Why Pittsburgh and why now

This is PennInvest Group's market commentary as of August 31, 2026. Part of the material is verifiable facts with reference to the source; part is our forecast, that is, an opinion that may not come true. We separate these two things in the text and ask that they be read differently.

One line abstract. Pittsburgh is trading for about 60 cents on the national dollar - while the world's largest technology companies and the construction of a 3 gigawatt data center are pouring tens of billions into the region. Nashville and Denver once had exactly the same gap. The macrocycle that has kept housing flat since 2022 is now, in our estimation, closer to the end than the beginning. We believe the gap will close and then reverse.

To our investors. You asked that the argument be presented in concrete numbers, not adjectives, and that it be deeply researched, not just stated. That's exactly what's presented here: six related arguments, each backed by numbers we can substantiate and back up with sources, from the simple fact of the price gap to the specific companies and capital remaking the city right now. Together they describe a market that is cheap for reasons that are actively disappearing.

Argumentation in six steps:

• 1 - Break. Pittsburgh sells for about 60% of the national median—as Nashville and Denver once did.
• 2 - Swing. Gold and housing alternate in long liquidity cycles: gold rose, housing waited.
• 3 - Turn around. The four-year housing plateau coincides with a regime change at the Fed.
• 4 - Engine. Google, Amazon, Microsoft and Meta are all expanding in and around Pittsburgh.
• 5 - Foundation. A 3 GW data center in New Kensington and a $90+ billion energy and industrial revival.
• 6 - Support. The most affordable major metropolitan area in the United States with a young, growing job-driven buyer base.

01Gap: About 40% discount to the rest of America

Let's start with a single number that sets the frame for everything else. A typical home in the city of Pittsburgh sells for approximately $245,000 - $260,000. A typical home in the United States sells for approximately $400,000 - $429,000, depending on the method. That's a discount of nearly 40% for a city with two world-class research universities, a major-league sports culture, and a diversified economy in health care, education, finance, and—increasingly—artificial intelligence.

01.1What exactly is the median: city, agglomeration and what we measure

It's easy to get confused here, so let's call things by their proper names. The range of $245,000 - $260,000 applies to the City of Pittsburgh. The "Pittsburgh Market by the Numbers" section on the site's home page lists a different figure: $277,137, which is the median sales price for the Pittsburgh metropolitan area for July 2026, according to the Redfin Data Center. The agglomeration includes suburbs and is therefore more expensive than the city. The third figure that appears in our materials is approximately $240,000, which is the median listing price, which is the asking price in listings, according to Realtor.com via FRED. Three different values, three different methods, none of them cancels out the others. Further in the text we always write which one we mean.

Disruption itself is not a thesis - a city can be cheap because it is dying. The thesis is that Pittsburgh is cheap for reasons that are actively unfolding. Let's look at the precedents of cities that have closed exactly the same gap:

• Nashville was an affordable Southern secondary market for decades, with prices noticeably below the national median, then sharply revalued as healthcare, music industry money and corporate relocations concentrated.
• Denver was a "value" market until the early 2010s, when a wave of technology and population influx lifted it above the national level - which it is now significantly higher.

In each case, the mechanism was the same: a city with strong institutions and low prices attracted high-paying industry, a new wage base boosted scarce housing, and the discount to the country first closed and then reversed. Pittsburgh, in our estimation, is at the beginning of the same curve, and arguments 4 and 5 name specific companies that are driving it.

Why this is important to our work. Buying at a deep discount to the national median means we are entering a market where there is no price bubble to burst. This is not the same as capital protection: property prices can fall even in cheap markets, and no discount will prevent this. The point is different - the growth potential here is tied to the revaluation that has already happened twice this century in other places. We are buying the gap rather than betting that it will persist.

02Liquidity swing: gold and housing alternate

Now about macro timing, because the entry point is no less important than the entry price. Over the long term, the two main stores of household wealth—gold and residential real estate—follow each other. This is seen most clearly through a single relationship: how many ounces of gold does it take to buy the average American house. When the ratio is high, gold is cheap and then outperforms for years; when it's low, gold is expensive and housing leads. This is one of the most consistent patterns of mean reversion in macroeconomics—but it is a pattern, not a law.

The story is specific. At gold's peak in 1980, the average home required about 100 ounces. By 2001, gold fell to about 450 ounces per home and then rose about 600% as U.S. housing peaked in 2007 and crashed. By 2011-2012, the relationship flipped - gold became expensive, about 88 ounces per home - and housing powerfully outpaced gold from 2012 to 2019. The swing has swung back, just like after 1980.

