Is the AI Boom a Productivity Revolution or the Most Expensive Hype Cycle in History?
Nvidia at $3 trillion. Hyperscalers spending $300bn on AI infrastructure. Strong jobs killing rate-cut hopes. Here's how to tell a genuine transformation from a bubble.
|
Finance Explained Simply The Deep Dive Full analysis · Week of 7 June 2026 · 12 min read
|
|||||||||||||
|
In this issue
| |||||||||||||
|
01 — The Jobs Report 245,000 Jobs in May. Wages Up 4.1%. The Fed Just Got Its Rate-Cut Hopes Pushed Back Again The US economy added 245,000 jobs in May — well above the 180,000 economists had forecast. The unemployment rate held at 3.9%. Average hourly earnings rose 4.1% year-on-year. By every metric, the labour market is running far too hot for a central bank trying to bring inflation down to 2%. The mechanism is straightforward. Strong employment means strong consumer spending. Strong spending means businesses can raise prices. Businesses raising prices means inflation. The Fed cannot cut rates while the labour market is adding quarter-of-a-million jobs a month — doing so would pour fuel on an already hot fire. The labour market’s resilience has surprised almost every forecaster. When the Fed began raising rates in 2022, most economists predicted unemployment would hit 5–6% by 2024. It never did. The US economy absorbed the fastest rate-hiking cycle in four decades without a meaningful rise in joblessness — a genuine historical anomaly.
| |||||||||||||
|
02 — Nvidia and the AI Trade Nvidia Just Crossed $3 Trillion in Market Value. What Does That Number Actually Mean? Nvidia’s market capitalisation briefly touched $3.1 trillion this week, making it one of the three most valuable companies ever, alongside Apple and Microsoft. The company’s revenue has grown from $27 billion in fiscal 2023 to a projected $145 billion in fiscal 2026 — driven almost entirely by demand for its H100 and Blackwell AI accelerator chips. At $3 trillion, Nvidia trades at roughly 35x forward earnings — expensive, but not absurdly so for a company growing revenues at 90%+ annually. The market is pricing in continued AI infrastructure investment from hyperscalers (Microsoft Azure, Amazon AWS, Google Cloud) who collectively plan to spend over $300 billion on capital expenditure in 2026, much of it on AI compute.
| |||||||||||||
|
03 — Revolution or Bubble? How to Tell Whether AI Is Genuinely Changing the Economy — or Just Changing Stock Prices The honest answer is: both, but unevenly. The infrastructure build-out is unambiguously real — billions are being spent on data centres, power grids, and chips, generating genuine revenue for Nvidia, TSMC, and the hyperscalers. What is less proven is whether AI will translate into broad productivity gains across the economy. Productivity data — the measure of output per hour worked — has so far shown only modest acceleration. History offers a useful frame: the internet boom of the late 1990s saw massive infrastructure investment (fibre, servers, routers) that destroyed enormous amounts of stock market capital when the bubble burst — but the underlying infrastructure proved enormously productive for the decade that followed. The lesson: the technology can be transformative and the stocks can still be overvalued simultaneously. The signals to watch: corporate profit margins in sectors adopting AI (law, finance, software development), labour productivity figures from national statistics offices, and whether hyperscaler capex ever starts to generate commensurate revenue from AI services. The case for AI as revolution rests on those numbers delivering. | |||||||||||||
|
◆ Concept of the Week Price-to-Earnings Ratio — The Single Number Every Investor Watches A P/E ratio tells you how much investors are paying for each £1 or $1 of a company’s annual earnings. If a company earns $10 per share and its stock trades at $300, the P/E is 30x. A higher P/E implies investors expect faster earnings growth — they’re paying a premium for the future. A lower P/E implies more modest growth expectations or higher uncertainty. The S&P 500’s current P/E of ~28x is high by historical standards (the long-run average is ~16–18x), which is why many investors argue markets are pricing in a lot of future growth that may not materialise. | |||||||||||||
|
● — Finance Explained Simply · Hit reply with any questions | |||||||||||||