Global Market: AI investment story shifts from spending to picking the winners

The latest earnings season has shifted the debate around artificial intelligence investments from whether Big Tech's massive spending spree will generate returns to which companies are best positioned to capture the long-term gains from the AI boom, as per a Reuters report.

Strong results from Microsoft and Amazon have reassured investors that demand for the infrastructure supporting artificial intelligence remains robust. Accelerating cloud growth and persistent shortages of computing capacity have reinforced expectations that AI-related spending will remain elevated.

However, investors are increasingly looking beyond the current supply constraints. The focus is turning to which companies can maintain strong profit growth once additional computing capacity comes online and the market becomes less constrained.


According to Reuters, many large asset managers continue to hold substantial positions in semiconductor companies despite a sector sell-off in July, when concerns emerged over the sustainability of AI spending and intensifying competition from China. At the same time, investors have been increasing their exposure to hyperscalers, the major cloud providers with the scale and infrastructure needed to expand AI capacity rapidly.

Hyperscalers Gain Investor Attention

The four largest AI capital spenders have underperformed the broader semiconductor sector, even as the Philadelphia Semiconductor Index has surged about 75%.

By comparison, Nvidia-backed neocloud providers such as CoreWeave and Nebius have delivered much stronger gains. Their businesses involve renting computing capacity to AI laboratories and enterprises, allowing them to benefit from elevated prices for scarce computing resources.

According to the report, some investors believe hyperscalers will eventually benefit from the enormous investments they are making today. As new data centres become operational and AI services generate more revenue, the largest cloud companies could see earnings and cash flow growth accelerate relative to capital spending.

A Reuters analysis estimates that hyperscalers could generate around $340 billion more in annual operating cash flow in 2027 than in 2025, while capital expenditure is expected to increase by roughly $534 billion over the same period.

AI Investment Is Becoming an Ecosystem Bet

The investment opportunity is not necessarily a choice between semiconductor companies and cloud providers. Instead, investors increasingly view artificial intelligence as a broad ecosystem involving chips, data centres, cloud infrastructure, software and AI applications.

Data centres generally take between 12 and 18 months to move from construction to revenue generation. This means the enormous capital spending reported by technology companies today could translate into stronger revenue and cash flows as infrastructure becomes operational.

This dynamic could create opportunities across multiple parts of the AI supply chain rather than concentrating returns in a single group of companies.

Investors Look for Sustainable Winners

The next stage of the AI investment cycle could favour companies that control both computing infrastructure and the software layers that allow customers to deploy AI efficiently.

Companies such as Amazon, Microsoft and Google have potential advantages because of their enormous scale, established customer relationships and ability to control significant portions of their infrastructure.

Neocloud providers, meanwhile, have benefited from tight computing capacity and high prices. But investors are increasingly questioning whether those advantages can persist once additional computing capacity enters the market.

Some strategists expect neocloud providers to face greater pressure if computing supply expands and pricing normalises, as per the report. Their reliance on debt and elevated valuations could make them particularly vulnerable to a decline in AI infrastructure prices.

Valuations Remain an Important Consideration

The valuation picture also favours a selective approach. Hyperscaler valuations have declined from their post-pandemic peaks, although significant differences remain across the group.

Microsoft trades at roughly 24.6 times forward earnings, while Meta is valued at around 17.6 times forward earnings. The divergence suggests that investors are no longer treating every major technology company as an identical AI beneficiary, the report stated.

For some investors, this has created opportunities to favour companies with stronger balance sheets, deeper customer relationships and more diversified technology portfolios.

AI Boom Could Produce Fewer Winners

The biggest challenge for investors may ultimately be identifying which companies can convert enormous AI investment into sustainable profits.

Swiss wealth manager LGF+ZEST estimates that AI monetisation may need to increase substantially to justify current spending plans. That raises the possibility that the eventual financial winners could be far fewer than the number of companies currently benefiting from the AI investment boom.

Investors expect competition to intensify as the market matures. Companies with broad technology portfolios, strong customer relationships and greater control over their own infrastructure could gain an advantage over more specialised competitors.

The AI investment story, therefore, is moving into a new phase. The question is increasingly not whether companies will spend heavily on AI, but which businesses will turn that spending into durable revenue, stronger cash flows and long-term shareholder returns.