Anthropic's $750 Billion Compute Bill Could Delay Its Path to Profitability
The company is committing an unfathomable amount of money to compute capacity over the next decade.
Anthropic, the artificial intelligence company, anticipates a staggering $750 billion investment in compute capacity over the next decade to support its Claude AI system, as disclosed in a draft of its upcoming S-1 filing ahead of the anticipated IPO. This significant expenditure is primarily driven by the need to run AI queries and train Claude, with each user interaction requiring a specific number of tokens, each commanding a high price, often reaching up to $8,000 in total.
A substantial portion of this compute expenditure, about $450 billion, is tied to noncancelable cloud compute spending contracts, ensuring Anthropic's ongoing commitment even if usage fluctuates or market conditions change. Despite a rapid acceleration in revenue, reaching $16.3 billion in the first half of the year, Anthropic still faces substantial losses, with the first half of the year seeing a $80 billion loss, primarily influenced by an accounting adjustment related to convertible notes.
While the company reported a slight profitability on a non-GAAP basis in the third quarter, GAAP profitability remains elusive in the foreseeable future, primarily due to its steep, non-GAAP-adjusted spending on compute. This financial strain, coupled with potential risks from losing key customers or a slowdown in revenue growth, underscores the considerable risk for investors considering Anthropic post-IPO.
Written by urgent.news from Motley Fool's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
