Everything About AI
NextFin News - Harvey Schwartz, chief executive of The Carlyle Group, expects the Federal Reserve to cut interest rates three to four times over the coming year - and he is drawing a line at five. In a wide-ranging interview covering the US economy, Fed policy, and an artificial-intelligence partnership with the Massachusetts Institute of Technology, the head of one of the world's largest private-capital platforms made a calibrated bet: a Fed that eases gradually because growth is stable, not a central bank slashing rates because something is breaking. Data as of September 15, 2026.
The distinction is the whole thesis. Three cuts in a soft-landing environment preserve corporate earnings while lowering financing costs - the ideal backdrop for private capital. Five cuts would signal an economy that "requires a lot of attention from the Fed," a reactive cycle that destroys the earnings denominator even as it lowers the discount rate. Schwartz has long framed the optimal path as "fine tuning," arguing the economy is "in better shape than people are giving credit for." The market usually cheers faster easing as unalloyed good news; he is arguing the opposite, that the boring path is the best one.
Carlyle's vantage point gives the view weight. The firm reported $485 billion of assets under management as of June 30, 2026, spread across private equity, credit, and real assets, with a credit platform alone spanning roughly $203 billion and about 1,000 borrower relationships. Its portfolio companies report payment behavior and cash flow in near real time - a higher-frequency read on the economy than GDP, and one that moves weeks before official data prints. When the owner of that much private-market exposure calls for gradual cuts, it is a signal worth separating from the noise of daily rate-probability swings.
"But the core economic engine globally feels pretty good," Schwartz said, while acknowledging there are "real issues for us to be concerned about."
The Rate Path: Why Four Cuts Is the Ceiling, Not the Floor
The central number in Schwartz's outlook is a range, not a point estimate. His expectation of three to four cuts over the next year sits just below the more aggressive easing some market participants have priced, and the gap is deliberate. A central bank cutting at that pace is normalizing policy after a hiking cycle - bringing rates down from restrictive toward neutral as inflation cools. A central bank cutting five or more times in twelve months is, by definition, reacting to deterioration: falling employment, cracking credit, or a growth shock that markets have not yet fully acknowledged.
This inversion - that more easing can be worse news, not better - is the second-order insight most rate commentary misses. Rate cuts work through the economy with long and variable lags, so the full effect of 2026 easing will not be felt until 2027. The Fed is effectively steering blind for the next several quarters, which is why gradualism aligns with its own data-dependent posture. Cut too fast and inflation re-accelerates, forcing a reversal that shatters credibility. Cut too slow and a preventable recession takes hold. Three to four cuts over twelve months threads that needle.
For Carlyle's credit book, the difference between three cuts and five is not incremental stimulus - it is the difference between a refinancing cycle and a credit cycle. In a fine-tuning environment, borrowers roll debt at moderately lower rates while revenues hold. In a five-cut environment, the same borrowers face falling top lines just as their debt service resets. Floating-rate structures and covenant protections matter far more in the second scenario than the headline rate does.
The Real Economy Beneath the Headlines
Schwartz's optimism rests on four concrete reads rather than a model: the economy is doing well, inflation appears under control, government policies are being implemented effectively, and cash flows at Carlyle's portfolio companies are increasing. That final point is the one that carries the most information. Public companies report quarterly and restate; a diversified private-market platform sees payment behavior across hundreds of companies in near real time. Rising cash flows across a base exposed to industrials, services, healthcare, and technology indicate that the consumer and the corporate capex cycle have not cracked - even as headlines fixate on tariffs, government shutdowns, and geopolitical shocks.
The counter-read is that private-equity cash flows are managed, not market-clearing. Portfolio owners can cut costs, delay maintenance, and draw committed credit lines to smooth earnings through a downturn. That caveat is legitimate, but it cuts both ways: if cash flows are rising while costs are held disciplined, the underlying revenue trend must be at least stable. Managed earnings that are still growing are not a warning sign - they are evidence that demand is doing the heavy lifting.
One figure anchors the firm's own confidence. Carlyle has set a target of $200 billion in capital inflows from 2026 to 2028, up from $158 billion over the 2023-2025 period - an increase of about 27%. Institutional and wealth investors do not commit capital on optimism alone; they commit when they believe deployment opportunities will clear their hurdle rates. A fundraising target of that size, from a base of $485 billion in AUM, is a market signal that the investors writing the largest checks expect private-market returns to remain attractive through the rate-cutting cycle.
Geopolitical Risk: The Gap the Market Cannot Price
If the rate call is Schwartz's macro thesis, the geopolitical read is his risk thesis - and it is where he believes investors are most exposed. Markets, he has argued, responded roughly seven times more violently to the collapse of Silicon Valley Bank than to Russia's invasion of Ukraine, even though the invasion was far more consequential. The reason is structural: geopolitical events "are hard to price and they are hard to hedge," so markets underweight them until they arrive, then repricing in jumps.
This creates a persistent asymmetry. Financial risk - rate moves, credit spreads, earnings misses - is priced continuously by liquid markets. Geopolitical risk - wars, supply-chain severance, energy shocks - is priced discontinuously. Portfolios therefore look well-diversified in calm periods and become highly correlated exactly when diversification is needed. For a long-horizon owner, that gap is a source of return: assets that are under-owned because their risk is hard to model can compound at attractive rates while the risk remains unpriced.
The sectors where this plays out are the ones Schwartz has identified as the themes of the next decade: national defense, energy security, and data protection. National security has been redefined beyond defense contracting to include energy independence and data sovereignty, and governments running high deficits will increasingly rely on private capital to fund all three. That is not a cyclical trade. It is a structural reallocation of capital toward assets whose returns are underwritten by government need rather than consumer discretion - a "demand for private capital" that Schwartz has described as set to be "extraordinary."