Where we are in 2026. We're deep in gold mode. Gold made a historic surge—about 60% in 2025 alone—to an all-time high of about $5,589 an ounce in January 2026, before correcting to about $4,040 by the end of June. Housing in the United States, meanwhile, has been practically flat for four years. In the language of the swing, gold is stretched and housing is compressed. Historically, this is precisely the configuration that preceded the housing turnaround.

03Late cycle compression comes to an end

Argument 2 says that housing is a lagging asset. Argument 3 explains why the lag, in our opinion, is about to be interrupted. The median price of secondary housing in the United States has been almost flat since 2022 - exactly the year when the Fed began quantitative tightening. Prices soared into 2022 on pandemic liquidity, then stalled when the Fed pulled it out. The four-year plateau is not a peak, but a pause. Prices did not fall; supply remained historically tight; and owners who have fixed a mortgage below 4% refuse to sell, blocking supply. This is a market holding its breath waiting for a reversal in the value of money.

03.1Regime Change at the Fed

Kevin Warsh is sworn in as Fed Chairman on May 22, 2026, and explicitly promises "regime change" at the central bank. His stated agenda, according to financial media reports, points in one direction over time: lower rates, a smaller balance sheet and an overhaul of how the Fed measures inflation. His signature intellectual argument is that AI-driven productivity growth will expand the economy's productive capacity and structurally reduce inflation, which in his logic creates room for rate cuts.

Why hasn't he cut the rate yet? Because he inherited a temporary energy shock. The conflict with Iran in late February 2026 drove up oil and gas prices, pushing headline inflation to a three-year high of about 4.2% and forcing the Fed to pause—and even signal possible hikes—until mid-2026. But here is the key conclusion for our timing: this limitation is external and, in our opinion, temporary. A US-Iran framework agreement with the opening of the Strait of Hormuz would relieve energy price pressures, core inflation is already rising more modestly than headline inflation, and Warsh's structural easing bias would then be unblocked.

The setup laid out straight. A four-year flat housing market, historically tight supply, a wall of buyers on the sidelines, and a new Fed chairman whose explicit thesis - that AI performance justifies lower rates - is now held back only by an energy shock that is already easing. When the value of money turns around, it is the lagging asset that moves, that is, housing. We want to own it before the reversal, not after.

03.2The Lock-in Effect: Why Supply Can't Flood the Market

The most important structural fact in the US housing market today is that people who own houses cannot afford to sell them. During the 2020-2021 boom, an entire generation of owners refinanced to the lowest rates in history. Now they are financially “locked in” in a good way: to sell means to give up the mortgage at 3% and buy again at about 6.3%. The numbers are impressive: About 80% of all US homeowners have a mortgage rate below 6%, about 52.5% have a rate below 4%, and about one in five have a rate below 3%. The average rate on existing mortgages in the US is only 4.4% versus the market rate of about 6.3%.

This is the lock-in effect, and it is what keeps supply historically tight, no matter how strong the demand: an owner with a loan at 3% will not put the house up unless life forces it. It is critical that blocking is resolved asymmetrically. Cutting rates doesn't release a flood of new supply—most owners are still well below any actual new rate, so they stay put—but cutting rates does release a flood of new demand as deferred buyers can suddenly afford the payment. Short supply meeting freed-up demand is a textbook recipe for a price rally.

03.3Demand is here: investors at five-year high

Demand is not hypothetical - it is visible right now, even with rates above 6%, in the behavior of investors just like our company. The investor share of U.S. home purchases has risen to a five-year high of about 30-34%, up more than 8 percentage points for the year by some estimates. And it's not Wall Street: private and small investors owning ten or fewer properties account for more than 90% of investment single-family housing; the largest institutional landlords are about 2%, and they have actually been net sellers for six straight quarters. The capital flowing into rental properties - private, local and operator - is the same profile as PennInvest.

The driver is exactly the dynamic we are positioned for: primary buyers are priced out, more households are renting, rental demand is rising and investors are coming out to meet it - often buying below the listing price, which is where value is created. As Cotality's chief economist put it, fundamentals continue to favor rental demand while affordability is tight for owner-occupiers. We don't wait for the market to form; We are operating in an already active market and buying in anticipation of a rate cut that will expand it.

03.4Compressed Spring: What Happens When the Fed Cuts Rates

Fold the two halves together. The proposal is structurally blocked. Demand is strong, growing and—among potential owner-occupiers—locked behind a single variable: the mortgage rate. The industry's own leading indicators show that this dam is under strain, with pending deals up 3.8% in the month of May 2026 even with rates unchanged, and purchase mortgage applications running about 31% above last year's level. NAR's chief economist calls this "significant pent-up demand" and projects existing home sales could rise as much as 14% in 2026 as rates ease.

