Amazon Can't Decide Why It's Cutting Jobs, and That's the Story
Amazon confirmed on October 7 that it had cut fewer than 1,000 white-collar jobs, concentrated in the Stores division that runs its main e-commerce business — customer service, selling-partner support and retail engineering roles in the US, UK and India. The timing was awkward: the cuts landed during Prime Big Deal Days, one of the company's biggest shopping events of the year. The company's own explanation was characteristically bland. A spokesperson said Amazon had adjusted parts of the Stores business because it believed the new structure would better enable it to deliver on its priorities.
Two Executives, Two Explanations
What makes this round worth a second look isn't the headcount — a thousand jobs is a rounding error next to the roughly 30,000 corporate positions Amazon has eliminated since October 2025. It's that two of the company's most senior voices offered contradictory accounts of why the cuts keep coming. Jeff Bezos, in an interview with Fox News' Bret Baier, pinned the blame on pandemic-era excess, saying the company had simply overhired and 'really grew our head count.' There's no AI in that explanation at all — it's a story about correcting a hiring binge, not about machines doing the work.
That account sits awkwardly next to what Andy Jassy told shareholders in June 2025, when he wrote that generative AI's efficiency gains meant Amazon expected, in his words, that 'in the next few years, we expect that this will reduce our total corporate workforce.' Jassy's statement was explicit and forward-looking. Bezos's is retrospective and AI-free. Both men are describing the same trend line of shrinking corporate headcount, and they're not telling the same story about it.
The CFO's Version, Nine Days Earlier
The contradiction is sharper because of what AWS chief financial officer John Felton said just days before the Stores cuts were confirmed. In an Axios interview published October 16, Felton made the case that Amazon's AI infrastructure spending is not just affordable but already paying for itself, citing a nearly $500 billion order backlog and server investments recouping their cost in under three years.
We're seeing paybacks of under three years on those servers.
Felton also described AI internally as 'an amazing muck reduction tool' that frees staff to 'reinvest their time into more value-added activities' — a framing that is explicitly about doing more work with fewer people, even as he was making the investor-facing case for AI capex rather than talking about layoffs. Put the two interviews side by side and Amazon is simultaneously telling Wall Street that AI is driving internal efficiency gains and telling the public that its job cuts have nothing to do with AI.
Why the Framing Matters More Than the Number
For decision-makers watching the AI-labor story, the headcount itself is almost beside the point. What matters is that Amazon is running two separate narratives for two separate audiences, and it will have to reconcile them in the open on October 29, when the company reports third-quarter earnings. Investors will be scrutinizing roughly $220 billion in planned 2026 capital expenditure and a trailing-twelve-month free cash flow that has turned negative by $7.6 billion, even as AWS revenue climbed 37% to $42.2 billion last quarter. A CFO who argues AI spending pays for itself in under three years is implicitly arguing that AI is already changing how much labor the business needs — which is a harder message to square with 'we just overhired in 2021.'
There's a broader bifurcation underneath both stories: Amazon's total global workforce was still up 3% year over year as of June 30, at roughly 1.595 million employees, driven by warehouse and logistics hiring even as corporate ranks have been cut by tens of thousands. That split — AI-adjacent efficiency rhetoric at the white-collar layer, continued physical-labor growth at the operational layer — is the more durable signal for anyone trying to read what AI is actually doing to Amazon's org chart, regardless of which executive's talking points get repeated in the next earnings call.