AI, Data Centers, and Energy: Where the Bottleneck Really Is

Flex Capital – Investment Perspectives | December 2024

Founder's Note: At Flex Capital, we focus on identifying constraints before they show up in prices. Cycles are shaped less by narratives than by bottlenecks, and disciplined capital earns its edge by positioning where demand becomes inelastic and supply cannot respond quickly.

For a long time, data centers felt predictable. Cloud adoption drove steady demand, power was assumed to be available, and capital flowed freely. If you believed in cloud, the rest largely took care of itself. That framework broke the moment AI entered the system—not because of hype, but because of physics.

AI workloads are fundamentally more energy intensive. Traditional enterprise data centers typically operate at 5–10 kW per rack. AI training clusters now routinely exceed 40–80 kW per rack, with leading-edge deployments pushing beyond 100 kW. A single hyperscale AI campus can require 300–500 MW of continuous power—roughly equivalent to the load of a mid-sized city.

The constraint is no longer compute. It is energy, land, and time. In the U.S., power interconnection queues now exceed 2 terawatts of proposed generation capacity—more than double the country's installed base. In core data center markets such as Northern Virginia, Dallas, and Phoenix, vacancy has fallen below 2%, yet new supply is constrained not by capital availability, but by grid access, transmission timelines, and permitting friction. Power has become the gating factor.

This shift is already reshaping energy markets—particularly natural gas. As AI-driven electricity demand accelerates, natural gas has emerged as the marginal and most reliable fuel source capable of scaling on data-center timelines. According to the EIA, over 40% of U.S. electricity generation already comes from natural gas, and incremental load growth through 2030 is expected to be met disproportionately by gas-fired generation due to its dispatchability, lower capital intensity, and permitting advantages relative to nuclear or large-scale renewables. Hyperscalers increasingly rely on gas-backed grids or direct gas-to-power arrangements to ensure uptime, even when renewable power purchase agreements are layered on top. This dynamic is beginning to tighten regional gas markets, particularly near major load centers, and is shifting capital back toward long-life, low-decline natural gas basins where supply reliability matters more than spot pricing. In other words, AI is quietly converting natural gas from a cyclical commodity into a strategic infrastructure input—supporting sustained drilling activity and midstream investment even in a volatile price environment.

This changes asset evaluation fundamentally. Energy costs, historically around 20% of operating expenses for data centers, now represent 40–60% for AI-oriented facilities. Power is no longer just an operating input—it is the asset itself. Markets with available substations, transmission redundancy, firm gas supply, and regulatory alignment are capturing pricing power, while others remain stalled despite overwhelming demand.

At Flex Capital, we view this as an infrastructure repricing rather than a technology cycle. When demand becomes inelastic and supply is slow-moving, value migrates toward the bottleneck. AI models will continue to improve efficiency over time, but infrastructure timelines do not compress. That mismatch—between exponential demand growth and linear supply response—is where durable, risk-adjusted returns are being created.

Happy Investing,
Flex Capital

Hotels After COVID: What the Operating Data Really Says

Flex Capital – Investment Perspectives | March 2025

Founder's Note: Hospitality cycles don't move in straight lines—they overshoot in both directions. The most compelling opportunities emerge when sentiment lags fundamentals, and when operating data quietly contradicts the prevailing narrative.

In 2020, hospitality was written off entirely. Revenues collapsed, leverage looked terminal, and business travel was declared structurally obsolete. Capital exited fast—and often indiscriminately.

Four years later, the operating data tells a very different story. According to STR, U.S. hotel ADR surpassed 2019 levels by more than 15% by 2023, with RevPAR exceeding pre-COVID peaks across many Sunbelt and leisure-oriented markets. Occupancy recovered more gradually, but margins improved meaningfully as operators adopted leaner staffing models, reduced service intensity, and enforced pricing discipline. What was once an efficiency experiment became a permanent reset.

Demand didn't disappear—it reconfigured. Leisure travel led the recovery, followed by sports, medical, and localized group demand. Business travel returned selectively, but in a different form: shorter stays, higher rates, and more intentional trips. The industry now produces fewer room nights, but materially higher revenue per occupied room.

