Diversification in the AI era

05 August 2026

• Wherever you side in the AI ‘bubble’ debate, portfolio diversification remains vital
• Traditional government bonds have been curiously failing in this role
• A slew of alternative fixed income approaches is available to investors

It is no exaggeration that artificial intelligence (AI) is dominating investment discourse today, with debate raging over whether we really are on the precipice of a revolution that will power the profits of both the enablers and beneficiaries of AI for many years to come. Proponents point to the actual, real revenue being earned by the large language builders, with Anthropic recently hitting USD 30 billion of annualised run rate (ARR)¹ and OpenAI not far behind². Both are expected to publicly list within the next twelve months³. Detractors point to the hype – SpaceX proposing data centres in space, as well as the sheer gains witnessed in the equity market by previously staid stocks like memory maker Micron, which is up a nosebleed-inducing 216% year to date as at 27 July this year⁴.

Yet whichever side of the argument appeals more, one thing hasn’t changed: serious investors across asset classes need to consider how they diversify the volatility that has naturally characterised equity investing for hundreds of years and which will doubtless continue to do so. For a long time, the standard solution was “60:40” portfolio, typically a blend of large-cap equities alongside a sleeve of medium-dated government bonds designed to smooth out the bumps. And this was a hugely successful combination.

For several decades, US stocks in particular have compounded 6.5-7% per annum over a multi-decade period, an outcome which has become known as the “Siegel Constant” after Professor Jeremy Siegel, who famously described it in his book Stocks for the Long Run⁵, even though returns were not positive in every calendar year. And, from the early 1980s through to the early 2020s, US Treasury bonds also rallied and generally provided diversification benefits when equities stumbled⁶. The benign bond markets owed much to US Federal Reserve chair Paul Volcker, who in the early 1980s tightened monetary policy to finally bring the price shocks of the 1970s under control. Then in the 1990s and 2000s, globalisation lowered the cost of manufactured goods as Asia was able to capitalise on its labour force and export more freely.

Government bonds enjoyed a long structural tailwind as yields fell and prices rose to reflect this low inflation environment which became known as ‘The Great Moderation’. So far so good for investors. But then in the aftermath of the Covid-19 pandemic, the pattern shifted. Bonds stopped rallying consistently and began to disrupt the 60:40 portfolio concept at awkward moments. Russia’s invasion of Ukraine in 2022, Donald Trump’s Liberation Day tariffs announcement in April 2025, and latterly the war with Iran beginning in February 2026 all saw US Treasuries sell off at the same time as stocks. The “40” was failing, and it did not go unnoticed.

In April this year, respected researchers at AQR published an extensive paper on the phenomenon, noting that “the correlation between stocks and government bonds, once fairly reliably negative, turned positive a few years ago.” The key questions for those involved both in constructing and investing in enduring portfolios are: why has this changed, will it revert back again and, if not, what alternatives should be considered?

Chart 1: Naughty forty – ‘classic’ bonds are no longer diversifying from stocks

Five-year correlation between the S&P 500 and Bloomberg US Aggregate Bond Index, from 31 Dec 1999 to 27 Jul 2026

The reasons for this change are not of purely academic interest. Understanding what has changed for bonds is key to determining whether their diversifying characteristics will ever ‘return’ and whether it’s time to seek alternatives. The first point to make here is that the macroeconomic environment has become decidedly more inflationary and bonds have adjusted accordingly. Since the end of the Covid-19 pandemic, inflation has been consistently under-estimated by central banks such as the Federal Reserve. In the aftermath of that emergency, the US central bank was continuously questioned for (not unreasonably) believing that the spike in inflation was merely ‘transitory’. Then Russia invaded Ukraine, lifting oil prices amid the West’s sanction-based response to Russian aggression. Donald Trump’s second term as US President began barely three years later and its protectionist stance coalesced around “Liberation Day”, which at a stroke single-handedly sought to reverse the globalisation era and, by implication, the low inflation that accompanied it.

Yet another energy shock followed in February 2026 with the war involving Iran, which saw Brent crude futures spike to nearly USD 120 per barrel in March and April of this year (as at 27 July they remained as high as USD 85)⁷. Central bankers often fret about the ‘de-anchoring’ of inflation expectations, in which the public see inflation around them and then come to expect higher inflation, which in turn creates more of it in a self-fulfilling prophecy. This is hard to contain, not least because in the case of supply shocks central banks cannot magic up more oil. Instead, they are left with the blunt tool of interest rates which, if used to control inelastic demand for energy, would cause a brutal and unacceptable recession. It’s therefore hard to argue for a return to a more benign inflation environment from here, given the unpredictability of geopolitics and consumers’ increasingly fatalistic view of future price increases.

