27JUL

Amazon cuts inside its own AGI unit

1 min read

10:02UTC

Amazon eliminated roles in model customisation and post-training inside its AGI unit on 22 July, disclosing no number, while committing roughly USD200 billion to AI infrastructure.

← Back to SAP freezes R&D headcount as others deny AI Jump to analysis ↓

EconomicDeveloping

Key takeaway

Amazon cut model-builders inside its AGI unit while committing USD200 billion to AI infrastructure.

Amazon made cuts on Wednesday 22 July inside its Artificial General Intelligence (AGI) unit, the internal frontier-research organisation rather than a customer-facing product team ¹. The roles eliminated sat in model customisation and post-training, two of the unit's research Teams. Amazon disclosed no number and framed the round as sharpening focus on what matters most to customers. Affected US staff receive 90 days of pay and benefits, outplacement support and transitional cover.

Post-training is the stage where a trained model is shaped for a specific task or customer, sitting between research and deployment. Cutting there while committing roughly USD200 billion to AI infrastructure suggests Amazon expects the next dollar of return from shipping what it already has rather than from pushing the frontier. January's round took 16,000 corporate roles across the company. This one takes the people who build the models.

Amazon published no headcount figure, and no regulatory filing will supply one. ServiceNow cut hundreds in June citing AI efficiencies without triggering a notice either . The US Worker Adjustment and Retraining Notification Act, known as the WARN Act, requires 60 days' warning of a mass layoff at a single site, a threshold that rounds of this shape, spread thinly across research Teams, rarely reach. Whatever Amazon cut this week will not surface in the American statistics Congress reads.

Deep Analysis

In plain English

Amazon has a research group working on advanced AI, called the AGI unit. On 22 July, it cut jobs inside that group specifically, in teams that fine-tune and adjust AI models after they are built. Amazon has not said how many people lost their jobs. At the same time, the company is spending about USD200 billion on AI infrastructure like data centres and chips. So even a company investing heavily in AI is cutting some of the specific jobs inside its own AI division.

Deep Analysis

Root Causes

Post-training and model-customisation teams scale with the number of distinct model variants a lab must fine-tune.

As Amazon consolidates its AGI unit around fewer, more general foundation models rather than many narrow ones, the headcount needed for that specific stage of the pipeline shrinks even as the company's total AI infrastructure spending rises.

What could happen next?

Meaning
Amazon cut jobs inside its own AGI research unit's post-training teams while committing roughly USD200 billion to AI infrastructure, showing AI investment and AI-team job cuts happening at the same company simultaneously.

Sources:Outlook Business

Mentions:Amazon →ServiceNow →WARN Act →

First Reported In

Update #18 · SAP freezes R&D headcount as others deny AI

Outlook Business· 27 Jul 2026

Read original →

Different Perspectives

European Commission

The European Commission's draft Annex III guidelines, closed for comment on 23 July, treat algorithmic scoring in recruitment, pay and termination as high-risk regardless of whether a human signs off, echoing Spain's Audiencia Nacional ruling 101/2026 on concealed scheduling algorithms. Brussels is shifting the fight from counting AI job losses to assigning legal liability for the tools themselves.

Office for National Statistics

The Office for National Statistics recorded UK vacancies rising to 712,000 on 21 July, the first quarterly increase this beat has tracked, with payrolled employment down 85,000 on the year against May's 210,000 fall. The bulletin names no AI cause anywhere, and that is the point: nothing in the release confirms the displacement story it gets cited to support.

Christian Klein, SAP

Christian Klein told investors on 23 July that SAP's research headcount will not grow for twelve months because AI agents and their token costs are absorbing the work, not because SAP is cutting jobs. He frames it as commercial arithmetic: the cost of AI-assisted coding tokens plus the salaries specialist AI hires command, not people being replaced by machines.

Betsey Stevenson, University of Michigan

Betsey Stevenson argued that the 187,000 jobless-claims reading describes a market that hires little and fires little, not one AI is emptying. She said the real damage hides in eligibility rules and suppressed job postings, not in the headline layoff counts employers keep denying.

Comisiones Obreras, UGT and Concentrix's A Coruña works committee

Comisiones Obreras, UGT and Concentrix's A Coruña works committee blamed Microsoft's push toward AI self-service for the 80 redundancies unions signed off on 22 July, not unavoidable business cause. A second Coruña procedure covering 80 more jobs runs to a 31 August deadline, and the unions want the state, not the employer, setting the pace of AI-driven cuts.

Stanford's 'We Must Act Now' signatories

More than 200 academics, including 16 Nobel laureates, published a 13 July letter warning of AI-driven labour disruption, citing Daron Acemoglu's NBER estimate that AI's total factor productivity gain stays under 0.66% over ten years. The letter's own cited economics sit well below Goldman Sachs Research's 1.5-percentage-point estimate published the same week.