Confronting growing market concerns that trillion-dollar AI capital expenditures could collapse into a bubble, Blackstone Group LP President Jon Gray has answered with striking real-world return data: an efficiency leap in the physical economy is triggering exponential demand, and the acute, tangible shortages of computing power, electricity, and data centers now represent the most critical investment opportunity.

On September 20, during a closed-door presentation to over 100 of the world's top investors, Jon Gray, President and Chief Operating Officer of Blackstone Group LP, unveiled for the first time a comprehensive look at the firm's proprietary data and core logic behind its artificial intelligence (AI) investments.

Addressing deep skepticism in the market over whether massive AI capital spending can deliver returns for investors, Gray borrowed a classic Super Bowl advertising tagline—"Where's the Beef?"—to pinpoint the central pain point of current AI investment. He asserted that AI deployment has already translated into tangible productivity gains and revenue growth for enterprises, and the ensuing demand is now colliding forcefully with bottlenecks in real-world infrastructure.

An "1870 Dawn": Similarities and Differences from Past Industrial Revolutions Looking back at history, Gray compared the current macro environment to the United States of the late 19th century.

"I think we are closer to 1870 today, with a major industrial revolution on the horizon," Gray noted. He pointed out that between 1870 and 1900, driven by the proliferation of steel, railroads, and electricity, annual U.S. labor productivity doubled, GDP grew fourfold, and the stock market rose sevenfold. "I'm not saying the same thing will happen today, but it does show the power of technological change and innovation."

However, unlike the painful lesson of that era—when 200 railroad companies went bankrupt due to "highly leveraged, build-ahead-of-demand" expansion—today's AI infrastructure build-out operates under a fundamentally opposite supply-demand dynamic. "Today's situation is completely different—demand far exceeds supply, and whether it's energy or data centers, most signed contracts are with companies that have very low leverage."

Billions in Revenue and Trillion-Dollar Valuations: The "Beef" Comes from Clear Returns on Investment Addressing the market's biggest worry—whether capital expenditures are just a web of circular financing—Gray used a series of eye-opening data points to prove the "beef" is real:

Explosive Growth in Frontier Model Revenue: Large model companies like Anthropic and OpenAI have seen revenue grow 100-fold in a short period. As of July 2025, their annualized revenue reached a staggering $105 billion. This momentum has driven their combined valuation from $683 billion to over $2.3 trillion.

Accelerating Enterprise Monetization: The 14 AI-related companies in Blackstone Group LP's portfolio have seen annualized revenue surge 21-fold, from $25 million in September of last year to $525 million.

Macro Productivity Leap: The average U.S. annual productivity growth rate over the past decade was 1.5%, but over the last two and a half years, it has reached 2.6%. EBITDA margins for both S&P 500 companies and Blackstone Group LP's portfolio companies have expanded by 500 to 700 basis points over the past four years.

Gray stated bluntly: "Average earnings growth over the past 12 months has been 15%, and even excluding one-time items, it's 32%. There's no doubt that as this technology spreads through the economy, it will have a massive impact."

Real-World "Hard Shortages": A Super Cycle of Compute, Chips, and Electricity As technology moves beyond the desktop into broader physical applications like robotics and autonomous driving, the true obstacles emerge. Gray stressed: "The real bottleneck is in the physical world, in chips, electricity, and data centers."

To meet surging computational demand, global hyperscale data center operators are on an unprecedented capital spending spree. Data shows that the capital expenditures of the five largest hyperscalers have doubled this year to $820 billion, equivalent to 2.5% of U.S. GDP. Blackstone Group LP's own data center leasing is projected to surge from 1 gigawatt in 2024 to an expected 6 gigawatts this year, corresponding to nearly $100 billion in capital expenditures.

Facing such massive investment, market "bubble" concerns are inevitable. But Gray forcefully rebutted this using memory chip giant SK Hynix as an example: "SK Hynix currently trades at just 4 times earnings. This is not 2000, when Cisco had a P/E of 150. There's still great skepticism about the current situation."