Here is the mechanism that closes the argument. The reason rates have remained high is the energetic surge in inflation. Once this resolves and headline inflation turns downward, the Fed will have room to taper, and the new chairman's structural propensity to ease will do the rest. Redfin's chief economist puts it starkly: If things tip toward a recession, "the Fed will have to cut, and you could see a much sharper drop in rates, which could lead to a much sharper rise in home sales." This is the release of a compressed spring: a supply shortage, a wall of waiting buyers and a rate cut as a trigger. We want to already own the properties when it fires.

Timing thesis in one paragraph. Four out of five homeowners are locked in below 6% and won't sell - so supply remains tight. Investors are already buying at a five-year cap to meet rental demand. A large pool of residential buyers is locked behind one variable: the mortgage rate. This rate is high largely due to the energy shock, which we estimate is being resolved. When inflation turns downward and the Fed cuts rates, the supply shortage will be met by freed up demand. Our position is to own assets before the trigger, rather than chase after them.

04Engine: Google, Amazon, Microsoft and Meta are expanding here

Here is the local engine - and this is no longer a story only about promising startups. The reason we expect Pittsburgh to lead the next phase of housing is because the planet's largest tech companies are concentrating high-paying jobs and capital in and around the city. It's the machinery of Nashville and Denver, but driven by four giants of the AI ​​age.

04.1Anchor: Carnegie Mellon and the Talent Magnet

It all starts with Carnegie Mellon University, whose School of Computer Science consistently ranks among the best in the world for artificial intelligence and robotics. In February 2026, CMU opened the Robotics Innovation Center, an approximately 14,000-square-foot facility. m in Hazelwood Green, on the site of a former 19th-century steelworks, with a state-funded Physics AI Accelerator forecast to create 150 jobs in the first year and attract three anchor companies within three years. Its first tenant is robotics company FieldAI. This one building is the whole thesis in miniature: a metallurgical site, a robotics laboratory, high-paying jobs, demand for housing.

04.2Google: a ten-year-old engineering hub

Google has maintained one of its large engineering offices for more than a decade in Pittsburgh's Bakery Square, a converted Nabisco factory in East Liberty, where it employs about 800 people in about 30,000 square feet. Pittsburgh ranks alongside Cambridge among the eastern US engineering hubs that Google's CEO has named for continued investment. His presence is widely credited as the catalyst that transformed Pittsburgh's East End into a technology corridor—and the campus is also home to the University of Pittsburgh and CMU's Software Engineering Institute.

04.3Amazon: Largest private investment in Pennsylvania history

In June 2025, Amazon committed to invest at least $20 billion to build artificial intelligence "innovation campuses" across Pennsylvania—what Governor Shapiro called the largest private investment in the state's history. The first sites directly combine computing with energy: one in Salem Township near the Susquehanna nuclear power plant, the other in Falls Township on the site of a former steel plant. The plan adds about 1,250 high-skill jobs plus thousands in the supply chain — on top of the more than $26 billion Amazon has already invested in the state since 2010. Locally, AWS opened a Cloud Innovation Center with the University of Pittsburgh focused on medical research and sports analytics, the first in the region.

04.4Microsoft and Meta: Energy, Computing and Capital

Microsoft has tied itself directly to Pennsylvania's energy and artificial intelligence sector: it is the contracted energy buyer for the restart of the former Three Mile Island nuclear power plant, renamed the Crane Clean Energy Center with a target date of 2027, which will provide 835 MW of carbon-free electricity for its data centers. Meta, in turn, committed $2.5 million to support early-stage startups at CMU through the Schwartz Center for Entrepreneurship, along with larger commitments to data centers and energy across the regional network. Along with Google and Amazon, all four hyperscalers now have a significant and growing presence in the Pittsburgh area.

Why four hyperscalers matter more than one. When a city is supported by one company, the local economy rises and falls with it. When Google, Amazon, Microsoft, and Meta expand simultaneously—for talent, energy, and land—the base of high-paying jobs becomes structural and diversified. It's exactly the kind of broad, sustained wage growth that lifts housing and closes affordability gaps.