At the same time, a new demand engine has quietly emerged. The rapid expansion of AI infrastructure, data centers, and advanced manufacturing corridors is creating durable, underappreciated lodging demand. These developments bring a steady flow of engineers, technicians, contractors, vendors, and executive teams—often on rotating schedules and extended project timelines. Unlike traditional corporate travel, this demand is less discretionary, less seasonal, and often rate-insensitive due to enterprise reimbursement structures.

Markets adjacent to major data center clusters, semiconductor fabs, cloud infrastructure hubs, and AI research campuses are already showing distinct lodging patterns: consistent weekday occupancy, premium ADRs, and repeat corporate users without the volatility of convention-driven demand. Hotels near these corridors function less like cyclical hospitality assets and more like mission-critical infrastructure supporting digital and industrial expansion.

Meanwhile, new supply remains constrained. Construction costs are up 25–35% versus 2019, financing remains tight, and underwriting has turned conservative. In many high-growth markets—particularly those attracting tech infrastructure investment—new hotel supply is running below long-term averages despite strong population, employment, and capital inflows.

At Flex Capital, we evaluate hotels as operating businesses anchored in tangible assets. Post-COVID winners are not generic boxes, but well-located, differentiated properties aligned with durable demand drivers. Upscale select-service and boutique assets are especially well positioned—offering pricing power and margin expansion without the cost burden of full-service operations.

Hospitality remains cyclical—but cyclicality is not the risk. It is the opportunity. Particularly when capital re-enters late, narratives lag data, and structurally improved assets continue to be priced as if conditions never changed.

The data is clear. The cycle has shifted. And in the shadow of AI and digital infrastructure, a new class of hotel demand is just beginning to be understood.

Happy Investing,
Flex Capital

Commercial Real Estate After COVID: A Capital Structure Reset

Flex Capital – Investment Perspectives | May 2025

Founder's Note: We distinguish between asset risk and balance-sheet risk. This piece reflects our view that cycles rarely destroy real estate demand—but they often expose capital structures that were built for a different regime.

Commercial real estate did not collapse during COVID. It was stress-tested.

Years of low interest rates masked weak fundamentals and aggressive capital structures. Assets were priced for liquidity rather than durability. COVID exposed the mismatch.

Office vacancy across major U.S. markets now averages roughly 18–20%, but dispersion is extreme. High-quality, well-located buildings remain occupied, while commodity stock struggles. Retail followed a similar path: necessity-based centers stabilized quickly, while discretionary formats continued to erode. Multifamily performed operationally but repriced sharply as interest rates reset valuations.

The common thread was not asset type—it was leverage. Short-term debt financed long-duration assets under the assumption of perpetual refinancing. When rates moved, equity disappeared even where operating cash flows remained intact.

At Flex Capital, we separate asset risk from capital structure risk. Many properties remain fundamentally sound but are constrained by over-levered stacks and near-term maturities. That dislocation creates opportunity for patient capital willing to engage complexity.

Commercial real estate recoveries are balance-sheet driven and time-dependent. Assets will reprice not because they are obsolete, but because their capital stacks no longer work. Selectivity, conservative leverage, and conviction in underlying use now matter more than cycle timing.

Happy Investing,
Flex Capital

Airline Leasing: Essential Capacity in a Constrained System

Flex Capital – Investment Perspectives | May 2025

Founder's Note: We focus on essential infrastructure embedded in global systems, where supply cannot be created quickly and replacement is uncertain. Our approach prioritizes assets with contracted cash flows, structural protections on the downside, and the flexibility to perform across cycles.

Airline leasing is often misunderstood as a bet on passenger growth. It is not. It is a bet on capacity discipline, replacement cycles, and asset scarcity.

Global passenger traffic has largely recovered, with 2024 volumes approaching or exceeding 2019 levels in many regions. Yet aircraft supply has not kept pace. OEM backlogs exceed 10,000 narrowbody aircraft—nearly a decade of production—while engine reliability issues, labor shortages, and supply-chain disruptions continue to delay deliveries.

As a result, mid-life narrowbody aircraft have become more valuable, not less. Lease rates for in-service narrowbodies are up 20–40% versus pre-COVID levels despite higher interest rates. Airlines need lift today, not promised efficiency years from now.