Beyond data centers and chips, the electricity bottleneck is equally severe. Building one gigawatt of power capacity requires up to $55 billion, and ordering a single General Electric turbine has a queue stretching to 2031. This capital intensity and incredibly high barriers to entry form the core logic behind Blackstone Group LP's heavy investment in foundational infrastructure.

"The real 'core'? It's chips, data centers, and electricity. And above that, the real core is the return on investment," Gray concluded.

Unignorable Risks and "AI-Proof" Scarce Assets While embracing the AI wave, Gray maintained objective risk awareness. He warned that excessive valuations are a genuine issue, "Companies that are just starting out with no revenue are valued at $10 billion, and some defense tech companies have staggeringly high valuations. These need close attention." Additionally, cybersecurity, potential regulatory constraints on the technological frontier, and geopolitical tensions all loom as risks over the market.

Interestingly, amidst AI reshaping everything, Blackstone Group LP has not limited its focus solely to technology. Gray specifically highlighted the investment value of "scarce assets"—from a cricket team in the 1.4 billion-person Indian market, to specialty coffee chains, to beachfront properties and Rome's airport. "AI can't replace cricket. These experiences won't go away, and these assets are irreplaceable."

Key Excerpts from the Presentation Transcript Jon Gray 00:02

I love this song. It's the 2024 Blackstone holiday video, and the 2026 one is upcoming. I'm thrilled you've all joined us from around the world, thank you. We wouldn't be where we are without your support, and we are eternally grateful. Thank you.

Jon Gray 00:27

Now, we want to talk about the questions you care about most when considering investing significant capital. I think it's best to start with the most important one—how to become a thought leader in global financial fitness.

Jon Gray 00:49

This is very important to me. I have five tips.

First, you need to sweat profusely. Luckily, I was born to sweat, but it makes you look very authentic.

Second, you need to angle the camera up. My daughters always remind me, otherwise you get a terrible double chin.

Third, shamelessly display your brand. Want to feel like a NASCAR driver? Here's a logo for you. You matter.

Fourth, have that goofy dad energy. Perfect! Add five emojis, it shows great enthusiasm and makes filming fun.

Alright, now, let's get down to business.

Jon Gray 01:38

We're going to try to answer four questions this morning.

First, what's happened since most of us last gathered a year ago?

Second, what can we learn from the past, especially the distant past?

Third, where are we today?

Fourth, where are we heading, and what will the future bring?

Jon Gray 02:02

We'll, of course, approach this from the perspective of those two letters: AI. This doesn't mean we're ignoring geopolitical situations in Ukraine and the Middle East—they've driven up commodity prices, causing inflation and higher short-term rates, as we saw yesterday. There are concerns about deficits, long-term rates—all important to us as investors. But when we think about what ultimately drives portfolio value, what's happening in the economy and where it's going, we have to think about those two letters.

So, I want to go back a year, to September 17, 2025, when we were in this room. What were we thinking?

Jon Gray 02:53

What was the main reason for our enthusiasm? Primarily, it was artificial intelligence. Our style hasn't changed much in a year.

Jon Gray 03:12

When we look back at the AI evolution over the past year or two, take Google tokens for example. Volume went from nearly zero in 2024 to 480 trillion tokens per month today. By May 2026, that volume had reached 32 quadrillion tokens—a word I learned for this presentation. By September, these numbers will clearly be even higher.

Jon Gray 03:41

Large model companies like Anthropic and OpenAI had almost no revenue a few years ago. By July 2025, revenue grew 100-fold. As of July, their annualized revenue had reached $105 billion. This is unprecedented in human history.

Jon Gray 04:03

Blackstone Group LP has a great vantage point through our portfolio companies. These figures come from our portfolio companies, our GP-held portfolio companies, and our borrowers. Last September, these 14 companies had annualized revenue of $25 million. Today, that's grown 21-fold to $525 million in annualized revenue. This is what's happening in the market, as these companies are achieving extraordinary returns on investment.

