Meta: Compute Eats Jobs
Meta is cutting roughly 16,000 roles out of 79,000 employees, about 20% of the workforce. Atlassian cut 1,600. The narrative is that AI is replacing humans. That misses what is actually happening.
Neither company has to lay off anyone. Atlassian has strong free cash flow. Meta's operating margins sit in the 40s. These are not distress moves.
The real driver at Meta: you spent tens of billions on compute infrastructure. The depreciation is coming due. You do not have the operating cash flow to support both Nvidia hardware and existing headcount. Compute eats jobs. That is the actual equation.
Four types of cuts are happening across tech:
- Over-hiring corrections: Companies that added headcount in 2021 and never justified the efficiency metrics, now using AI as cover.
- Growth-to-profitability pivots: Revenue growth dropped from 20% to 2%. Wall Street wants either growth or profitability. If you cannot deliver growth, cut to margins.
- Genuine AI displacement: Coding roles where AI tooling has replaced functions that previously required humans.
- Capex trade-offs: The Meta scenario. Infrastructure spend forces workforce reductions.
The fifth category matters most for sales teams: talent reshuffling. Companies are not shrinking to save costs. They are cutting 20 mid-tier engineers to hire 8 elite ones at double the salary. They need people who can deploy AI tools, not talk about them.
The AI Fluency Test
One question tells you if someone is actually AI fluent: What commercial AI tool did you deploy in your organisation this month?
Not what they read about. Not what demos they watched. What did they buy, configure, and put in front of their team in the last 30 days. Anyone on the bleeding edge has done this repeatedly.
Even at top startups, maybe 30% of management teams meet this standard. The role that matters now is not "prompt engineer" (that job lasted about a year). It is agentic deployment expert: someone who can evaluate, deploy, and scale AI tools across teams.
NVIDIA: $1 Trillion and a Shrug
NVIDIA CEO Jensen Huang announced a $1 trillion revenue forecast at GTC. The stock moved less than 1%.
Why the flat reaction? NVIDIA did $215 billion last fiscal year. Analyst consensus for next year already sat in the mid-$300 billions. The "trillion dollar" headline was two years of consensus estimates rounded up for the stage. The market processed it in ten minutes.
What the forecast confirms: capex investment at these levels continues for four to five more years. When NVIDIA hits $600 billion in revenue, global capex spend supporting it likely exceeds $1.2 trillion.
The tension: token consumption may grow 3,000x over five years, but if price per token falls 6x simultaneously, revenue growth is not linear. The bull case requires demand to outrun price compression at massive scale. So far, every data point says it is. But inference prices keep falling. Customers are building on Google TPUs and Amazon chips. The probability of something breaking over five years sits around 30%.
Not enough to bet against NVIDIA. But worth naming as a bet.
What This Means for Sales Teams
If you are hiring, test for deployment fluency, not buzzword fluency. If you are being hired, show what you shipped, not what you read.
If your company is cutting 20% of headcount while booking record revenue, it is not about AI replacing you. It is about capex priorities or talent restocking. Know which one you are in.