Leasing reallocates risk away from fuel prices and ticket demand toward counterparty credit, maintenance discipline, and residual value. Properly structured leases include maintenance reserves, return conditions, and insurance protections that preserve downside while retaining upside.

At Flex Capital, we view aircraft leasing as infrastructure investing. Narrowbody fleets are embedded in global mobility and commerce. When new supply is constrained and replacement cycles lengthen, the value of existing, serviceable aircraft rises. That dynamic—rather than travel sentiment or macro headlines—is what ultimately drives returns in aviation leasing.

Happy Investing,
Flex Capital

Investing in Retail Strip Malls: The Asset Class That Refused to Die

Flex Capital – Investment Perspectives | August 2025

Founder's Note: Retail is often discussed as a single category. That is a mistake. Cycles do not eliminate demand—they redistribute it. Strip malls illustrate how everyday necessity, location, and capital discipline continue to matter.

Retail was declared dead long before COVID made it fashionable to say so. E-commerce, shifting consumer behavior, and department store bankruptcies drove a decade-long narrative that physical retail was structurally obsolete. COVID appeared to confirm the thesis—stores shuttered overnight, rent collections collapsed, and capital fled the sector. Yet, as with most real estate cycles, the data tells a more nuanced story.

The retail that failed was not retail broadly. It was discretionary, overbuilt, and poorly located retail—particularly enclosed malls dependent on apparel anchors and destination foot traffic. Strip malls anchored by necessity-based tenants followed a very different trajectory.

Grocery stores, pharmacies, discount retailers, quick-service restaurants, medical clinics, fitness concepts, and personal services proved resilient because they sit closest to daily life. You can delay buying a suit. You cannot delay groceries, prescriptions, or a haircut indefinitely—no matter how good your Zoom camera is.

During COVID, necessity-based retail centers recovered faster than almost every other commercial real estate asset class. By late 2021, national strip center rent collections exceeded 95%. According to CoStar, vacancy for neighborhood and community retail centers fell below pre-COVID levels by 2023, reaching roughly 6–7%, while enclosed malls continued to experience double-digit vacancy in many urban cores.

The recovery was not driven by leverage or financial engineering. It was driven by operating fundamentals.

Supply is structurally constrained. New strip mall development remains limited due to zoning restrictions, high construction costs, and lender conservatism. Unlike multifamily or industrial, retail has not overbuilt its way into trouble this cycle. In many suburban and infill markets, replacement cost is now meaningfully above existing valuations, creating a natural floor under pricing.

Tenant demand has also evolved favorably. Retailers today prioritize smaller footprints, high-traffic visibility, and proximity to rooftops rather than regional draw. Omnichannel strategies have made physical locations complementary to digital sales, not competitive. Many tenants now underwrite stores as last-mile distribution nodes, marketing platforms, and customer acquisition channels rolled into one.

Strip malls benefit disproportionately from this shift. They are easy to access, visible, and embedded in residential corridors. They trade architectural prestige for functionality—and functionality compounds quietly over time.

At Flex Capital, we underwrite retail strip malls as service infrastructure, not discretionary commerce. The value lies in tenant mix durability, daily-use relevance, conservative leverage, and long-term land positioning. These assets rarely make headlines, which is precisely why they tend to be mispriced.

Retail strip malls are not exciting but they collect rent through cycles, re-tenant efficiently, and age far better than narratives suggest.

In a world chasing the next transformation story, necessity-based strip malls continue doing what they have always done—serving people where they live. Boring, perhaps. Effective, unquestionably.

And as a bonus, they survived both Amazon and sweatpants!

Happy Investing,
Flex Capital

The Warehouse Shortage: When No New Supply Becomes the Opportunity

Flex Capital – Investment Perspectives | November 2025

Founder's Note: Industrial real estate is rarely dramatic. It is the quiet infrastructure beneath global commerce—functional, efficient, and often overlooked. But every cycle has a moment when the mundane becomes strategic. In 2025, that moment belongs to warehouses. Not because demand is surging, but because new supply has quietly stepped aside. And in real estate, value is reshaped far more by scarcity than by growth.

