News & Stories
Uncertainty is certain for 2026
UNC Kenan-Flagler Business School faculty weigh in on what to expect in 2026 – tariffs, AI, leadership, consumer spending, firm disclosures and more.
The cost of uncertainty about tariffs: John Gallemore
Tariffs were the big U.S. tax-and-trade story of 2025, and they are unlikely to disappear in 2026. The bigger challenge for firms, however, might be tariff uncertainty — changing tariff rates, shifting implementation dates (and delays), exemptions and the potential for retaliation.
My research on firm-level policy expectations suggests that this kind of uncertainty has a real cost: When managers are unclear about the future policy path, firms tend to delay or scale back investment until such uncertainty is resolved. This is a major concern for overall economic growth, and it also undercuts a key goal of the tariffs: inducing firms to move production to the U.S. to avoid the levies. Shifting the location of production is expensive and sometimes irreversible, and if firms cannot forecast whether the current tariff regime will persist, “wait and see” might be the rational choice.
For business leaders, the practical takeaway is to plan for additional tariff changes, while watching for signals that the policy process might be stabilizing. Regarding the latter, the Supreme Court has heard arguments over the current administration’s authority to impose certain tariffs under emergency powers. If the Supreme Court decides to limit the administration’s unilateral authority to impose tariffs, this could reduce one source of policy uncertainty, even if some tariffs remain. This would allow firms to more confidently deploy capital and make long-run changes to their supply chains.
AI feedback for real learning: Elad Sherf
Monday morning, an analyst opens a document and AI is already talking: “Your last three client emails were long, buried the ask and used hedging language.” By lunch, the same tool has summarized her meetings (“You interrupted twice.”), suggested a “stronger executive presence” and offered a checklist for next time. By Friday, she has accumulated a small novel of feedback — most of it plausible, some of it useful, all of it easy to receive because it arrived without the awkwardness of asking a person.
That’s the opportunity AI creates: Feedback becomes cheap, constant and personalized. The trap is subtler. When feedback arrives as a stream of tips, reflection can become a thin ritual —nodding along, tweaking a sentence, promising to be more concise — without ever challenging the assumptions that keep producing the pattern.
AI can be excellent at fueling single-loop learning: small corrections that reduce errors while leaving the underlying logic intact. But the bigger gains often require double-loop learning: stepping back to ask what you are optimizing for, what you are trying to avoid and which “rules of thumb” are quietly running the show.
Forward-looking organizations will use AI to push beyond surface processing. Instead of ending with generic advice, AI feedback can be designed to generate better questions: “What were you trying to avoid by hedging?” “What did you assume the client needed that made you bury the ask?” “If your goal is to sound thoughtful, what’s a clearer way to do that?” Those prompts shift people from editing behavior to interrogating their mental model — where double-loop learning actually lives.
My prediction for 2026 is that we’ll see a split between organizations that bolt AI onto existing HR and feedback routines and those that integrate it into a feedback system built for double-loop learning.
In the first group, AI mostly accelerates noise: more tips, more dashboards, more private coaching — plus more quiet disagreement about what any of it means.
In the second group, AI becomes a bridge to shared sensemaking: It helps employees design small experiments in the world (“Try X twice this week.”), then nudges them to triangulate with the people who can add context and disconfirm comforting stories (“Ask your manager and one peer what they noticed.”).
The difference won’t be access to AI. It will be whether AI is used to produce advice — or to surface assumptions, invite perspective and turn feedback into coordinated action.
Certainly uncertain about an AI bubble: Gil Segal
The year 2025 concluded with the Nasdaq index soaring by 21%, marking its third consecutive year of double-digit returns. That performance was fueled by tech-heavy firms deeply integrated into the AI ecosystem. Over the last three years, the gains have been staggering: Nvidia surged by over 1,200%, Google by 250% and Tesla by nearly 265%.
This raises an urgent question: Are we in an AI-driven bubble?
