JAS
← All insights
· 9 min readAI LayoffsAgency BusinessAustralian TechHuman Cost of AI

WiseTech's "$100 vs $2" Moment: What Happened After a Sydney CEO Said Humans Are Too Expensive

WiseTech founder Richard White told investors that paying $100 for human labour when AI costs $2 is "stupid." The company then cut 2,000 jobs. One employee threatened the CEO's family. The human cost of the AI transition is becoming visible in ways that no executive memo anticipated.

In February 2026, Richard White, founder and executive chairman of WiseTech Global, a $20 billion ASX-listed Sydney logistics software company, stood at a Macquarie Bank investment conference and made a calculation visible that most executives only run in spreadsheets.

"It doesn't take much effort to convince people, in the end, that they're stupid to be paying $100 for labour when you can pay $2 for the AI."

WiseTech then cut 2,000 jobs: 30% of its workforce. Shares jumped 11% on announcement day.

590 employees signed a petition calling for fair redundancy packages. The company extended consultation deadlines and then ignored them. Sacked staff were reportedly banned from joining four competitor firms.

On 24 May 2026, White sent an all-staff email: WiseTech's CEO Zubin Appoo had received a handwritten threat of violence. The letter contained personal information about members of his family. Police were called. The Sydney office security was upgraded.

The $100 vs $2 calculation had produced an outcome no investor presentation had modelled.

Why White's Math Is Correct

The $100 vs $2 framing is financially accurate for a significant range of cognitive tasks in 2026. AI tools, Claude, GPT-4o, Gemini, and the specialist tools built on top of them, can produce work that previously required human labour at a fraction of the cost per task.

A piece of structured analytical work that a junior analyst would spend four hours producing can be generated in two minutes. A candidate brief that a recruitment coordinator would spend an hour compiling can be structured and formatted in seconds from raw interview notes. A weekly client report that an account manager would spend three hours composing can be generated from dashboard data automatically.

The ratio of human cost to AI cost for these specific tasks is not $100 to $2 in every case. In some it is $100 to $5. In some it is $100 to $20. In the cases White was describing, software engineering tasks at scale, the ratio genuinely approaches the numbers he cited for routine coding, testing, and documentation work.

The executives who condemned White's language, and there were many, did not contest the math. They contested the framing of humans as interchangeable with AI at the task level. Their argument, correctly, was that the human doing the $100 task is not only doing the task.

What the Spreadsheet Does Not Capture

The standard defence of human labour against AI cost comparison is qualitative: humans bring creativity, judgement, empathy, and relationship skills that AI cannot replicate. This defence is true at the level of abstract description and often unconvincing to a CFO who is looking at a line item.

The more precise defence, and the one that is actually defensible in a finance model, is institutional knowledge as an economic asset.

The human doing the $100 task does not just perform the task. They carry the context of 200 previous versions of the task, including the 40 times it went wrong and what was learned from each failure. They carry the client relationship that prevents the task from being re-opened after it is delivered. They carry the institutional knowledge of why the process is the way it is: information that is not in any system and that the AI prompt cannot access.

This context has measurable economic value. The recruitment consultant who has placed 50 candidates into a specific client over three years does not just find candidates: they pre-screen based on a detailed mental model of what that client's hiring manager actually wants that has never been written down. The AI can search. It cannot access the unwritten model.

The task is $2. The context is not on the spreadsheet. The context is what White's calculation could not see.

The Threat Was Not Random

It is important to understand the context of the WiseTech escalation before reducing it to an extreme individual response. The threat of violence against CEO Zubin Appoo did not emerge from a single employee's instability. It emerged from a specific sequence of events that created the conditions for extreme responses.

White made the "$100 vs $2" statement publicly at an investor conference: the statement that most executives make in private financial modelling was made on the record, in language that the 2,000 people being made redundant could read and interpret as "you are being eliminated because you are stupid to expect $100 when you cost what you cost."

The company then ran a consultation process that 590 employees described as inadequate, extended deadlines without substantive engagement, and reportedly attempted to prevent departing staff from joining competitors: a restraint of trade concern that further inflamed a group of people who had just lost their jobs and were trying to find new ones.

The threat against the CEO was an extreme consequence. The sequence that produced it was a predictable response to a transition process that consistently communicated the following: your skills are worth $2, your consultation is a formality, and your future options may be restricted.

What This Means for Agency Owners Managing AI Transitions

The WiseTech case is an outlier in consequence but a signal in pattern. The pattern, AI cost justification + inadequate transition management + frustrated workforce, is not specific to a $20 billion ASX company. It is present at any scale where an agency is reducing headcount or restructuring roles in response to AI automation.

Three implications for agency owners managing their own AI transitions:

The public framing of cost comparison damages the team, even if the calculation is correct. White's "$100 vs $2" statement was accurate. It was also the most damaging communication decision he made during the transition, because it framed the people being made redundant as economically irrational rather than as contributors whose roles were being changed by technology. The agency owner who tells their team "AI is cheaper than you" will experience the same trust damage at a smaller scale, with lower legal risk but the same retention and morale consequences.

The transition process is as important as the transition decision. The 590-employee petition was not primarily a response to the job cuts. It was a response to a consultation process that employees experienced as inadequate. For agency owners reducing headcount through AI-driven restructure, the quality of the transition process, genuine consultation, adequate notice, realistic redeployment options, outplacement support, determines whether the process produces cooperation or adversarial dynamics.

Making the $100 visible is the alternative to the $2 argument. The most effective response to the "$100 vs $2" calculation is not to argue that the AI is wrong. It is to make the $100 visible: to document specifically what the human contribution contains that the AI output does not, the institutional knowledge, the relationship value, the error-prevention built from experience, and to present that documentation to the decision-maker before they run the comparison.

The Calculation Your Clients Are Running

White made the "$100 vs $2" comparison visible by saying it at a conference. Every client CFO with AI tools available is running the same calculation privately. The difference is that they apply it not to their own workforce, but to their agency spend.

The question is not "should our employees cost $100 when AI costs $2." It is "should this agency cost $15,000 per month when HubSpot Breeze can generate the same content for $500 per month, or when LinkedIn Hiring Assistant can do the same sourcing for $3,000 per seat per year."

The agencies that survive this calculation are not the ones that argue the AI cannot do the work. Some of it, the AI demonstrably can. They are the ones that have made the "$100 visible" for their specific service: the documented evidence of outcomes the AI does not produce, the institutional knowledge of the client's specific context that the AI cannot access from a prompt, and the relationship value that the invoice does not capture.

WiseTech's founder was right about the math. He was wrong about what the math contained. The agencies that thrive in the "$100 vs $2" world are the ones that know exactly what is not in the math, and have built the evidence that makes it visible.