For most of the last decade, warehouses were built at a relentless pace. E-commerce growth, just-in-time logistics, and abundant capital fueled development across every major corridor. Developers delivered millions of square feet quarter after quarter, lenders competed aggressively, and tenants signed long-term leases with confidence. Even as supply surged, vacancy remained near historic lows. The system appeared self-reinforcing.

That system has now changed.

According to the U.S. Census Bureau, warehouse and industrial construction activity is down more than 20% from recent peaks. In some regions, starts have fallen far more sharply. But the decline is not merely cyclical—it is behavioral. Developers are not just slowing; many have stopped entirely. Financing costs remain elevated, construction pricing has reset higher, entitlement timelines have lengthened, and lenders have grown increasingly cautious on speculative industrial projects.

The most important shift, however, is happening on the tenant side. In a tariff-driven and geopolitically uncertain trade environment, companies are reluctant to commit to three- to five-year leases when trade policy, sourcing routes, and landed costs can change with a single announcement. As one industrial investor put it, you cannot justify breaking ground today if you are unsure whether a tenant's supply chain will look the same twelve months from now. Flexibility has become more valuable than expansion.

And yet, even as new development stalls, existing warehouses are filling up. The Logistics Manager's Index (LMI), which tracks real-time sentiment among supply-chain executives, shows warehousing utilization rising steadily throughout 2024 and into 2025. But this is not the exuberant demand of the e-commerce boom, It is a defensive demand. Importers are pulling inventory forward, accelerating shipments ahead of potential tariff changes, and maximizing the space they already control. They are not taking additional square footage; they are densifying, backfilling, and sweating existing assets.

This creates an unusual tension in the market.

Warehouses are full, but tenants are cautious. Operators are busy, but developers are frozen. Demand is real, but conviction is fragile. And yet, from an ownership perspective, the outcome is unmistakable: vacancy continues to tighten, utilization is rising, and rents are increasing—not because of exuberance, but because there is no slack left in the system.

The LMI confirms that warehouse pricing has increased throughout 2025, driven directly by high utilization colliding with a shrinking supply pipeline. Replacement costs remain materially above in-place valuations, making new construction economically unattractive even before financing is considered.

Thomas Goldsby, a leading supply-chain scholar at the University of Tennessee, summarized the moment succinctly: warehouse operators have realized they can operate at very high utilization, maintain low vacancy, and allow price to do the work. Scarcity, not expansion, is now driving returns.

This is a rare form of tightening. Not one powered by accelerating demand, but by the absence of new inventory. Construction costs remain structurally higher than pre-COVID levels. Capital is selective. Tariff uncertainty discourages long-term commitments. And developers, facing elevated risk with limited visibility, are choosing to wait.

The result is a market where existing warehouses are becoming more valuable simply because the next generation is not being built.

For owners, this is a powerful position. Fewer new deliveries mean less competition. Lease renewals tilt in favor of landlords. Downtime shortens. Pricing power improves. And asset values rise not because sentiment is euphoric, but because the arithmetic of scarcity is unavoidable.

At Flex Capital, we view this environment not as a temporary distortion, but as a structural opportunity. Industrial real estate is entering a phase where the most valuable assets are not those imagined on future pro formas, but those already standing—particularly infill warehouses, cross-border logistics hubs, and assets in markets where replacement cost sits meaningfully above current pricing.

These are not speculative bets. They are scarcity assets in a cycle defined by constraint.

Two things can be true at once. Warehouses can be full because tenants are uncertain. And that very uncertainty can create one of the strongest ownership environments in a decade.

Industrial real estate is tightening. And in real estate, tightening often proves far more durable than growth.

Happy Investing,
Flex Capital

The Preferred That Stopped Being Preferred

Flex Capital – Investment Perspectives | January 2026

Founder's Note: Preferred equity is often described as a "low‑risk, low‑reward" instrument. That framing works—until the market reminds us that the middle of the capital stack is not the middle of the risk spectrum. Cycles don't change the rules; they expose where we misunderstood them.

Preferred equity has long been marketed as the safe compromise between debt and common equity. It offers contractual returns, sits above the sponsor, and avoids the volatility of pure equity. For years, that narrative held up. Then the cycle turned, and the illusion cracked.

