Science & Tech · AI · 3 hrs ago
AI computing race is adding pressure to mobile phone prices
The race to build and run AI systems is straining computing capacity across the technology industry.
The shortage involves chips, data centres and energy, and is affecting companies including Anthropic and major cloud providers.
Anthropic’s Claude Code drew so much demand that its services faced repeated outages, and the company later secured computing capacity from Elon Musk’s xAI.
AI companies are competing for a limited supply of computing power, while the largest US cloud firms are sharply increasing infrastructure spending.
Renting an Nvidia H100 GPU under a one-year contract has become 60% more expensive in a year, according to Semianalyst.
The chip shortage is also putting pressure on consumers’ wallets, and the story says mobile phone prices are expected to keep rising.
The squeeze is likely to continue until something changes, with a fall in demand identified as one possible turning point.
The race to secure computing power for artificial intelligence is adding pressure to mobile phone prices.
A shortage of chips, data centres and energy is limiting the supply of computing capacity.
US cloud giants invested $380 billion in AI infrastructure in 2025, up from $217 billion in 2024.
Forecasts suggest their investment could reach $725 billion in 2026.
Rising demand has pushed up the cost of renting Nvidia H100 GPUs and led cloud providers to seek longer contracts.
- Who
- Technology companies seeking computing capacity, including Anthropic and major US cloud providers, and mobile phone consumers.
- What
- The AI computing race is intensifying shortages and raising costs, adding pressure to mobile phone prices.
- When
- The article was published on 21 October 2026; it describes events from early 2026 and investment figures for 2024 to 2026.
- Where
- The article discusses the global technology sector and a meeting at xAI’s offices in Hawthorne, California.
- Why
- Demand for computing capacity is outstripping supply, amid shortages of chips, data centres and energy.
This story does not have two clearly opposing sides.
Amazon, Google, Meta and Microsoft invested $217 billion in AI infrastructure.
Anthropic sought additional computing capacity from Meta, which decided not to provide it.
Anthropic added $6 billion in annual recurring revenue, according to the article.
Anthropic co-founder Tom Brown met Elon Musk in Hawthorne, California, to seek a computing-capacity rental agreement.
An agreement for xAI to rent more than 300 megawatts of computing power for $45 billion over the coming years was confirmed.
The four major US hyperscalers invested $380 billion in AI infrastructure.
- Investment in 2024
- $217 billion by Amazon, Google, Meta and Microsoft
- Investment in 2025
- $380 billion by Amazon, Google, Meta and Microsoft
- Forecast investment in 2026
- Up to $725 billion
- Nvidia H100 rental price
- Up 60% in 12 months for a one-year contract, according to Semianalyst
- xAI computing-power agreement
- More than 300 megawatts; $45 billion over the coming years
- Anthropic revenue added in February
- $6 billion in annual recurring revenue
What people are saying
There is a sense that if you do not build this capacity now, you will lose the opportunity to remain relevant in the long term.