The warning signs are hard to ignore. The cyclically adjusted price-to-earnings (P/E) ratio has climbed to 40x — a level last seen during the peak of the dot-com era in late 1999. Market concentration is equally extreme, with the “Magnificent 7″ now accounting for 35% of the S&P 500. Furthermore, systemic risks are mounting: “Circular” deals between AI firms raise questions about revenue quality, data center investments have reached trillions in debt-financed capital, and the entire ecosystem faces an energy demand that threatens to outpace the power grid.
Perhaps the most puzzling aspect of this bull market is that it is accompanied by an unprecedented spike in uncertainty. The Baker, Bloom and Davis measures of fiscal and monetary policy uncertainty are currently off the charts, reading two to four standard deviations above historical norms.
Conventionally, high uncertainty cripples economic activity and depresses equity prices — indeed, real investment growth remains negative as of Q3 2025.
Yet, this unusual combination of high uncertainty and high valuation might be rational. By accounting for the sheer variance in AI’s prospects, we can explain why these lofty multiples might not represent a bubble at all.
To understand a P/E of 40, we can use a simple Gordon Growth Model: PE = 1/(r–g)
Where r is the discount rate and g is the nominal earnings growth rate.
- The baseline: If we set r at 9% (4% risk-free plus a 4.5% equity premium) and g at 5% (matching historical nominal GDP growth plus corporate leverage), we get a P/E of 25. This is nearly identical to the post-1990 average of 26.6.
- The average AI boost: Economists at Goldman Sachs and JPMorgan estimate AI will boost GDP by roughly 7% over the next decade, adding about 0.7% to annual growth. Adjusting g to 5.7% brings the P/E to 30.3 — a significant jump, but still far from 40.
The missing piece is the radical uncertainty regarding AI’s actual impact. At one end of the spectrum, Nobel laureate Daron Acemoglu projects a modest 0.15% annual boost. At the other, PwC forecasts a massive 2% annual boost. This implies a standard deviation for AI’s contribution to growth as high as 1.5%.
If we assume two equally likely scenarios, we can see the power of convexity:
- Success (50% probability): g = 7.2% (Average boost + 1.5% Std Dev) PE = 1/(9%-7.2%) = 55.5.
- Failure (50% probability): g = 4.2% (Average boost – 1.5% Std Dev) PE = 1/(9%-4.2%) = 20.8.
The average of these two scenarios is a P/E of 38.2 — remarkably close to the current market level of 40.
Because the P/E ratio is convex in growth, uncertainty acts as a catalyst for value. A “success” scenario has a disproportionately positive impact on prices compared to the negative impact of a “failure.” While high uncertainty usually raises the discount rate (r), this current uncertainty is largely tech-specific and dominates the right tail of potential outcomes with limited systemic downside (“good” uncertainty).
In the current market, uncertainty isn’t a bug: It’s a feature. For the modern investor, uncertainty might just be their best friend.
Comfortable with ambiguity: Jim Johnson and Stephanie Mahin
One of today’s most consequential leadership shifts isn’t about making faster decisions or finding clearer answers. It is about how leaders operate when clarity is unavailable. Across industries and institutions, ambiguity is no longer an occasional disruption: It is the condition under which leadership increasingly happens.
Ambiguity has become structural. Leaders are expected to act while facts are incomplete, timelines are fluid and public scrutiny is constant. In these environments, waiting for certainty is often indistinguishable from inaction. Leaders who build trust are not those who eliminate ambiguity, but can work visibly and responsibly inside it.
This shift has changed how communication functions. Communication is no longer a soft skill or a downstream activity. It is performance infrastructure. It is the system through which decisions are explained, uncertainty is navigated and credibility is sustained over time. When conditions are unsettled, stakeholders look less for polished messaging and more for evidence of judgment: how leaders are thinking, what they are weighing and how they will adapt as circumstances evolve.