Preferred Equity's Safe Haven Years

From 2015 to 2021, multifamily development operated in a world of tailwinds. Cheap debt, demographic momentum, and institutional capital created an environment where even average projects produced extraordinary outcomes. Syndicator's raising capital during this period joked that underwriting was optional. Equity multiples doubled in three years. Construction delays were inconveniences, not existential threats.

In that world, preferred equity looked like the perfect instrument: steady yield, limited downside, and a sponsor eager to protect your position.

But cycles don't ask whether an asset class is popular. They simply reprice risk.

The Reset No One Could Ignore

The 2022–2023 rate shock wasn't a speed bump — it was a structural reset. Capital suddenly became expensive, insurance costs surged, and labor tightened across every trade. Construction inputs turned volatile almost overnight. Lumber became the clearest example: prices exploded more than 350%, rising from roughly $320 per thousand board-feet in April 2020 to over $1,500 just a year later. Framing costs followed the same trajectory, climbing from the mid-$300s to nearly $1,480 per mbf by mid-2021. By 2024, the cumulative impact was unavoidable — the carrying cost of a multifamily construction loan was 40–60% higher than it had been only two years earlier. What had once been a predictable cost stack became a moving target.

And valuations? They didn't drift lower — they snapped to a new reality. A project underwritten at a 4% exit cap in 2021 suddenly faced a 5.5%–6% market. That single shift erased 30–40% of value before a project even opened its doors.

This was the environment in which Flex Capital stepped into what appeared to be a straightforward preferred equity opportunity.

The Investment That Should Have Been Routine

The project was a nearly completed Class A development in a growing metro. It had strong pre‑leasing interest, a clear path to TCO, and a capital stack that looked textbook:

The original budget was $99 million, later revised to $111 million due to inflation and rate shocks — a familiar story in 2024.

• $79 million senior loan
• $27 million common equity
• Finally a Preferred equity in 2025 just below the lender to bridge the final gap on top of the common equity stack

The plan was simple: finish → stabilize → refinance → return capital. Preferred equity was supposed to be the safest seat in the room.

Then the room moved and tables turned.

When the Finish Line Moves Faster Than the Project

Within months, the project encountered a cascade of setbacks:

• New compliance requirements from the city delayed TCO.
• The senior lender tightened its posture as timelines slipped.
• The refinancing market froze for transitional assets.
• And most critically, the valuation collapsed.

This same property which was appraised at $140 million in 2022 by Colliers was now revalued at $90 million in 2025 by the very same third party— the exact same project, not because the building changed, but because the market did.

Suddenly, the capital stack inverted. The senior loan was protected. Everything beneath it was exposed.

The project had two options: a fire sale with millions in loss or a new capital infusion.

Faced with limited alternatives, the sponsors chose survival—the only rational option. They raised $20–25 million of rescue capital for this ready for lease-up property. By necessity, this new capital was structured senior to the preferred equity. Cash flow to the pref investors was paused, the original 18-month timeline collapsed, and the investment horizon reset. What had been underwritten as a short-duration hold is now likely a decade-long position, with returns accruing on paper but no liquidity until a future sale or refinancing.

At the bottom of the capital stack, the original LPs—the project's earliest believers—now face no practical path to recovery until the next bull market.

What This Cycle Re‑Taught Us

Cycles rarely teach new lessons. They simply force us to relearn the ones we softened during the good years.

• Preferred equity behaves like equity when valuations fall.
• Late‑stage construction is still a construction risk.
• Regulatory delays can be as damaging as cost overruns.
• Rescue capital always sets the rules.
• Liquidity assumptions must be conservative, even for "short‑duration" structures.

Preferred equity felt safe when cap rates were compressing. In a normalized market, it behaves exactly as the capital stack says it should.

The Duality of the Present

Two things can be true:

Preferred equity can be a powerful tool in a dislocated market.

And preferred equity can become unexpectedly risky when valuations reset faster than projects can finish.

The genius of the "and" applies here too.

We can acknowledge the pain of this investment while also recognizing the opportunity ahead — a market where disciplined capital and sober expectations will define the next generation of winners.

Happy Investing,
Flex Capital