Can Big Tech’s $1-trillion AI bet pay off? Jefferies explains why it remains bullish

The artificial intelligence race is driving one of the largest investment cycles in technology history. Major hyperscalers, including Microsoft, Amazon, Alphabet, Meta and Oracle, are expected to spend about $818 billion in 2026 and $1.12 trillion in 2027 on data centres, chips, networking equipment and other AI infrastructure, according to Jefferies.

The investment boom is putting pressure on balance sheets and cash flows. Aggregate debt for these companies is projected to increase from $226 billion in 2024 to $502 billion in 2026. Their combined free cash flow is expected to fall from $238 billion to a marginal deficit of $2 billion over the same period.

Despite the near-term financial pressure, Jefferies remains constructive on the AI investment cycle. Here are three reasons why:


1) Customer commitments are growing faster than investments

Hyperscaler backlogs have increased by approximately $1.67 trillion since December 2024, nearly four times the $453-billion increase in capital expenditure over a comparable period.

Backlogs, also known as remaining performance obligations, represent contracted revenue that companies expect to recognise in the future. Their rapid growth suggests that customer commitments, rather than expectations alone, are supporting the infrastructure expansion.

Anthropic's growth illustrates the speed at which this demand is being monetised. Jefferies estimates that Anthropic could reach $200 billion in annual revenue within seven years of its launch, a milestone that AWS is expected to take about 25 years to achieve. Its quarterly revenue surged more than 14-fold year-on-year to $11.5 billion in the second quarter of 2026, significantly outpacing the estimated fourfold annual growth in available computing capacity.

Demand also remains strong even after excluding large contracts from frontier artificial-intelligence companies, according to the brokerage.

2) Cloud growth and operating cash generation remain strong

Combined cloud revenue growth among the hyperscalers accelerated to 49% in the second quarter of 2026 from 25% in the first quarter of 2025 as demand for AI computing increased.

Aggregate operating cash flow rose 36% year-on-year to $140 billion during the June quarter. This provides the large technology companies with greater capacity to fund their investments, even as heavy capital expenditure weighs on free cash flow.

Jefferies believes the current cash-flow pressure reflects the timing of investments: companies must build data centres and install computing capacity before these assets can generate revenue.

3) AI infrastructure economics are improving

Cloud operating margins expanded for four consecutive quarters, rising from 33.8% in the second quarter of 2025 to 38.9% in the second quarter of 2026. This indicates that hyperscalers have been able to absorb or pass on higher memory and component costs.

Long-term customer contracts and improving infrastructure utilisation are also supporting returns. Jefferies noted that some AI servers can recover their costs in less than three years while remaining operational for five to six years.

Meanwhile, AI adoption is still at an early stage. An estimate cited by Jefferies puts the potential AI addressable market at around $26 trillion. However, the median company spends only about $12 per employee each month on AI, compared with average software spending of $777 per employee each month, leaving considerable room for expansion.

However, the pace of expansion could be constrained by power availability, shortages of skilled workers and delays in obtaining regulatory permits. Enterprise adoption could also remain uneven because of security concerns, fragmented corporate data and uncertain returns from some AI applications.

Jefferies' top AI stock picks

Among mega-cap companies, Jefferies prefers Microsoft, with a price target of $575, for its position across Azure and Copilot; Amazon, with a $330 target, on expectations of stronger AWS growth; and Alphabet, with a $445 target, for its integrated stack of custom chips, Gemini models and Google Cloud.

Its principal large-cap picks are Oracle, with a target of $290; CoreWeave, at $150; and Snowflake, at $385. These companies offer exposure to AI computing and the data infrastructure needed to build and operate AI applications.

Jefferies also identified Intuit, with a $500 target, as its contrarian software pick, arguing that its proprietary data and established tax and accounting platforms provide protection against AI disruption.

(Disclaimer: Recommendations, suggestions, views and opinions given by the experts are their own. These do not represent the views of The Economic Times)