Investors start every year with a set of assumptions about the key drivers of corporate earnings. Few years end with those assumptions intact, and 2026 has been no exception.
At the start of the year, the case for stronger earnings looked well supported. Central banks were expected to keep cutting interest rates, fiscal stimulus was due to support US demand, European governments had committed to higher defence spending and corporate investment in AI was still rising.
While forecasts pointed to accelerating earnings growth in the US, Europe and emerging markets, the risks were visible. The US jobs market was cooling with more unemployed workers than vacancies, and inflation was expected to remain above central bank targets.
What changed
The conflict in the Middle East pushed oil prices sharply higher, and although they retreated from their peak, they have stayed volatile. Economists' inflation forecasts, which had been drifting lower early in the year, were revised up.
Growth forecasts, which had been raised to optimistic levels, were trimmed. The debate over interest rates moved from how many cuts would come, to whether rates might need to rise, while long-dated government bond yields kept climbing. While this combination would normally weigh on earnings forecasts, we have instead seen them rise.
An unusual pattern
Over the past two decades, expectations for global earnings growth have on average been revised down by c.10% over the course of the year. Analysts tend to be optimistic at the start of a forecast period, which is why forecasts made in January deserve some caution.
However, this year is an exception. Expected earnings growth for global equities has roughly doubled since January, with the MSCI World’s earnings per share growth estimates growing from 16.3% at the beginning of the year to 33.5% at the end of the year. The size of the upward revisions sits at the extreme end of the historical range.
The pattern is not limited to the US: similar upgrades have appeared in Europe and emerging markets. Revisions are useful to watch because they show how expectations are changing, not just where they stand.
Where these earnings upgrades have come from
Energy and technology have been the key drivers of upward earnings revisions. Higher oil prices have lifted earnings expectations for energy companies. But technology has been the larger driver, especially in the semiconductor industry, which supplies the chips, memory and equipment used to build AI infrastructure.
Upwards earnings revisions in technology companies can be traced to capital spending. Forecasts for capital expenditure of the largest technology companies have been revised up repeatedly, and expectations for this year are now well over double what they were little more than a year ago.
The sums are large, with annual spending by the five largest companies now running into the hundreds of billions of dollars, comparable to the output of a mid-sized economy. Of course, one company’s spending is another company’s revenue, and the technology companies supporting the AI build-out are on the receiving end.
Earnings are not the same as cash
This demonstrates why earnings should be read alongside cash flow. The biggest spenders generate substantial operating cash flow, but they are reinvesting almost all of it back into the AI build-out.
On current forecasts, their free cash flow, the cash left after capital spending, will fall sharply and even turn negative for a period before recovering later in the decade. Meanwhile, the suppliers receiving that spending are seeing free cash flow rise significantly.
Both groups can report strong earnings, but the durability and character of that growth differ. Where the cash is going, and how long the spending lasts, matter as much as the headline profit figures.
The risk of a narrow set of drivers
When one theme accounts for a large share of upgrades, aggregate figures can mask how narrow the underlying drivers are. AI's influence is not confined to technology and has reached industrial companies through the demand for electrical equipment, construction and power infrastructure.
Earnings growth that leans heavily on a single theme is more exposed to changes in that theme, whether from shifts in spending plans, regulation or the pace of adoption. Diversifying across sources of earnings growth is the natural response.
Conclusion
Uncertainty is inherent to earnings forecasting. Forecasts are meant to be revised; the focus should be on the drivers of those revisions and on the conditions required for forecasts to be met.
The extraordinary path of company earnings this year has three key lessons. First, forecasts made at the start of a year are provisional rather than authoritative, and the biggest driver of the year may not be on the original list at all.
Second, the direction and scale of changes to expectations can tell investors as much as the forecast itself. Third, with earnings growth now expected to be far stronger than in January, there is more room for disappointment, particularly if inflation and interest rates stay higher for longer.
The dominance of the AI theme over the continued upwards earnings revisions for 2026 and 2027 calls, in our view, for technology exposure selected with an awareness of where the cash is going.
Matthew Page is co-manager on the Guinness Global Equity Income and Guinness Global Innovators strategies. The views expressed above should not be taken as investment advice.