This uncertainty feeds another driver of higher bond yields (and lower prices), namely the rising term premium in US Treasuries. Put simply, this is additional yield which lenders to the US government demand in return for the lack of clarity around not just future inflation and interest rates but also fiscal (tax) policy. The US is carrying debt of nearly USD 40 trillion and rising. According to Bloomberg Economics, fully 0.7% of the 4.6% yield on US Treasuries as at 27 July is derived from this uncertainty premium which bondholders demand as the price for lending to Uncle Sam against an increasingly volatile macroeconomic and geopolitical backdrop.

Other factors are also at play in pushing down bonds and at this point it’s probably worth mentioning that China is weaning itself off the habit of parking money in US Treasuries. Its holdings have gone from over USD 1 trillion in 2011 to just over USD 600 billion today⁸. As these bonds are slowly dumped into international markets, they have the inevitable effect of lowering prices. And China could continue to unwind its holdings for another decade should it so choose. Taken together, the factors discussed feel far more structural than temporary in nature.

But investors need not despair. Breathless declarations that Markowitz’s original concept of portfolio diversification is dead are far too dramatic and simply not the case. Investors instead need to consider effective alternative fixed income approaches, which do exist. Three examples in particular stand out, namely insurance-linked catastrophe (cat) bonds, mortgage-backed securities and emerging market debt.

Cat bonds are securities that allow insurers and reinsurers to transfer the financial risk of major natural disasters, such as hurricanes or earthquakes, to investors. Investors receive regular coupon payments in return, but if a predefined catastrophe occurs, they may lose some or all of their principal, which is used to cover insured losses. A skilled cat bond manager can build a well-diversified portfolio of these bonds which ensures a steady collection of coupon income over time without excessive exposure to a single type of natural disaster. In addition, they can trade around the aftermath of events to pick up unwanted bonds at lower prices. The main advantage for portfolio builders is the relative independence of these bonds to what is going on in traditional capital markets such as equities or government bonds.

Mortgage-backed securities (MBS), meanwhile, offer another source of diversification, receiving cash flows generated by groupings of responsible homeowners’ mortgage payments. They differ from US Treasuries because they offer an additional yield to compensate for changes in housing conditions and mortgage refinancing activity. Again, an expert MBS manager will be adept at managing a portfolio that anticipates these factors and the result will be a return profile that is refreshingly independent from traditional US Treasuries.

Finally, emerging market debt refers to bonds issued by governments or corporates in developing economies. Higher yields than developed market equivalents are the reward on offer for taking on risks such as political instability, economic uncertainty and currency volatility. Once again, specialist managers in the field can extract these higher yields from a portfolio while successfully managing the risks described in such a way as to preserve capital.

Chart 2: Better because they’re different – ‘cat’ bonds tend to outperform amid equity instability

From 28 December 2001 to 28 July 2026

The need for diversification is an ever-present fact in investing. However positive the equity journey appears to have become in the last few years, serious allocators of capital cannot afford to become complacent. Apart from anything else, most investors desire some degree of smoothing along their investment journey, with very few opting to simply absorb the volatility equities naturally generate, unless their horizon is truly multi-decade in nature. With that in mind, the smoothing requirement has become something of an awkward investment challenge today. The 60:40 approach which worked so well for so long finally appears to have run out of steam in the last few years after repeated inflation shocks disrupted the world economy. Given the polarised state of politics and international relations today, it is not unreasonable to assume that normal service is not going to assume anytime soon for US Treasury bonds, meaning upward pressure on yields and downward pressure on prices looks set to continue for some time.

Happily, there are alternative fixed income approaches available, with cat bonds, MBS and emerging market debt being just three examples that come to mind. All can offer diversification in their different ways but there is one important caveat here: none of them are ‘fire-and-forget’ solutions in the way that US Treasuries had become in investor portfolios. They require careful and expert management given their unique features and characteristics. But, assuming the right provider can be identified, they potentially offer an effective new take on portfolio diversification.

Julian Howard is Chief Multi-Asset Investment Strategist at GAM Investments. This article represents the views of GAM’s Multi-Asset team.