To evaluate a well-managed liquid alternative fund, he highlighted four topics for asset allocators: long-short exposure management that aligns the portfolio construction process with the defined risk targets; alpha generation from security selection; asset protection strategies so there’s a low correlation to traditional asset classes; and having a multi-asset class focus, which allows an investment manager to source the best risk-adjusted opportunities.
Read: Back to the basics on liquid alternatives
The most dominant, important theme in financial markets, according to Murray, is the artificial intelligence infrastructure capital spending cycle. It’s heavily impacting equities, has a rising impact on public and private credit and the leverage to the cycle is rising significantly, he added.
By the end of 2026, the five largest hyperscalers — Apple Inc., Amazon.com Inc., Google, Meta Platforms Inc. and Oracle Corporation — are going to spend close to $800 billion, marking a 90 per cent growth year over year, he said, noting their capital expenditure in 2027 is expected to reach between $900 billion and $1 trillion.
“The real number is going to be between $1.2 and $1.5 trillion, so another 20 to 50 per cent upside to current consensus expectations. This is liquidity moving from financial assets to the real economy. A significant amount of AI infrastructure [capital expenditure] is driving earnings. Earnings are driving the equity market. Most of that is coming from AI infrastructure stocks.”
There are many ways to measure the evidence of return on all of this capital, said Murray, adding the most important is to follow the financial key performance indicators from the AI companies, such as Anthropic. Its annual recurring revenue has increased from $9 billion at the end of 2025 to $44 billion at the end of May this year.
Read: Data centre, AI growth trajectory facing risks that could impact institutional investors: expert
“That’s the fastest growth in ARR in the history of the world. . . . They’re adding more than $5 billion of ARR every two weeks. The buy side expects this number to be between $80 and $100 billion at year-end and I think it will comfortably surpass that.”
While Murray encouraged investors to embrace this theme, he also noted risks are rising. The hyperscalers are all seeing accelerating cloud revenue growth and expanding margins, he said, highlighting that more than 50 to 60 per cent of their capital expenditure is being put into AI infrastructure hardware with short life spans.
“If the revenue doesn’t show up, you’re going to have a material earnings hit from a high level of depreciation.”
In addition, hyperscalers are spending more than 100 per cent of cash flow from operations on capital expenditure in 2026, he said, and he expects that trend to continue across 2027. “Again, it’s the change in liquidity that I would be most focused on as the primary risk, not necessarily open-source models versus frontier, not token maxing versus token minimizing.”
Read more coverage of the 2026 Alternative Investment Conference.