New Delhi [India], November 12 (ANI): The massive investment in artificial intelligence infrastructure is facing a reality check as concerns grow over whether technology companies will be able to generate sufficient returns from their spending, with the next 18 months emerging as a crucial period for the sector.
According to a Goldman Sachs report, the AI investment cycle continues to expand, with technology companies committing large amounts of capital to data centres, computing infrastructure and advanced chips.
However, questions are increasing over the ability of businesses to convert this spending into sustainable revenue and profits.
Chamath Palihapitiya, Founder and CEO of Social Capital, said the sustainability of the investment cycle will ultimately depend on whether companies using AI can generate meaningful returns from it.
Highlighting the gap between productivity gains and actual business performance, he said the “number of companies that can attribute earnings/profitability gains to AI is essentially zero today”.
He said, “The next 18 months are critical. We will not have the final answer by then, but we will have set the table for a clearing process that distinguishes sustainable AI business models and infrastructure commitments from unsustainable ones”.
The report also raises concerns about the business models of specialised cloud infrastructure providers, or neoclouds, which are making large commitments to GPUs, power and data centre capacity.
Their business models depend on keeping computing infrastructure sufficiently utilised to generate revenue and meet long-term financial obligations. However, demand from end customers remains a key uncertainty.
The financing side of the AI expansion is also becoming more selective.
Miriam Wheeler, Goldman Sachs' Global Head of Leveraged Finance, said investors are increasingly focusing on the quality of projects, developers, leases and tenants.
She described the emerging trend in the market as “Up in credit quality”.
She said, “Structural complexity will likely come out of the market as investors move up in credit quality without needing to stretch. In the short term, leases with non-IG tenants may remain more challenging. Developers with an unproven track record, limited equity, or inferior leases will likely struggle to access capital”.
The report noted that around 30 AI-related deals came to the high-yield debt market over the past 12 months, but new deal activity contracted sharply after the end of September, with no new deals coming to the market over the last month.
At the same time, the report presents a more positive outlook for large technology companies, which have stronger balance sheets and established businesses to support their AI investments.
Eric Sheridan, Goldman Sachs' lead US Internet and Interactive Entertainment analyst, identified three major factors that could slow the investment cycle: disappointing demand, lack of access to capital and insufficient power.
He cautioned that “If any one of those is missing, current trough of disillusionment could become permanent contraction.”
The report highlighted that six major US hyperscalers -- Alphabet, Amazon, Microsoft, Meta, Oracle and SpaceX -- could achieve returns on invested capital of 15 per cent or more, according to Goldman Sachs' estimates.
Overall, the report does not point to an immediate collapse of the AI investment cycle but highlights the possibility of a correction and consolidation if returns fail to match the scale of spending.
The coming 18 months could therefore prove critical in identifying which companies have sustainable business models and which may struggle to meet their financial commitments. (ANI)
AI investment faces reality check as returns remain unproven, next 18 months critical: Goldman Sachs