Tolerating discomfort, the ability to stay engaged and make sound judgments when clarity is unavailable is now a leadership differentiator. In ambiguous moments, the impulse to rush to closure often reflects a leader’s need for relief rather than the situation’s readiness for resolution. Leaders who resist that impulse and communicate without overclaiming can signal steadiness to those watching.
In 2026, effective leadership will belong to those who can groove in ambiguity by using communication not to manufacture certainty, but to support alignment, trust and forward movement when certainty isn’t yet possible.
The risk of unwinding institutional order: Christian Lundblad
Business leaders spend most of their time debating interest rates, growth, inflation, technology and competition. Those matter. But the far bigger, underpriced risk for 2026 and beyond is institutional. The post-World War II order in which we have lived, with predictable rules, enforceable contracts and credible fora for settling disputes, is an aberration in human history, not a birthright. On the heels of the bloodiest time known to humanity, it was deliberately built through institutions and norms that made commerce easier and conflict costlier.
Here is the controversial claim I will say to friends on the right and the left: No issue is worth the unwinding of our institutional order. We always will have future disputes and injustices to fight over. That is precisely why the rules matter. We can prefer Duke or Carolina, but if we collectively decide the referee is illegitimate, we do not get a cleaner game. We get no game at all, just chaos, retaliation and a permanent rise in the cost of doing business.
To those most excited about a return to a world where power drives outcomes, be careful what you wish for. Capitalism’s superpower, its ability to create opportunity for talent, entrepreneurship, risk taking and old-fashioned hard work depends on predictability, property rights and contract enforcement. When those weaken, investment horizons shorten, financing costs rise, supply chains retrench and innovation becomes more cautious. Last, we should be honest about the arithmetic of the last 80 years. The United States authored much of this system and was the greatest beneficiary of its success. Don’t let anyone tell you otherwise.
The importance of consumer spending: Camelia Kuhnen
Last year, American consumers helped the economy expand at a surprisingly high rate in spite of the very high levels of policy uncertainty in the country (think tariff uncertainty). While households expressed pessimism about the future of the economy, as shown by survey data from the University of Michigan and from the Conference Board, they continued to spend.
Unfortunately, there have been more and more signs of weakness in the labor market, which have been registered by U.S. households. The big question for 2026 is whether we will see a contraction in spending, as households finally start engaging in precautionary behaviors due to concerns about their employment status in the near future.
If we see the rate of layoffs increase or the unemployment rate go up more, American consumers will likely feel very jittery and pull back spending, especially those without significant asset holdings. (Those invested in equity markets should feel the wealth effect: They will spend a bit more this year because they experienced high returns last year.) This could harm economic growth. Let’s see just how resilient U.S. consumers will be this year.
Rethink the firm’s disclosure strategy: Brad Hendricks
Technology is rapidly changing business economics this year. Managers are investing heavily in automation, AI-enabled workflows and cloud infrastructure to lower the cost of coordination, experimentation and scale.
In response, many investors and analysts across industries are forecasting broad margin expansion. But cost reductions don’t automatically translate into durable profit gains, particularly in competitive markets. That uncertainty has shifted the conversation from whether margins will rise to how much, over what horizon and under what conditions.
Firm leaders need to think carefully about how they provide information that helps stakeholders address those uncertainties. Forward-looking disclosures are one of the most effective ways to bridge the information gap with investors and other stakeholders, and credible disclosure is often associated with greater liquidity and a lower cost of capital.
Yet most firms are still focused on short-term guidance around next period earnings or revenues, metrics that reveal little about how technology reshapes long-run economics. Longer horizon targets — such as steady state margins, cost ratios, scalability assumptions or operating leverage — can provide clearer insight into how managers expect today’s investments to translate into long-term performance.
While such disclosures are more commonly used by IPO firms to explain new or evolving business models, structural shifts enabled by new technologies mean that even mature firms are undergoing profound changes to their business models. By complementing traditional guidance with longer horizon financial target disclosures, firms can help investors, employees and other stakeholders understand not just what the firm earned last quarter, but what it is becoming.