Somewhere in the math of modern Big Tech, 28,000 people lost their jobs so that servers could take their place. Since January, Amazon, Meta, and Cisco have collectively eliminated roughly that many positions, and each company has pointed to artificial intelligence as a driving reason. Over the same stretch, the three firms have committed to spending a combined total approaching $300 billion on AI infrastructure in 2026, funding data centers, custom chips, and networking hardware at a pace that dwarfs anything the technology industry has ever attempted.
The pattern is hard to miss: the same companies writing the largest capital expenditure checks in corporate history are simultaneously telling thousands of employees their roles are no longer needed. In each case, executives have drawn a direct line between the two, arguing that AI-driven efficiency makes certain jobs obsolete while demanding massive new investment in computing power.
Amazon: 16,000 jobs and counting
Amazon’s reductions are the largest of the three. The company cut approximately 16,000 corporate roles in its latest round, a figure based on company communications and reporting by the Associated Press, though the precise total has not been confirmed in a single public filing. The layoffs hit teams across operations, human resources, and internal tools, areas where Amazon leadership has said AI-powered automation can now handle work that previously required human judgment.
These cuts followed earlier rounds in 2024 and late 2023 that had already trimmed thousands of roles, making the cumulative reduction one of the deepest sustained workforce contractions in Amazon’s history outside of seasonal warehouse adjustments. CEO Andy Jassy has repeatedly framed the restructuring as necessary to fund the company’s generative AI ambitions through Amazon Web Services, which competes directly with Microsoft Azure and Google Cloud for enterprise AI workloads.
Amazon has not disclosed a specific capital expenditure figure exclusively for AI in calendar year 2026, but the company signaled in late 2025 that cloud and AI infrastructure spending would exceed the roughly $100 billion it deployed the prior year. Analysts at Morgan Stanley and Evercore have estimated Amazon’s 2026 capex could land between $130 billion and $160 billion, with the vast majority directed at data centers and the custom Trainium chips that power its AI training clusters.
Meta: record spending, quieter cuts
Meta’s investment plans are the most precisely documented of the three, anchored by regulatory filings that carry legal liability for material misstatement. In its first-quarter 2026 earnings disclosure filed with the SEC, Meta raised its full-year capital expenditure guidance to between $125 billion and $145 billion. That range is devoted overwhelmingly to AI data centers, custom silicon, and the computing clusters required to train and deploy the company’s Llama family of large language models.
The workforce picture is murkier. Meta confirmed layoffs earlier this year, with reporting from Bloomberg and The Verge placing the figure at roughly 5,000 positions when combining a “lowest performers” culling announced in early 2025 with additional reductions in subsequent months. No single SEC filing specifies a precise headcount or contains a direct executive statement attributing cuts to AI by name. Internal communications used broader language about “efficiency” and “reprioritization,” making it difficult to isolate AI-related restructuring from other cost discipline.
The strategic direction, though, is unmistakable. CEO Mark Zuckerberg told investors during the company’s April 2026 earnings call that Meta’s “single largest investment theme” is AI infrastructure, and that AI-generated content recommendations and advertising tools would drive the revenue growth to justify the spending. The gap between that optimism and the thousands of workers no longer on the payroll has drawn scrutiny from labor advocates and some shareholders who question whether efficiency gains are being shared or simply extracted.
Cisco: 7,000 roles cut and a new restructuring wave
Cisco’s approach is the most explicitly framed as a trade-off. In a Form 8-K filed with the SEC in May 2026, the company described a restructuring plan designed to enable investment in key growth areas including silicon, optics, security, and artificial intelligence. The filing estimates up to $1 billion in pre-tax restructuring charges, split between Cisco’s fiscal fourth quarter of 2026 and fiscal year 2027.
