Summary

Apollo’s slightly-over-$1 trillion business is predominantly investment-grade credit, not private equity. About 80% of AUM is credit; the remaining $200 billion is split roughly equally between hybrid equity and traditional private equity. Marc Rowan’s larger claim is that scale requires serving a “fundamental good”: retirement income, industrial financing, and diversification beyond public markets.
Public portfolios are becoming concentrated expressions of the same technology trend. Ten stocks represent nearly 50% of the S&P 500, while Rowan says global fixed income is about to be dominated by five large banks and five large technology companies. With companies including Anthropic, OpenAI, SpaceX, Cognition, and Cursor still private, he argues that “there’s no place to get” genuine diversification other than private markets.
Apollo considers origination capacity—not available capital—the binding constraint on growth. Unlike a conventional manager that can immediately deploy new money into listed securities, Apollo “can only invest as fast as we originate, as fast as we create.” That scarcity supports a capital-heavy model: retain principal exposure, align with clients, and use the balance sheet to guarantee outcomes for issuers and retirees.
Private markets must acquire public-market infrastructure before they can serve retirement and wealth portfolios at scale. Individuals, insurers, traditional managers, institutional debt and equity allocations, and 401(k)s will not reorganize themselves around drawdown funds, so “we are going to have to conform to them.” Apollo plans daily estimated valuations for its private investment-grade products by June 30 and its entire credit business by the end of September, alongside standardized data, price disclosure to other dealers, data warehouses, and market-making.
AI’s capital intensity is creating a financing opportunity far larger than venture equity can absorb. Rowan cites $800 billion of 2026 capex from only four public companies, before private-company spending, and thinks investor concentration limits will eventually widen spreads. The investable mechanism is to parcel data-center, chip, energy, manufacturing, defense, and robotics projects into venture risk, equity risk, and reusable hard-asset credit rather than financing “every dollar since the invention of fire” with equity.
Rowan expects AI to impair enterprise-software valuations even when the underlying companies survive. Roughly 30% of private equity over the past decade went into enterprise software, often at prices reflecting “a future that did not have AI in it”; he personally expects private-equity returns in aggregate to be “disastrous,” while explicitly saying the conclusion does not apply to every company. For lenders, the response is familiar: diversify, demand seniority or hard collateral where warranted, and underwrite for three, five, or seven years—not 20 or 30.
Apollo’s attempt to remain entrepreneurial rests on institutionalizing rapid admission of error. Rowan says he is right “60% of the time, max,” and employees are not fired for bad decisions but can be for failing to recognize, own, and repair them; every senior professional belongs on the firm’s “wall of shame.” The cultural objective is a durable financial institution built around clean-sheet thinking, intellectual challenge, “do right over easy,” and humane treatment during the moments that shape a career.

Deep dive

1. Drexel taught Rowan to underwrite businesses and invent markets

Rowan joined Drexel in 1984 because financing entrepreneurs and below-investment-grade companies demanded more knowledge of businesses than of conventional finance. Without reliable third parties, “business first” analysis became the foundation of every credit decision.
Markets now treated as established barely existed: high-yield bonds, leveraged loans, ETFs, and securitized products did not exist. PIK instruments, highly confident letters, bridge financing, and other structures emerged through “problem solution, problem solution”—the clean-sheet mentality Rowan says still powers Apollo.
Michael Milken would ask Rowan one question he could not answer at the end of each trading day, teaching him to connect geopolitics, technology, markets, and personalities. The lasting maxim: “You either accept change or change is visited upon you.”
Drexel’s sudden 1990 collapse also supplied Apollo’s risk doctrine. Rowan describes 1990 as a global recession, banking crisis, Texas and New York real-estate crises, and savings-and-loan crisis. Financial firms die from “heart attacks or cancer”: heart attacks come from borrowing short and lending long; cancer comes from accumulating bad assets, which Apollo tries to prevent by admitting mistakes, taking losses, and refusing to double or triple down.

2. Apollo emerged from crisis with improbable scale

After leaving Drexel with a cardboard box amid multiple overlapping crises, Rowan and colleagues kept serving clients without a firm or expectation of payment. A cold call from Crédit Lyonnais initially proposed an M&A boutique; their response was that 1990 looked like “an awesome time to deploy capital.”
The group received $800 million from the French government-owned bank even though its members had never invested money before and the institution itself was not an investor. By year-end it had $6 billion—extraordinary scale for 1990—and subsequently generated Crédit Lyonnais $3 billion-plus annually for several years.
When the bank later needed capital, it sold Apollo to its largest client, François Pinault, whom Rowan says initially thought he was buying portfolio companies such as Samsonite, Culligan, and Vail Resorts rather than an investment firm. Apollo’s track record then enabled a gradual move toward broader institutional capital.

3. The private-equity label obscures Apollo’s actual balance sheet

Apollo now manages slightly more than $1 trillion across retirement services and asset management. Credit represents 80% of AUM, predominantly investment grade; of the other $200 billion, half is hybrid or “partner-like” equity and half traditional fund-based private equity.
Rowan argues a large financial firm needs a societal purpose or regulation and public pressure will constrain it. Apollo identifies three: providing retirement income, financing a global industrial renaissance, and diversifying portfolios whose public holdings have become highly concentrated.
The operating flywheel matches undersaved retirees seeking income with mostly investment-grade companies financing infrastructure, energy transmission, advanced manufacturing, AI, defense, and data centers. Capital demand runs continuously, and Rowan compares Apollo’s position to standing at “First and Main” while traffic flows 24/7.

