detection, and synthetic media space, this capital wave has several direct consequences. First, the GPU supply that trains frontier video models — Sora, Veo, Runway Gen-4, Kling — is being expanded by hyperscaler capex. More compute means larger, more capable video diffusion and autoregressive models, longer coherent outputs, and dramatically lower per-second generation costs over the next 18 months.
Second, the labor displacement story is moving into creative industries. Animation studios, VFX houses, voiceover, and stock media are all seeing budget compression as generative tools mature. Netflix's recent disclosure of AI-generated content in production, alongside the broader Hollywood AI debate, sits directly inside this capex-versus-labor dynamic.
The Return-on-Capital Question
The unresolved question is whether $725 billion in annual AI spending can generate proportional revenue. Current AI revenue across the major labs and platforms is estimated at $50–80 billion annually — an order of magnitude below the capex. Analysts at Goldman Sachs, Sequoia, and others have flagged the gap. The bet is that productivity gains, enterprise automation, and new categories (agents, generative video, synthetic data) close the deficit before investors lose patience.
If the bet pays off, expect another wave of capex in 2026 and a continued reorganization of knowledge work. If it doesn't, the correction would hit GPU demand, model training budgets, and the venture funding pipeline that supports companies like Runway, ElevenLabs, Pika, and the long tail of synthetic media startups.
Strategic Takeaways
For operators and investors tracking AI video and authenticity tools, three signals matter: (1) hyperscaler capex commitments through 2026 — these directly determine model capability ceilings; (2) enterprise AI deployment metrics, particularly in media, marketing, and customer-facing video; and (3) the labor-to-automation conversion rate in creative roles, which sets the demand floor for generative tools.
The $725 billion figure isn't just a financial milestone. It's a structural commitment that locks in the trajectory of generative AI — including the synthetic video and audio capabilities that will continue to challenge digital authenticity frameworks for years to come.
Stay informed on AI video and digital authenticity. Follow Skrew AI News.
