Global equity markets have spent 2026 climbing from one record high to the next — the S&P 500 has set more than two dozen of them and is up around 11% for the year — and yet, for all the talk of stretched valuations, share prices have risen far more slowly than the profits supporting them. That is an unusual state of affairs, and, we would argue, an encouraging one. A market that makes new highs while its earnings grow faster still is not becoming more expensive; it is becoming better supported. The gap between what companies are earning and what investors are willing to pay for those earnings has, if anything, been closing.

The scale of the earnings expansion has been extraordinary. According to LSEG I/B/E/S data compiled by MRB Partners, S&P 500 companies grew second-quarter earnings by 52% year-on-year, on revenue growth of more than 15%. This is the strongest pace since the recovery from the pandemic and the seventh consecutive quarter of double-digit growth. Even after stripping out the mark-to-market gains that flattered the headline number, chiefly the large unrealised profits on the stakes Alphabet and Amazon hold in Anthropic and SpaceX, underlying growth was still close to 30%. More than 85% of companies beat earnings estimates, profit margins reached a record, and consensus now expects earnings to rise 33% across 2026 as a whole, followed by 15% in each of the next two years. Earnings growth has, quite simply, exploded.

Because profits have outrun prices, valuations have quietly de-rated even as the index has made new highs. The S&P 500 began the year on a forward price-earnings multiple of around 22 times — one of its richest readings on record — and that multiple has since slipped back as earnings caught up with, and then overtook, the share-price gains. Capital Economics made the point neatly in a note on 27th August: big-tech companies’ share prices, it observed, have “grown much more slowly than their earnings have”, and the sector’s valuation premium “has narrowed further”. MRB Partners, in its 4th September asset-allocation review, went further, arguing that “better value and improving growth momentum” now justify overweighting select markets outside the United States. Valuations, in other words, are beginning to look compelling rather than dangerous.

Nowhere is this de-rating more evident than in the very part of the market that has attracted the most anxiety, notably AI. Artificial intelligence remains the driving force behind a deepening and broadening of global growth — the capital spending on data centres, semiconductors and power generation is now the single largest swing factor in the United States economy — and yet the shares most exposed to it have lagged. Capital Economics notes that the ten most profitable companies in the S&P 500, which between them account for roughly a third of the index’s earnings, would actually have underperformed the market this year. The engine of the expansion has been re-rated downwards even as it does the heavy lifting.

That the market disbelieved the AI investment case is not in doubt: through the first half of the year, credit-default-swap spreads on the hyperscalers widened and their multiples compressed to their worst stretch of relative performance since 2022. The second-quarter results allayed the concerns. BCA Research, in its strategy note of 3rd August, showed that the return on incremental invested capital for the hyperscalers is running near 30%, with cloud revenues up 43%, and concluded plainly that “the ROI on AI capex is showing up”. In a companion piece the following day, BCA’s Marko Papic argued that, with the AI concerns now largely addressed, “only the Ayatollah and Putin can save the bears” — that only a genuine geopolitical shock, rather than anything in the fundamentals, could derail the advance. Crucially, technology’s forward earnings multiple has fallen relative to the start of the AI boom, relative to the rest of the market, and relative to the dotcom peak of 2000: the sector’s equity risk premium is a positive figure of around 2% today, against roughly minus 2% at that earlier summit. The excess that ended the last cycle is simply not present in this one.

As always there are significant risks, of which two stand out: a rapid rise in long-term bond yields, and a fresh bout of geopolitical turbulence. The bond-market case deserves to be put fairly. MRB warns that global yields are “poised to climb further over the next 6-12 months”, with central banks possibly needing to tighten by more than investors expect, and with the enormous capital demands of the AI build-out — alongside heavy government borrowing — adding to the upward pressure. The US 10-year Treasury yield, hovering at 5.0%, is already at its highest since October 2023, and a Federal Reserve under more hawkish leadership could lift the policy rate by 75 basis points according to fed funds futures. Against that, Capital Economics doubts the Treasury market will “be the straw that breaks the US stock market’s back”, noting that monetary policy is most unlikely to be tightened anything like as aggressively as it was into 2000. Two facts reinforce the calmer prognosis. Bond-market volatility, as measured by the MOVE index, remains muted, and inflation expectations well anchored. And monetary conditions are still, on balance, accommodative: with both the policy rate and the 10-year yield sitting below nominal GDP growth of around 5%, and far below the pace of earnings, the cost of capital remains comfortably beneath the return on capital.

The geopolitical risk is, for now, chiefly an energy risk. Brent crude has climbed back above $100 a barrel — most recently on attacks against Saudi facilities — and stands some 50% above its level of a year ago. Yet the world is coping with elevated oil prices far better than the headlines imply. MRB judges the current level “roughly neutral” for global growth, observing that today’s prices merely remove a previous tailwind rather than impose a fresh drag, and its earlier work on energy sensitivity cautioned against underestimating the economy’s resilience. With the US midterm elections in November, moreover, the administration has little appetite to let fuel prices run higher into the vote, and the incentives point firmly towards relief. We would expect oil to drift lower rather than spike from here.

It is worth adding that rising bond yields are not, in themselves, a threat. Part of the increase reflects rising productivity — US productivity growth is running well above its long-run norm enabling a higher potential growth rate for the economy; a bond market repricing to reflect stronger real growth and stronger earnings is a rational and healthy development, not a warning. In the end it is earnings, not multiples, that drive markets, and the earnings picture is strong enough that even a considerable disappointment could be absorbed while leaving the investment case intact — the forecasts carry a wide margin of error in the investor’s favour. Valuations are undemanding relative to where they have stood at comparable moments in the past, and the greater danger, as ever, lies in calling the bear market too early rather than in staying invested.

Overberg Asset Management therefore remains fully invested, holding to its long-standing tilt towards the technology leadership driving this cycle while diversifying steadily towards the better-value markets outside the United States. And where are we in that cycle? BCA’s work is instructive: technology capital spending has only just reached the levels, relative to the economy, that it attained in the 1990s, and grafting the current investment boom onto that earlier trajectory suggests it could run for another three to five years. If the 1990s are any guide, this bull market is closer to 1995 than to 2000 — and there is a great deal of road between the two.

Sources: MRB Partners (August–September 2026); Capital Economics (August–September 2026); BCA Research (August 2026); LSEG I/B/E/S.