The conference in 5 bullets

• Real estate looks like the bottom of the cycle. Four panels (Global Capital Markets, Private Wealth, Multi-Asset Management, Public vs Private) made the same call from different vantage points: new supply is down 50-75% from four years ago, base rates are coming down, and AI-related demand for power-adjacent real estate (logistics, industrial outdoor, student housing, data-center sites) is real. Public REITs underperformed the S&P in 9 of the last 10 years; the contrarian setup is unusually clean.
• Power generation, transmission, and grid infrastructure are the most consistently flagged under-allocated category for institutional portfolios. TCW cited a need for 250 GW of new US generation in coming years (about Germany plus France combined) and a shortage of 500K electricians annually. Multiple panels endorsed core-plus regulated and contracted power on both equity and credit sides.
• The single biggest unresolved disagreement was the AI-capex equity question. Most CEOs treated trillion-dollar build-outs as a productivity-cycle setup. Apollo's Jim Zelter broke from that consensus to ask whether equity holders will actually capture the returns, citing his own miss on SVB as a reason for humility. Worth tracking which framing wins, because it changes the trade direction.
• Private credit has bifurcated. Sponsor-direct lending was repeatedly described as "vanilla, beta, and saturated." The high-alpha sub-strategies are new-vintage senior with stronger covenants and higher base rates, asset-based finance (especially mezz on the right collateral), private-credit secondaries (the EQT/Coler deal), and rescue finance. The 1-, 3-, and 5-year track records on existing direct lending will look bad as write-downs flow through.
• Public markets are pricing close to zero probability of left-tail events. Goldman Sachs and TCW both said it: equity-risk premium negative, credit spreads as tight as they have been. Combined with the over-allocation to USD assets in most institutional portfolios (even non-US LPs hold around 70% of equity in US names), this is the cleanest argument for adding ex-US (especially Asia) and downside-protected exposure now.

Where consensus is forming

• Real estate at the bottom of the cycle, with the strongest sub-sectors being logistics, industrial outdoor storage, and student housing rather than the office and retail headlines.
• Power generation and grid infrastructure as the under-allocated category, on both equity and credit sides. Core-plus regulated and contracted power is the highest-demand segment.
• New-vintage senior private credit and asset-based finance as the entry points; sponsor-direct lending as a saturated trade.
• 401(k) and retail access to private markets is happening: DOL rulemaking is moving forward, the SEC publicly supports it, and the bottleneck is now manager wrapper readiness.
• Public-private boundaries continue to dissolve. The secondary market is going from $250B last year to a projected $500B in four to five years, and is now an active portfolio-construction tool rather than a distress-only venue.
• Asia is the cleanest case for ex-US allocation. Diana Choyleva's first-time-positive A-share call combines with Gaw Capital Partners' "China-adjacent" Japan, Korea, Vietnam thesis. India demographics and Hong Kong's role as an institutional gateway are reinforcing forces.

Where smart people disagree

• AI-capex equity returns. Most CEOs assumed equity holders capture the productivity gain. Jim Zelter at Apollo did not.
• PE allocation direction. Temasek and EQT are reducing their PE share or rotating it; Bain and Barings are staying full-throttle. The disagreement is real, not stylistic.
• China expression. Diana Choyleva at Enodo wants direct A-shares; Goodwin Gaw at Gaw Capital Partners prefers Japan, Korea, and Vietnam. Both are bullish on the same theme.
• AI's timeline to fundamental change in financial services. Kamal Bhatia at Principal said three to four months. David Gross at Bain said four years. The range on a single panel was wide.
• Whether the US has a retirement crisis. PGIM's David Blanchett says no, citing decumulation behavior as the actual issue. Abacus' Jay Jackson said sharply that the framing is wrong.
• Pricing of public-equity tail risk. TCW says zero probability is priced in. Other panelists were comfortable with current pricing.
• Higher education and AI. Sian Beilock at Dartmouth treats AI as the moment requiring institutional response now. Charles Isbell at Illinois argued AI in 2026 is the consequence of compute-distribution decisions made in the 1980s and 90s, and that we are miscalibrating to the noise.

Top 3 things to consider for the portfolio

1. Add or up-weight power-generation infrastructure (equity and credit) and real estate in logistics, industrial outdoor, and student housing. This is the cleanest cross-panel call. Multiple credible operating platforms made the same case from different vantage points. If one new private-markets commitment is in budget for the next quarter, this is where to look.
2. Reassess the Asia underweight, with a barbell expression. Direct China A-shares (anchored on the Sep 2024 consumer-rebalance and the structural household-savings funnel into equities) plus a "China-adjacent" sleeve in Japan, Korea, and Vietnam (semi supply chain, manufacturing reshoring). Avoid pure China megacap tech. Consumer and financialization themes have higher conviction than industrial exposure.
3. Add downside protection to public-market exposure. TCW's framing of the equity-risk premium combined with the AI-capex equity-return uncertainty argues for explicit tail-protection. Concrete tools: structured products with downside floors (Calamos's 80%/90%/100% protected Bitcoin notes are a useful template for sizing similar wrappers around equity indices), volatility hedges, and tilting public-equity exposure toward secular themes (power, defense, real assets) rather than passive index.

3 most thought-provoking non-investment ideas

1. Michael Crow at ASU described the top third of US households as "Switzerland for 115 million people," doing better than ever on every measure, while the bottom third is declining on lifespan, education, and economic mobility for the first time in modern American history. He framed it as a sustainability-of-the-economic-model question rather than a fairness one. It explains why housing, healthcare, and education policy will keep absorbing capital in the next decade regardless of administration.
2. Charles Isbell at Illinois reframed the AI moment. AI in 2026 is the consequence of compute and connectivity decisions made in the 1980s and 90s. The change worth investing patiently around is what AI is setting up for the 2050s, not the chaos of the current adoption cycle. Worth holding in mind when evaluating any 12-month "AI is going to transform this category" pitch.
3. Yue-Sai Kan and Goodwin Gaw on Chinese consumer brands going global. Pop Mart did $5.4B last year on Labubu (a stuffed toy). BYD is the top-selling car in Singapore. Anta owns Salomon, Wilson, and Arc'teryx and is upgrading them with Chinese manufacturing tech, then exporting from US and Mexico locations. Yue-Sai Kan's anchor on operational density: 250 billion express packages per day in China, a single sunscreen launch doing €6M on day one. Most US and EU portfolios have no exposure to this theme.