Deep Dive · by Xiaohu

a16z's 14 charts make the case that AI investment isn't a bubble

Along the way, they debunk three widely misread metrics: the falling token chart, the six-month order backlog, and that 5% entry-level job stat.

The 60-second take
  • Vertiv, which makes data center cooling and power gear, fell $76 million short of its delivery targets. That got spun as a sign of an AI bust, so a16z checked public data against that narrative three ways.
  • Their conclusion: demand is unusually high, but the data can't say whether supply chains are also straining. Along the way, they unpack three widely misread numbers.
  • One of those — the viral "AI demand is collapsing" curve — isn't actually measuring demand at all.
⚑ This piece draws from a16z's weekly data column, Charts of the Week (July 31, 2026, by Moses Sternstein). The raw data for the 14 charts comes from the U.S. Census Bureau, the New York Fed, Indeed Hiring Lab, ADP Research, SiliconData, YipitData, Kalshi, and Capital IQ / LSEG Datastream / Yardeni Research. ⚠️ a16z is an investor with a stake in this space, and this entire issue argues that AI infrastructure demand hasn't cooled. It also flags three limitations in its captions: the numbers on two GPU charts were visually reconstructed by a16z from others' published charts; OpenRouter's spending data covers only a slice of overall demand and skews toward U.S. SMBs; and the backlog ratio and unfilled orders are two different statistical measures. This article also flags five spots where the text disagrees with its own charts; I go with the charts and note each one.
Infrastructure

AI compute and hardware demand is red-hot, but supply chains are strained

a16z (Andreessen Horowitz) just published a new issue of Charts of the Week, its weekly data column, with 14 charts, all drawn from public data. The process is the interesting part. Everyone's arguing about whether AI is a bubble, and most arguments rest on stock prices and vibes. This issue instead walks through Census Bureau, NY Fed, Indeed, and ADP numbers to check three specific questions: Is AI infrastructure demand fading? Are entry-level jobs being killed by AI? Is real corporate AI spending shrinking? Where the data runs out, it says so plainly.

A data-center cooling company just missed its shipments

Vertiv supplies cooling and power gear for data centers and has been one of the big winners in this build-out. Its Q2 2026 revenue came in at $3.27 billion, up 24% year over year.

But it missed its own guidance by about $76 million this quarter, and missed the Street's higher consensus by about $120 million. The $76 million gap is roughly 12% of its year-over-year increase of $630 million (that math is ours).

Management's explanation was "minor timing shifts" — temporary supply chain congestion, plus multi-phase project execution getting more complex as deployments scale. In other words: the orders are there; the gear just didn't ship on time.

Vertiv quarterly revenue from Q1 2023 to Q2 2026
Vertiv's quarterly revenue rose from $1.52 billion in Q1 2023 to $3.27 billion in Q2 2026, more than doubling in just over three years. The +24% marked on the chart is year-over-year growth. Data from Capital IQ, GAAP total revenue. ⚠️ The English phrase "added a substantial $3.27B in revenue" is easy to misread as "revenue increased by $3.27B"; per the chart's own title, "Vertiv Quarterly Revenue," $3.27B is the quarterly total, and the actual increase is $630M.

If Vertiv is hitting supply chain issues, others likely are too. Do those show up elsewhere? Three leads to chase.

Vertiv short by $76M Lead 1: Machinery ordersWho's ordering Lead 2: Backlog ratioHow long to clear Lead 3: Physical importsIs less arriving AC orders doubled in a decade;material handling spiked, then fell Takes ~6 months to clear — high,but flat since 2024 Unit counts haven't fallen;same goods just cost ~20% more Together: demand is unusually strong and outpacing supply; whether supply chains added friction, none of these three can say
Our diagram of this section's verification chain, following the order of the original post. Each lead is unpacked below.

AC orders doubled in a decade; material handling on the same chart spiked and fell back

The first lead is orders. The Census Bureau tracks manufacturing orders by category, and among the data center–adjacent categories, air conditioning stands out the most.

U.S. machinery orders, 2000 to 2026, four series
U.S. machinery orders, 2000–2026. Ventilation, heating, AC, and refrigeration (blue) rose from roughly 3.8 in 2016–2020 to about 7.1 now; turbines, generators, and power distribution (green) around 5.3; mining and oil & gas field equipment (orange) around 4.8; material handling (red) is around 2.85 now, was about 2.4 in 2016 and peaked near 3.7 in 2023. ⚠️ The y-axis units aren't labeled on the chart, so only relative changes can be described. Data from LSEG Datastream and Yardeni Research.

The AC category doubled (turbines, generators, and power distribution rose about 30% over the same stretch — no doubling). That's not hard to explain: high-performance computing is basically turning electricity into heat. More machines, more heat, tighter supply for cooling gear. Generation and distribution rising with it follows the same logic.