AI slowdown calls justified but collapse of bubble may be more immediate threat | Heather Stewart
Scale of debt issuance being used to fund the hectic pace of the datacentre rollout by tech firms is a worry Warnings about powerful and uncontrolled AI have dominated the headlines this past week. But even amid the omens of apocalypse , it is still worth asking if the entire AI-industrial complex is sailing towards a financial iceberg. It will hardly be our most pressing concern if the bots are…
The recent surge of warnings about potentially uncontrollable AI has sparked discussions, yet experts remain concerned about the possibility of a financial collapse within the AI-industrial complex. While the prospect of AI taking over the world may be alarming, the implications of a bubble bursting could have far-reaching consequences beyond the United States.
AI executives have recently joined forces to express concerns that their immensely powerful creations could either destroy or fail to destroy humanity. Some are advocating for limitations on the rapid advancement of AI technologies and their applications.
There is a growing awareness that AI requires regulation, as evidenced by incidents involving Meta's smart glasses recording without consent and the failure of safeguards allowing chatbots to engage in hacking. Proposed measures, such as independent AI model analysis, appear to be positive improvements. However, it is crucial to remain vigilant against the risk of a few intricately linked megafirms, heavily indebted, relying on state-backed support to maintain their dominance.
Beyond the potential dangers posed by AI products, tech firms face traditional financial challenges. The year 2022 alone saw massive debt issuance by hyperscalers, including Google, Amazon, Microsoft, Meta, and Oracle, reaching $132 billion, according to one estimate. With bond markets experiencing fragility and 10-year US treasuries yielding around 5%, these debt levels may trigger a market reevaluation.
Moreover, the "unit economics" of AI are questionable. Large language models' processing costs are plummeting, while building them remains expensive. OpenAI has continuously reduced fees to retain customers, and a recent report by Bloomberg stated that the cost of AI is collapsing, but the cost of building it is not keeping pace.
Silicon Data's index reveals that the price for a million tokens has more than halved since June, reaching less than $1. Nevertheless, the frenzied demand for hardware components like semiconductors in datacenters continues to keep costs high.
Some argue that AI businesses are leveraging complex financial mechanisms, making it difficult for laypeople to comprehend their operations. A recent analysis from financial analyst Groundbreaker draws a parallel with the prelude to the 2007-2008 global financial crisis, citing a "compute commencement wall" of $1.5 trillion facing AI labs over the next couple of years.
Much like the "teaser" mortgage rates that later escalated, datacenter contracts are often structured as "take or pay," with no initial payments due until specific deadlines, and the hyperscalers recording the value as future revenue.
Written by urgent.news from Guardian Business's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