05Foundation: 3 gigawatt data center and industrial renaissance

Beneath the corporate offices lies an even larger layer of physical infrastructure. About 30 miles northeast of Pittsburgh, data center developer TECfusions is transforming the former Alcoa and Arconic research campus in Upper Burrell into the "Keystone Connect" project: a 1,350-acre campus with a master plan for 3 gigawatts of capacity at full build-out, with on-site gas generation and microgrid capability. The first phase is already underway, with full construction expected to run until around 2031, and anchor AI client TensorWave has signed a contract to scale up to a full gigawatt of computing. And again the model is the same: dead industrial site, AI infrastructure, thousands of jobs.

Pittsburgh is also one of the US centers for physical artificial intelligence and autonomous vehicles:

• Aurora Innovation (Nasdaq: AUR), headquartered in the Strip District, with about 1,900 employees and a $14 billion capitalization, has the nation's first commercial driverless trucking on public roads on the Dallas-Houston route, is now expanding to El Paso and Phoenix, and plans hundreds of next-generation trucks in 2026 - in partnership with NVIDIA, PACCAR, Volvo, FedEx and Uber Freight.
• Astrobotic builds lunar landers and space robotics out of Pittsburgh, while Duolingo (Nasdaq: DUOL), a public company that grew up here, is bolstering a growing base of local tech capital and talent.
• The Pittsburgh Robotics Network includes more than 125 companies supporting more than 7,000 jobs, and NVIDIA has established a dedicated artificial intelligence research center with CMU and the University of Pittsburgh.

05.1Energy and Industry Layer

In July 2025, the first Pennsylvania Energy and Innovation Summit in Pittsburgh announced more than $90 billion in commitments for data centers, energy and artificial intelligence. Let's be honest: some of the projects predate the summit, and several large figures have not been independently confirmed - but even with a strong correction, the scale is exceptional. Main projects:

• Homer City Energy Campus is a former coal plant being redeveloped into a 3,200-acre AI and high-performance computing data center campus, fueled by a new approximately 4.4 GW natural gas plant, with first capacity by 2027, about 1,000 jobs on site and about 10,000 in construction.
• Westinghouse's goal of having ten large nuclear reactors under construction by 2030, an estimated economic impact of about $6 billion and about 15,000 jobs in southwestern Pennsylvania.
• Eos Energy is moving its world headquarters to Pittsburgh with an investment of approximately $353 million and approximately 1,000 jobs to produce batteries for artificial intelligence loads, directly referencing the CMU talent pipeline.
• Steel and Heavy Industry - Nippon Steel, Mitsubishi and Hitachi Energy expanded their presence in the Pittsburgh area in 2025, marking a significant renaissance of the region's industrial core.

The strategic logic is simple and robust: AI needs colossal energy, and Pennsylvania is the second-largest energy producer in the country, with an abundance of gas, nuclear and hydro within reach of half the US population. This is why data centers are clustered here. Each of these projects imports construction crews, engineers, electricians and operators who need housing, a perennial demand that's weighing on one of the country's cheapest major markets.

Conclusion for the investor. Arguments 4 and 5 are local fuel for the macro fire of Argument 3. As the Fed cycle unfolds, most US markets will be revalued on cheaper money alone. Pittsburgh, in our estimation, will revalue on cheaper money plus a real, capital-backed influx of high-paying jobs. These are two engines, not one.

06Pillar: Accessibility, Cash Flow and a Younger City

Finally, the demand side. In late 2025 and early 2026, Pittsburgh was named the most affordable major housing market in the United States by Realtor.com and ranked No. 1 in housing affordability among major markets worldwide by Demographia. The reason is the ratio of how much residents earn to how much housing costs.

Two features make this affordability a price underpin rather than a sign of weakness. First, Pittsburgh is the only major US metropolis where it is cheaper to buy than to rent, meaning rental yields directly support purchase prices. Second, according to Zillow's stress test, a median-income Pittsburgh household could still afford the median home even if mortgage rates rose to 8.9%—a margin of safety that essentially no other major market has.

Our own data tells a different story about the same thing. In a sample of eight metropolitan areas as of July 2026, Pittsburgh has the highest gross rental yield and lowest entry price. The gap with Cleveland is minimal, and we show it as it is.

06.1Rent: two different dimensions, and they diverge

Here we have a responsibility to show the contradiction, not hide it. Among the 50 largest U.S. metropolitan areas, Pittsburgh saw the largest six-year rental growth in the country, up 39.9% from 2019 to 2025, just ahead of Tampa at 39.1% and Indianapolis at 32.3%, according to Realtor.com data cited by Newsweek. What makes this figure remarkable is the fact that during the same period, the national median asking rent fell—annualized for 25 consecutive months through the end of 2025.