AI: The MIT Partnership and the Deployment Gap
The third pillar of the interview is artificial intelligence, and here the framing shifts from macro to operational. The MIT partnership discussed by Schwartz fits into a broader build-out at Carlyle: Chief Digital Officer Matt Anderson leads AI, machine learning, and data science across the platform, and the firm has established strategic digital partnerships with OpenAI, including early access to agent-based tooling designed to improve how data and insights inform investment activity. Carlyle has also run AI-focused programs for portfolio-company leaders, drawing on research and frameworks from institutions including MIT.
The significance of an academic partnership is what it is not. It is not a financial bet on AI models, and it is not a claim that frontier capability alone creates value. It is an effort to connect private-market deployment with research on how AI actually changes workflows inside companies - which is the unresolved question in the AI investment debate. That debate has split into two camps: those who see a capital-expenditure bubble and those who see a productivity revolution. Both are partially right, and the gap between them is where private capital operates.
The bubble concern is real at the infrastructure layer, and Carlyle's own credit research documented it. In early 2026, as AI anxiety spread, spreads on software loans widened to 783 basis points versus a broader index spread of 467 basis points - lenders repricing the terminal value of the sector almost overnight. That is the market pricing a bust scenario into AI-adjacent credit. Yet repricing creates opportunity for disciplined capital: Carlyle-backed supply-chain AI firm Exiger announced an AI-native platform upgrade in September 2026, and Carlyle agreed to acquire data-and-cloud AI consultancy Adastra to deepen its transformation capabilities.
The productivity argument rests on adoption, not capability. Research cited in private-markets analysis suggests only about 5% of AI initiatives succeed - a sobering figure implying the winning strategy is not to fund more AI startups, but to fund the companies that can integrate AI into actual workflows. That is the private-equity edge. A public-market investor can buy an AI winner; a private owner can install AI across a portfolio company's procurement, customer service, and back office and capture the margin expansion directly.
Alongside the opportunity sits a governance risk Schwartz has flagged directly: the rise of "vibe-coding," where employees use AI to write software without understanding what it does.
"I don't want people randomly coding," Schwartz said of the companies Carlyle has invested in, framing the issue as accountability - if code is generated outside the engineering function, someone must own the bugs, the security holes, and the technical debt.
For an investor, that is not philosophical. Unreviewed AI-generated code is a liability that compounds silently until it surfaces as an outage or a breach. The firms that win the AI cycle will be the ones that treat governance as part of deployment, not as an afterthought.
The Counter-Thesis: What If the Soft Landing Is Already Priced In?
The strongest case against Schwartz's view is that it is consensus dressed as insight. Markets have priced a soft landing and a gradual cutting cycle for much of 2026; technology valuations embed significant AI productivity expectations; and private-market fundraising targets assume a benign exit environment. If the base case is already embedded in prices, the upside to being right is limited while the downside to a surprise is large.
There is force to this objection. Equity markets have rallied on precisely the expectation that the Fed will cut without a recession, and AI-related capital expenditure is running at a pace that assumes demand will absorb the new compute capacity. A portfolio positioned for fine-tuning is vulnerable to two shocks: inflation that refuses to fall, forcing the Fed to hold or reverse course; or a growth shock that converts three cuts into five for the wrong reasons.
The answer is that Schwartz's position is an operating thesis for where private capital deploys, not a call on the S&P 500 over the next quarter. Private-equity returns are made over five to seven years and are driven more by entry pricing, leverage terms, and operational improvement than by the direction of public indices in any single year. A gradual-cutting environment with stable growth is the ideal backdrop for that model regardless of whether public equities have already discounted it. The real risk is not that the soft landing is priced in; it is that credit spreads are too tight for the geopolitical and AI-adoption risks sitting underneath. That is a spread-duration call, not a directional one.
What Would Break the Thesis
Schwartz's framework is falsifiable, and the falsifying signals are specific. On rates: if the Fed cuts five or more times within twelve months, the "fine-tuning" thesis fails and the cycle has shifted from normalization to reaction - warranting a defensive posture in credit and lower multiples on cyclical private-equity assets. On inflation: if core price pressures re-accelerate for two consecutive quarters, the gradual-cut path collapses and the Fed faces the worst version of its trade-off. On the real economy: if Carlyle's reported cash-flow growth stalls or reverses across a diversified portfolio base, the soft-landing assumption was too optimistic.
On AI, the signal is adoption success. If the share of AI initiatives delivering measurable productivity gains remains in the low single digits over the next two years, the deployment thesis weakens and capital rotates back toward infrastructure winners and away from enterprise adopters. If adoption accelerates while infrastructure bottlenecks ease, the productivity case strengthens and the current wide spreads on software credit look like an overreaction.
For investors, the implication is a barbell, not a single bet. The base case - three to four cuts, stable growth, gradual AI adoption - favors private credit with floating-rate protection, defense and energy-security assets with government-backed demand, and private-equity operators with genuine AI integration capability. The downside case - five cuts or sticky inflation - favors shorter-duration credit and cash. The upside case - faster-than-expected productivity gains - favors the companies that can install AI at scale, not the ones that merely talk about it.
The bottom line: Harvey Schwartz's three-to-four-cut forecast is not a bet on easier money; it is a bet on a Fed that gets the landing right. The market should listen less to the number of cuts and more to the reason for them - because three cuts born of stability and five cuts born of stress produce the same headline but entirely different portfolios.
Explore more exclusive insights at nextfin.ai.
