AXIOMA ROOF™ SCORE HIGHLIGHTS

Few technologies have attracted as much capital as artificial intelligence. From the outset, AI was backed by some of the deepest pools of capital in financial history, and investors were willing to finance the build-out at valuations that assumed not only technological success but rapid commercial success. The promise was enormous, but so too was the confidence that revenues, profits and cash flows would emerge quickly enough to justify the extraordinary sums committed in advance.

The market rout of the past two weeks suggests investors are beginning to reassess AI’s timeline. Most still appear willing to believe that AI will transform industries and economies, but what has become harder to ignore is the possibility that the path from technological breakthrough to profitable monetization may take longer and be more expensive than was widely assumed at the start of the year, when investors expected interest rates to keep falling and capital to become cheaper.

Instead, borrowing costs have moved higher, making it more expensive to sustain an AI build-out financed largely by debt while investors wait for revenues to arrive. Companies such as CoreWeave and Oracle have committed enormous sums in anticipation of future demand, effectively borrowing against revenues that have yet to materialize. Every quarter of delay now carries a higher cost than investors expected just a few months ago, increasing the pressure on companies whose financing needs remain immediate while their revenues remain delayed.

Seen through that lens, investors may be starting to ask if AI risks becoming a Cinderella story in reverse. Not because the magic is fading, but because the deadline was always unrealistic. Cinderella did not choose midnight; investors did. In their eagerness to own the future, they may have compressed decades of potential innovation into a few years, assuming that cheap money would remain available until the rewards arrived. The magic may still be there, but what has changed is the value of time, and with it the confidence that midnight remains far, far away.

The lesson is an old one: time has a variable cost. Financial history is littered with examples of investors who correctly identified transformative technologies but underestimated the time required for those technologies to reward their backers. The deterioration in investor sentiment over the past two weeks suggests growing concern that the gap between today's financing costs and tomorrow's revenues may be widening rather than narrowing. Investors are becoming less certain that some of the companies financing the AI build-out in their portfolios can comfortably survive long enough to reach their happy-ever-after ending before the clock strikes midnight.

That reassessment has left a visible fingerprint on investor sentiment. Aggregate ROOF sentiment deteriorated from Neutral a month ago to strongly Negative today, with last week's decline confirming the broader monthly downtrend. More telling than the aggregate score is the migration underneath it. A month ago, sentiment among investors in three of the ten markets we track remained Positive while only one market had crossed into Bearish territory. Today, investors in five markets are Bearish, Negative in another three, and only US and Australian investors remain Neutral.

The deterioration in sentiment was concentrated outside the United States. Sentiment among UK, Developed Markets ex-US, and European Developed Market investors deteriorated materially over the week and now sits firmly in Bearish territory, while investors in Emerging Markets and APAC ex-Japan remain among the most pessimistic in our investment universe after substantial deteriorations over the past month. By contrast, sentiment among US investors improved modestly over the week and the balance between risk tolerance and risk aversion remains close to equilibrium.

Investors did not all reach the same conclusion, they merely became worried about the same thing at the same time. The compressed nature of sentiment across markets reinforces that point. Cross-market dispersion remains near the bottom of its historical range at the 12th percentile, a level last seen in April and one that has persisted for more than a month. Dispersion widened modestly over the past week, but investors across our investment universe continue to exhibit an unusual degree of agreement about what dictates their risk appetite.

The week ahead will give investors an opportunity to revisit both sides of the equation. Earnings from Microsoft, Alphabet and Amazon will offer fresh evidence on when AI spending might translate into profits, while policy decisions from the Federal Reserve, European Central Bank, Bank of Japan and Bank of Canada will help determine the cost of waiting for that moment to arrive. GDP releases, inflation data from both sides of the Atlantic, elections in Brazil and Israel, and the approaching US midterms should ensure no shortage of competing distractions.

The week ahead will help investors decide whether midnight is still comfortably in the distance or closer than they thought. Faith is patient. Debt is not.

Aggregate ROOF Trend: The black line is the aggregate mean of ROOF scores across the ten markets; the shaded grey band shows the range from the most bearish to the most bullish market on each day, indicating the dispersion of sentiment across the universe. Reading it: a rising line indicates increasing risk tolerance; a declining line indicates increasing risk aversion. 

Bottom strip: cross-market dispersion percentile — the cross-sectional standard deviation of ROOF across the ten markets, percentile-ranked against its own full history. Shaded zones mark elevated (>80th percentile) and compressed (<20th) readings. Reading it: a spike into the top shaded zone is a material divergence episode (sentiment moving in different directions across markets), a slide into the bottom shaded zone is a material convergence episode on a common driver (sentiment moving in the same direction across markets).

Regime Distribution: Three horizontal stacked bars show how the ten markets are distributed across the five regimes — today, one week ago, and one month ago. Each bar is colored by regime: red (Bearish), yellow (Negative), grey (Neutral), blue (Positive), green (Bullish). The white vertical marker on each bar shows the aggregate ROOF score for that period. Tick marks below the bars align with the regime thresholds at ±0.20 and ±0.50. Reading it: a leftward shift over time indicates broadening deterioration; a rightward shift indicates broadening improvement. Widening tails signal rising dispersion.

Market Monitor: A table showing, for each market: current ROOF score, current regime, weekly and monthly changes, regime one month ago, and distance to the nearest extreme threshold. Color cues highlight regime deteriorations (red) and improvements (green). Reading it: scan the "1M Ago Regime" column for markets that have crossed regime bands over the past month; check the "Distance" column to find markets close to the ±0.50 extreme thresholds.

Correlation Matrix (Δ1M): A heatmap of correlations between monthly sentiment changes across all market pairs, over a rolling 60-day window. Green cells indicate that investor sentiment in the two markets changed in the same direction; red cells indicate opposite directions. The rightmost column shows each market's average correlation with the rest of the universe. Important: these are correlations of sentiment changes, not market returns or ROOF levels. High correlation means investor mood in the two markets moved in sync over the past 60 days.