Unlike Amazon, Cisco’s filing quantifies the financial cost of restructuring rather than the headcount. The company carried out a separate reduction of approximately 7,000 roles in late 2024 and early 2025, as disclosed in its SEC filings from that period. Those 7,000 cuts account for Cisco’s share of the 28,000 total referenced in the headline, with Amazon contributing roughly 16,000 and Meta roughly 5,000. The new restructuring charges suggest a continuation of that pattern and could affect several thousand additional employees, though Cisco has not confirmed a figure for the latest round.
In an accompanying earnings press release, Cisco highlighted strong AI-infrastructure order momentum in its fiscal third quarter, emphasizing customer demand for networking equipment capable of supporting large-scale AI workloads. CEO Chuck Robbins told analysts that the company is “reshaping our portfolio and our workforce” to capture what he described as a generational shift in enterprise networking driven by AI.
The $300 billion question
Adding up the three companies’ AI-related capital plans produces a number in the neighborhood of $300 billion for 2026, though the total requires some estimation. Meta’s guidance alone accounts for $125 billion to $145 billion. Amazon’s projected spending, based on analyst consensus and the company’s own directional signals, likely falls in the $130 billion to $160 billion range. Cisco’s capital expenditures are far smaller in absolute terms, typically in the low single-digit billions, but the company’s restructuring spending and its role as a key supplier to other firms’ AI buildouts place it squarely in the same investment cycle.
No single regulatory document confirms a combined $300 billion figure, and the number involves assumptions about fiscal-year alignment and how much of each company’s budget is truly AI-specific versus general infrastructure. But the order of magnitude is consistent with what each company has disclosed individually. For context, the five largest U.S. tech companies are collectively on track to spend more on capital expenditures in 2026 than many mid-sized national economies produce in a year.
Wall Street, for its part, has largely rewarded the strategy. Amazon and Meta shares have held up well through the layoff announcements, with investors treating the combination of cost cuts and AI investment as a sign of fiscal discipline rather than distress. That market reaction underscores a tension at the heart of the story: what looks like strategic clarity from a shareholder’s perspective looks very different from inside a team that just got dissolved.
What workers are actually facing
For the 28,000 people affected, the corporate framing of “reallocation” offers limited comfort. None of the three companies has published detailed hiring targets for AI-related positions that would numerically offset the reductions announced so far. Amazon has posted AI and machine learning roles on its careers page, and Meta has described plans to hire engineers for its AI infrastructure teams, but neither company has provided a figure that allows a direct comparison with the jobs lost.
The skills gap compounds the problem. Many of the eliminated positions were in corporate functions like program management, recruiting, and internal communications. Those roles do not translate directly into the machine learning engineering and data center operations jobs that AI investment creates. Retraining programs exist at all three companies, but their scale and effectiveness remain largely unaudited by outside observers.
Labor economists have noted that previous waves of automation-driven layoffs in sectors like manufacturing and telecommunications did eventually produce new job categories, but often at different companies, in different cities, and at different pay levels than the roles that disappeared. Whether the AI cycle follows that pattern or breaks it is one of the defining workforce questions of the next several years. So far, neither Congress nor federal agencies have introduced binding disclosure requirements that would force companies to report how AI adoption affects headcount.
Enormous checks, real losses, and no receipts
As of mid-2026, the verified record shows three things clearly. Amazon, Meta, and Cisco are spending at historically unprecedented levels on AI infrastructure. All three have cut thousands of workers and explicitly connected those cuts to AI-driven efficiency or the need to fund AI investment. And none has provided the kind of detailed, auditable disclosure that would let workers, investors, or policymakers trace a specific dollar of AI spending to a specific job created or destroyed.
That gap between the scale of the investment and the transparency of its human consequences is not a minor accounting detail. It shapes how workers evaluate their own career risk, how policymakers decide whether to intervene, and how investors assess whether these companies are building durable businesses or simply substituting labor costs for capital costs in ways that may not hold up over time.
The checks for servers and chips are enormous and verifiable. The job losses are real and documented. The promise that AI will create as many opportunities as it eliminates remains, for now, an assertion without a receipt.
