Mark Zuckerberg is on track to spend as much as $145 billion on AI infrastructure this year, a figure that would have looked absurd just two years ago. It is now merely his share of an industry outlay approaching $800 billion across the biggest U.S. tech names. Yet in a CNBC segment on August 10, 2026, Barclays Senior U.S. Economist Jonathan Millar delivered a sobering counterpoint: all that capital has not yet shown up as measurable productivity gains in the broader economy.
Millar’s framing is straightforward. “Spending on capex is not the same thing as productivity. And there’s a lot that comes in, in between in that process of actually making that new capital productive.” He argues that translating hardware and models into economic output requires organizational change, worker reskilling, and industry-wide restructuring, none of which happens on a quarterly cadence.
The 1980s Parallel
Millar reached back four decades for a template. “Typically it plays out over the course of years. So if you remember the experience from the IT boom, we had computers, say in the 1980s, we were seeing lots of imprints from IT on the stock market, on investment and so forth. But it really wasn’t affecting productivity. And it took more than a decade for that to really show up in the numbers.”
The BEA data backs this up in real time. The Information sector grew just 1.5% in Q1 2026, decelerating from 2.5% in Q4 2025. Millar added that “industries adopting AI more rapidly do not appear to be experiencing faster growth.” Adoption surveys show breadth climbing steadily, but sustained daily productive use lags well behind.
Meta: The $145 Billion Bet
Meta Platforms (NASDAQ:META | META Price Prediction) narrowed its full-year 2026 capex range to $130 billion to $145 billion. The strain is visible. Q2 free cash flow collapsed to $784 million, a -91.31% year-over-year decline, while long-term debt climbed to $83.66 billion. EPS of $6.18 missed the $7.2173 consensus, snapping a six-quarter beat streak, per the company’s Q2 8-K. Shares are down 9.72% year to date.
Amazon: $220 Billion and Still Short of Capacity
Amazon (NASDAQ:AMZN) is guiding to roughly $220 billion in cash capex. CEO Andy Jassy said “even at that amount, we will still not have enough capacity to meet all the demand we have in 2026.” AWS grew 36.7% in Q2, and the AI and chips businesses each cleared a $25 billion run rate. Shares are up 20.48% YTD.
Alphabet: $205 Billion and Supply Constrained
Alphabet (NASDAQ:GOOGL) raised its 2026 capex range to $195 billion to $205 billion. Google Cloud revenue grew 82% in Q2 with backlog reaching $514 billion. CEO Sundar Pichai conceded the enterprise story is early: “Now think about what percentage of workloads are really AI-native and AI-enabled. It again feels like very, very early.” Shares are up 14.37% YTD.
Microsoft: $175 Billion With FCF Intact
Microsoft (NASDAQ:MSFT) is set to spend around $175 billion. Azure crossed $100 billion in annual revenue for the first time, and Microsoft 365 Copilot passed 30 million paid seats. Commercial RPO stands at $625 billion. Shares are up 5.11% YTD.
What Investors Should Watch
Millar’s argument resets the timeline for AI’s payoff. If the 1980s pattern holds, the productivity payoff from today’s $800 billion in annual capex may not surface in national statistics until organizations finish rewiring workflows and workforces. The bull case for these four names depends on execution reaching customers faster than debt costs and depreciation schedules bite. For now, Meta’s collapsing free cash flow is the sharpest reminder that the bill is arriving before the productivity is.
Contact [email protected] for any questions or corrections.