Our own dataset paints a different picture. According to the Zillow ZORI index, which we publish in the “Pittsburgh Market by the Numbers” section, rent in the Pittsburgh metropolitan area has increased by 25.1% over five years—exactly the same as the national average. Both numbers are correct and measure different things: Realtor.com calculates the median asking rent for the city over six years from 2019, Zillow ZORI is a smoothed index for the entire metropolitan area, including the suburbs, over five years. We don't choose the one that is more convenient. Our data doesn't suggest that rents in Pittsburgh are rising faster than the rest of the country, and we're upfront about that.

The contrast with the Sun Belt is real. Pandemic-era hype cities—Austin, Denver, Phoenix, Nashville, Atlanta—are flooding themselves with new apartments, with multifamily housing starts nationwide reaching a forty-year record of about 600,000 units in 2024. Rents there are now falling: Austin is down about 13% from its peak, Denver is down 7% for the year, Nashville is down 5%, and landlords are offering free months and discounts to fill space. Pittsburgh hasn't built much, leaving supply tight, and the city's median rent rose 3.9% year over year through early 2026, while new leases in the U.S. fell about 1.5%.

Why is this important for a fix-and-hold strategy. The “buy, renovate, rent” strategy lives and dies on rental growth, and the geography of that growth has changed. Sunbelt redevelopment means falling rents and discounts; Pittsburgh's supply discipline plus real growth in high-paying jobs means the opposite. Buying one of the cheapest major markets in the country and collecting rising rent at the same time is the core of the asymmetry: return from day one, rather than betting that prices themselves will save the deal.

06.2Demographic turnaround - honestly

For years, the bearish argument against Pittsburgh was demographic: an aging, shrinking population. The accurate current picture is more favorable than this caricature - and we want to give it with complete accuracy. The city of Pittsburgh has begun to grow again, adding about 4,708 residents from 2020 to 2024 and growing again in 2025 due to domestic migration tied to jobs, with a net influx of about 1,600 people from other regions of the country on top of international arrivals. Realtors note a wave of first-time buyers aged about 30 - this is a key cohort for household formation, driving stable demand for housing.

Fair disclaimer: the broader eight-county metropolitan area is still showing natural decline—more deaths than births, a forty-year echo of the steel industry crash of the 1980s. But that's the point. The direction has turned where it matters to our thesis: the city and its young, well-paid newcomers are growing because the jobs are coming.

07Six arguments put together

Each argument reinforces the next, and together they describe a single, coherent arrangement:

• Cheap (argument 1): Big discount to the national median - the same gap that Nashville and Denver closed.
• Lags (argument 2): In the gold-housing relationship, gold is expensive and housing is cheap—historically a harbinger of the housing decade.
• About to unfold (argument 3): the four-year plateau meets regime change at the Fed, held back only by a waning energy shock.
• With the engine (argument 4): Google, Amazon, Microsoft and Meta are all expanding here, concentrating high-paying jobs.
• On deep foundations (argument 5): 3 GW data center in New Kensington and more than $90 billion in announced energy and industrial investments.
• On a rising pedestal (argument 6): one of the most accessible large metropolises, generating cash flow and - within the city - getting younger.

Synthesis. We're buying into one of the cheapest major housing markets in America, at a lagging point in the macrocycle, just as regime change at the Fed is poised to reverse the value of money—in a city where four of the planet's biggest tech companies and a 3-gigawatt data center boom are importing exactly the kind of buyers who are closing affordability gaps. The growth potential is a revaluation that has happened twice this century elsewhere, but here driven by artificial intelligence, energy and the return of manufacturing simultaneously.

Our conviction. We don't need each of these forces to work perfectly. We need the gap to be real, the cash flow to hold and one or two of those engines to fire. Everyone does part of the work; together they make Pittsburgh, in our judgment, one of the most asymmetrical opportunities in residential real estate in the country today.

08Important Disclaimers

This material has been prepared by Penninvest Group LLC for informational purposes and does not constitute investment, legal, financial or tax advice, or an offer or solicitation to buy or sell any security or interest. Real estate and market forecasts are inherently uncertain; past performance and historical patterns, including the gold-housing relationship and past urban revaluations, do not guarantee future results. Macroeconomic statements, including those regarding Federal Reserve policy, reflect public communications as of August 2026 and are subject to change. Corporate investment figures reflect statements by companies and authorities; Some announced amounts have not been independently confirmed and timing may be subject to change. Market data is obtained from third party sources believed to be reliable - Redfin, Zillow, NAR, US Census Bureau, Carnegie Mellon University, Realtor.com, Amazon, AWS, TECfusions and others cited - and has not been independently verified by us. Prospective investors should conduct their own due diligence and consult their own advisors.