# AI's Epic Flop: $700B Black Hole in the US Economy
Buckle up, developers—AI's trillion-dollar promise just got punked by cold, hard GDP math. Goldman Sachs chief economist Jan Hatzius didn't mince words: AI investment contributed 'basically zero' to US economic growth in 2025, even as Big Tech dumped hundreds of billions into data centers and chips. The US economy chugged along at 2.2% growth, with AI scraping a pathetic 0.2 percentage points—that's Joseph Politano's brutal tally from Apricitas Economics.
<> "There's been a lot of misreporting of the impact that AI investment had on GDP growth in 2025, and it's much smaller than it's often perceived." — Jan Hatzius/>
This isn't just underwhelming; it's a scathing indictment of the AI gold rush. Morgan Stanley and JPMorgan echo Goldman: the buildup's direct GDP punch is dramatically overstated, possibly zero. Why? GDP accounting is a buzzkill. Three-quarters of those fat AI data center bills flow to Asian chip fabs—Taiwan, South Korea, you name it. That spending subtracts from US output because it fattens foreign economies, not ours. Developers, think about it: you're coding the future on hardware that's economically offshore.
The Hype vs. Reality Smackdown
AI evangelists sold us productivity Armageddon—models scaling to godlike intelligence, juicing GDP like the internet on steroids. Reality? Thousands of CEOs admit zero impact on employment or productivity; 80%+ of firms see no gains. Hannah Rubinton at the St. Louis Fed calls her own rosy 39% growth attribution a ceiling—"It's not like AI is propping up the economy." Politano nails it: big deal, sure, but not the be-all-end-all.
Here's the analytical gut punch:
- Localized wins, national zilch: Construction booms, power grids groan under AI's thirst—but that's no broad growth engine.
- Measurement mess: No reliable way to track AI's sneaky business/consumer boosts yet.
- Future forecasts fizzle: Goldman pegs potential productivity bumps at 0.3-3% annually eventually, but 2025? Crickets.
Developer Wake-Up Call: Stop Drinking the Kool-Aid
As coders, we're the foot soldiers in this hype war. AI tools feel magical—agentic agents like Claude or OpenClaw hint at AGI vibes—but where's the ROI? We're building on promises of explosive gains that haven't materialized. Tech giants eye $700B in 2026 AI infra spend. If productivity doesn't spike, this becomes a sustainability nightmare. Investors poured in on faith; now faith's fraying.
My hot take: AI's no NFT scam, but the bets are wildly mismatched. It takes time—internet lagged 20 years for full impact—yet we're frontloading trillions like it's 1999 dot-com 2.0. Slowdown risks? Catastrophic for hyperscalers' capex addicts. Broader economy? Meh, it'll shrug.
Bottom line: Ditch blind faith; measure ruthlessly. Track your AI experiments' real output. If Goldman, CEOs, and GDP say zero, pivot. The real revolution? In code that ships value, not vaporware datacenters.
