Billions spent and hypothetical returns: the AI boom explained with six charts
AI spending, adoption and stock valuations are surging, but the returns, power supply and infrastructure may not keep up.
Intelligence analysis by GPT-5.4 Mini
The Guardian frames the AI boom as a capital-heavy race with fast-rising adoption, soaring tech stocks and huge datacentre build-outs. The article also warns that weak profits, high costs and infrastructure limits could expose bubble-like risks.
The AI boom is like building a giant amusement park before knowing if enough people will buy tickets. Money is being spent fast, but if the rides cost too much to run or the crowds do not pay enough, the whole thing can wobble.
Analysis
The AI boom has become a market story
The article says the AI surge is no longer just about chatbots or software. It is now shaping stock prices, capital markets and infrastructure spending, with the S&P 500 heavily influenced by the biggest AI-linked tech firms.
Spending is racing ahead of proof
Goldman Sachs is cited projecting AI-related spending, from chips to datacentres, rising from $765bn this year to $1.6tn in 2031. But the article stresses a central uncertainty: whether those investments will be built on time, used efficiently and translated into enough revenue to justify the cost.
Adoption is widespread, but the business model is still murky
McKinsey data in the piece suggests most companies are already using AI, while ChatGPT has reached a massive consumer audience. Even so, the article notes that firms still need to show AI improves outcomes enough to cover the bill, especially if they want to automate full workflows instead of just isolated tasks.
Competition, costs and infrastructure are tightening the pressure
Anthropic is gaining ground on OpenAI, helped by Claude Code and a shift toward autonomous AI agents. At the same time, token-based usage costs are rising, and the article says some users are quickly moving to expensive subscriptions. The datacentre build-out is also running into physical and political limits: more power, more grid capacity and more land will be needed, and governments may struggle to deliver.
Why the economy angle matters
The piece ends by arguing that datacentres are already propping up US growth. A Harvard economist is cited saying information-processing investment accounted for most of the US GDP growth in the first half of 2025, which means any slowdown in AI investment could ripple into the broader economy.
Key points
- AI-linked stocks have driven a large share of the recent S&P 500 gains.
- Goldman Sachs expects AI spending to keep climbing sharply through 2031.
- Most companies are now using AI, while ChatGPT has reached 1bn monthly active users.
- Anthropic is gaining ground on OpenAI as Claude usage grows quickly.
- Datacentre expansion may not keep pace with demand, creating a possible compute crunch.
If the spending plans are carried out smoothly, the article says they could unlock a new wave of AI demand. Wider adoption across companies and consumers could also make the technology useful enough to boost productivity and justify the investment.
The article warns that delays in datacentre construction, rising token costs and power constraints could break the economics behind the boom. If returns disappoint, the market could face a dotcom-style correction and the broader economy could feel the hit through slower investment and weaker growth.
