CEOs are unusually confident in their own companies and unusually cautious about the world outside. That gap is the story: how to turn it into practical advantage. The latest data shows how that tension is shaping decisions in the corner office.
Surveys from 2026 consistently show CEOs expressing confidence in their own companies and caution about the broader economy. The size of the gap varies by survey, but the pattern is clear. The data is less contradictory than it appears. Leaders are betting on what they can influence and hedging against what they cannot.
The gap shows up in survey after survey. KPMG’s 2026 Global CEO Outlook found that confidence in their own organisations’ three-year growth prospects rose to 84 percent from 79 percent a year earlier, while confidence in the global economy fell from 68 percent to 65 percent, its lowest since 2021. KPMG’s separate pulse survey of 100 US CEOs found 83 percent confident in their own company’s growth and 86 percent in their industry’s, against 55 percent for the US economy and 53 percent for the global economy. EY’s 2026 CEO Outlook, based on 1,200 CEOs, described a paradoxical mindset: confident in their own prospects, less so in the global economy. EY’s overall sentiment index slipped from 83.0 to 78.5 in that survey. A Fortune and Deloitte survey from autumn 2025 told a similar story, with 32 percent of CEOs pessimistic about the global economy and only 2 percent pessimistic about their own company.
The picture is not uniform. PwC’s 2026 Global CEO Survey, which questioned 4,454 CEOs in 95 countries between late September and early November 2025, found that only 30 percent were confident about their revenue growth over the next 12 months. That is down from 38 percent a year earlier and 56 percent in 2022, a five-year low. Yet 61 percent expected global economic growth to improve, up from 58 percent. This is close to the reverse of the pattern above. The difference lies in the question. KPMG asks about three-year growth prospects, while PwC asks about revenue over the next year. A CEO can believe in the company’s long-term position and still doubt next year’s top line. The paradox is real, but it depends on the time horizon, and it should not be quoted as a single number. For Brussels readers, that distinction matters when assessing policy exposure.
US sentiment adds a second split. The Conference Board’s measure of CEO confidence fell to 46 in the fourth quarter of 2026 from 52 in the third, its lowest reading since the second quarter of 2025. The index has swung through the year, from 59 in the first quarter to 47 in the second and 52 in the third, a stretch that included the US attack on Iran in late February and higher oil prices. In the latest survey of 131 CEOs, 41 percent expected the economy to worsen over the next six months, up from 19 percent the quarter before, and only 13 percent expected it to improve. The Business Roundtable’s CEO Economic Outlook Index told a different story in the third quarter. It rose 3 points to 94, its fifth consecutive increase and the highest since 2022, against a historical average of 83. The two measure different things. The Conference Board asks how CEOs see the economy and their industries, while the Business Roundtable collects their plans for hiring, capital spending and sales. One index tracks how leaders see the world, the other what they intend to do about it.
Betting on themselves means spending. Even as the Conference Board’s CEOs grew more pessimistic about the economy and their industries, 53 percent said they would maintain their capital investment plans, and 36 percent said they would increase them. Only 11 percent planned to cut. PwC found that more than 40 percent of CEOs have already moved into new sectors, and that among those planning major acquisitions in the next three years, 44 percent intend to do deals outside their current industry. KPMG’s US survey quotes its chief executive describing AI investment as required, not discretionary. The common thread is that leaders who cannot control the economy are investing in what they can control: technology, talent, and portfolio moves.
The risks they list are the ones they cannot control. In KPMG’s global survey, cybersecurity and data security were the most common threats to growth, cited by 31 percent. Global economic uncertainty followed at 28 percent, with geopolitical instability and regulatory pressure each at 26 percent. The Conference Board’s fourth-quarter survey had 71 percent of CEOs calling cyber threats high-impact, with AI-related and geopolitical risks at 64 percent and 58 percent. The Conference Board’s economist suggested the quarter’s weaker mood may reflect rising geopolitical tension and elevated oil prices. The pattern is consistent: confidence rests on levers inside the company, and unease attaches to forces outside it.
Self-belief can become a blind spot. What follows is our reading, not a study finding. Confidence in one’s own company is natural. Still, it is also hard to test, and much of a company’s results depend on things outside the building: energy costs, customer demand, supply chains, security and regulation. PwC found that CEOs spend 47 percent of their time on issues with a horizon of under a year and 16 percent on decisions beyond five years, which leaves little room for the long view that three-year confidence implies. The Conference Board also found 28 percent of CEOs seeing conditions in their own industry worsening, up from 13 percent a quarter earlier. Confidence in the company may be less insulated from the outside world than the headline gap suggests.
What leaders can do with the gap. The following steps are also our interpretation of the evidence. First, write down the external assumptions behind the company’s forecast, such as energy prices, interest rates, trade conditions and demand, and test the plan against at least one adverse case. Second, keep investing in the levers the company controls, but release capital in stages tied to evidence, so that confidence is checked as it goes. Third, protect time for long-horizon decisions, since the surveys suggest the calendar is dominated by the next 12 months. Fourth, treat cyber resilience as a condition for growth, not a cost centre, given that it tops KPMG’s list of threats. Fifth, be candid with boards and employees about what the confidence is based on: the company’s ability to adapt, not a forecast of the world.
Read the surveys with care. The samples differ. The Conference Board and Business Roundtable surveys cover large US companies, while KPMG, EY and PwC survey global CEOs, with different dates, wording and time horizons. Brussels readers should note that those differences affect how policy risks and economic views are read. Confidence is also sentiment, not performance, and leaders’ optimism about their own firms does not guarantee results. The sensible use of these figures is to see where views converge and where the question has changed the answer.
What this means: the confidence paradox is less a contradiction than a division of attention. Chief executives believe in what they can influence and worry about what they cannot, and the best of them will use that clarity to decide where to invest and what to protect. The leaders who turn the gap into growth will be the ones who keep testing the company’s confidence against the world it operates in.
Every 2027 plan rests on at least one external assumption. The real test is whether leaders know which one matters most, and how they will respond if it fails.