4. Scarce origination makes capital-heavy alignment an advantage

Rowan rejects AUM as the best measure for an alternative manager: public managers can invest every new dollar immediately, whereas Apollo will leave money undeployed if it lacks suitable originations. “We are not limited ultimately by capital. We are limited by our capacity to create.”
Because an originated asset is the scarce product, Apollo wants both fee income and as much principal upside as markets permit. Clients also value the alignment—“eating your own cooking”—when the manager is a partner alongside them.
Rowan is “unapologetic” in the capital-light versus capital-heavy debate. Brand matters, but so does the ability to guarantee outcomes for issuers, insurers, and people on the retirement-income side; capital enables those guarantees and lets Apollo partner directly with clients.

5. Private credit is being rebuilt for daily-priced portfolios

Alternative assets were designed around one institutional bucket and slow-moving drawdown funds. Apollo now serves five additional markets—individuals, insurers, institutional debt and equity allocations, traditional managers, and 401(k)s—and Rowan calls it “hubris” to expect those markets to conform to legacy private-fund structures.
Apollo instead plans daily estimated values for its private investment-grade suite by June 30 and all credit by the end of September. Pricing alone is insufficient: the ecosystem requires standardized information and identifiers, standardized data warehouses, regular price disclosure to other dealers, and market-making.
Rowan’s directional bet is categorical: “I’ve never seen a market in the world where you have transparency and price discovery that is not 10 times its size and change.” Execution will be imperfect initially, and equity may eventually follow, but “that’s not this year’s business.”
Credit still requires distinct discipline: lenders receive principal and interest rather than equity upside, so they should not generally be around risk-taking and should be fully diversified. Apollo pairs low-cost retirement liabilities with safe long-duration yield—not risky assets inappropriate for a regulated balance sheet.

6. The best assets sit between established allocation buckets

Institutions classify public equities, public fixed income, liquidity, real assets, and alternatives, leaving private but safe credit without a natural home. Apollo calls that hybrid; Rowan says it is the firm’s fastest-growing business because “in between is almost always the best asset class.”
Private investment-grade credit benefits from the same poor capital formation. Banks excel at short-term lending because they fund themselves with short-term deposits; public bonds provide standardized long-term financing, while private capital can structure complex projects involving energy, chips, and offtake.
Large private investment-grade issuers already include Intel, Air France, EDF, AT&T, Meta, and BP Energy. Apollo’s principal underwriting supplies the asset and demonstrates alignment, after which insurers, pensions, endowments, and potentially individuals can participate through its third-party credit business.

7. AI infrastructure will require finance to parcel the risk

Data centers, chips, robotics, manufacturing, and defense are becoming capital-intensive at a scale Rowan calls “unimaginable,” making equity financing inefficient and unable to meet the required scale.
The emerging structure separates company-level venture and equity underwriting from reusable infrastructure and hard assets that can enter credit markets at appropriate returns and ratings. In Rowan’s framing, 2025 proved that chips, data centers, and energy were necessary; in 2026, the market is recognizing the actual quantum of capital.
With $800 billion of capex from four public companies alone, Rowan expects investors to hit concentration limits and thinks spreads will widen. He expects really good entrepreneurs to pair with “financial entrepreneurs” who can democratize credit and hybrid-equity exposure.
Robotics extends the thesis beyond AI compute. If autonomous driving can solve the changing, safety-critical Waymo problem, Rowan reasons that construction equipment should be easier; equipment-rental-style financing can then provide much cheaper and larger-scale capital than venture equity.

8. AI breaks software valuations before it destroys software companies

Rowan’s response to the “SaaS apocalypse” is that “there’s no going back,” while stressing that the view is neither exhaustive nor applicable to every company. If AI weakens enterprise-software credit, the corresponding equity is even more exposed.
About 30% of private equity over the past decade went into enterprise software. Rowan personally expects private-equity returns in aggregate to be “disastrous,” not because every company disappears, but because purchase prices assumed a no-AI future and the prospects of selling those companies to public markets or other buyers have been reduced.
AI advances fastest where an answer can be checked: coding, accounting, and trade operations may see replacement on a “vertical line.” Judgment-heavy work resembles choosing the best Shakespeare essay—improving, but without one verifiable answer—so near-term change is more likely augmentation.
For lending, technological obsolescence is familiar: Yellow Pages, television, radio, cable, satellite, and mobile telephony all looked durable at some point. Rowan’s prescription remains diversification, seniority where risk warrants it, hard collateral, and decisions bounded to three, five, or seven years.

9. Apollo is codifying “right over easy”

Rowan frames his confrontation with Penn as opposition to “favored speech, preferred speech,” not free speech. He says donors’ shift to giving one dollar annually got the university’s attention; after university presidents’ testimony in Washington, D.C., failed to call terrorism murder reprehensible, the university’s chair and president ultimately resigned.
He extends the principle to business: climate policy means “make it better, not worse,” even when that includes hydrocarbon financing, while hiring follows “merit adjusted for distance traveled”—individual adversity and achievement rather than immutable group characteristics. His summary is “we do right over easy.”
Scaling to roughly 4,000 asset-management and 2,000 retirement-services employees forced a six-month negotiation over “what makes Apollo Apollo.” The resulting culture document, posted under Careers, is intentionally candid so recruits can decide whether the firm’s expectations fit them.
“Playing to win” requires resisting the point where fear of losing overwhelms ambition. Rowan says he is right only 60% of the time and “fail[s] quickly and fix[es] it quickly”; bad decisions are tolerated, but refusing to recognize, own, and repair one is not.
What must survive him is clean-sheet thinking, informality, “intellectual insubordination,” and an environment where the right answer wins and hierarchy can be challenged. Apollo also must remain human during employees’ “moments that matter,” because an experience-driven institution works only when partners choose to stay for entire careers.

2026 AI Gaming Landscape: Four Layers, Three Misconceptions | 筱宁 The AI Industry’s Money In, Money Out and Money Made — A Cross-Show Cyber Dialogue