Clearly, this impacts valuations. We were fortunate to invest in these two companies. Their valuation was $683 billion last September; today it's estimated to be over $2.3 trillion combined. As they begin to go public, their strong revenue models are driving value growth.

Jon Gray 05:05

We also discussed a lot last year: not just models and technology, but also the infrastructure needed to deliver them. Global capital expenditure has grown six-fold, and next year these numbers will be even bigger. Last year, the five hyperscalers' capex was $415 billion. This year, that number has doubled to $820 billion, equivalent to 2.5% of U.S. GDP.

It's not just the hyperscalers; we're also very active in data center leasing. We have the largest platform globally. We leased 1 gigawatt in 2024, 2 gigawatts in 2025, and we expect to lease at least 6 gigawatts this year, representing nearly $100 billion in capex, with tenants investing hundreds of billions more in chips.

Unfortunately, we don't invest in memory chip companies. Micron, SK Hynix—these three leading companies in the sector have seen their stocks rise 600% and 500% respectively. When you see charts like that, as an investor, your first reaction is: this feels like a bubble. But SK Hynix currently trades at just 4 times earnings. This is not 2000, when Cisco had a P/E of 150. There's still great skepticism about the current situation.

Over the past year, we've deployed significant capital into financing for GPUs and TPUs, building partnerships with Nvidia and Broadcom, investing heavily in new cloud computing, and creating a joint venture with Google. Through Crux, we've invested with Firmus in Australia, NASA, India, and numerous other data center companies globally. We believe strong demand for computing power will continue to generate solid returns.

It's not just data centers; electricity is becoming increasingly important. We talked about its critical role last year. Over the past five years, utility capex totaled $800 billion, and that's expected to almost double in the next five. I was in Canada this week, and they're discussing investing hundreds of billions to expand and upgrade the grid. This is happening globally.

Stock reactions here aren't as dramatic as memory chip makers, but the power sector used to be very boring. Now, companies making turbines, cooling equipment, and electrical gear have seen their stocks rise 44% to 81% in a year. Fortunately, we've also invested heavily here—we've provided significant credit to Sempra Infrastructure for a large LNG project; invested in energy data analytics company Enverus; invested in Williams to build utilities for power plants and data centers, they're working on the McLean project; and invested in European electrical equipment, wind and solar technology, and renewable energy companies. We've made massive investments in energy transition and infrastructure businesses.

What kind of returns has all this AI-related investment generated? We analyzed our top ten company-wide earnings investments for the second quarter of this year. Nine were focused on AI, including some industry bellwethers, with the only exception being India's Aster Care hospital. We expect the third quarter to look very similar.

Jon Gray 07:30

So, what can we learn from the past, especially the distant past? I want to turn the clock back to September 17, 1870. Most Americans lived on farms, in the post-Civil War era, with almost all buildings made of wood. To visit family in the next town, you'd ride a horse; to read at night, you'd light a candle.

Just 30 years later, everything had transformed. By 1900, many buildings were made of steel; you could take a train to visit family, and cars were appearing on the roads; and thanks to Thomas Edison, you could read in electric light.

I think we are closer to 1870 today, with a major industrial revolution on the horizon. What happened economically from 1870 to 1900? Annual labor productivity doubled, GDP grew fourfold, real manufacturing grew six-fold, and the stock market rose sevenfold. I'm not saying the same thing will happen today, but it does show the power of technological change and innovation, and what they ultimately mean for the economy and for investing.

But even in boom times, there are downturns. The key infrastructure underpinning that era was the railroad. During that period, 200 railroad companies went bankrupt or became insolvent. That's a cautionary tale. At the time, they built supply ahead of demand with high leverage, very similar to what happened later with telecom infrastructure. Today, the situation is completely different—demand far exceeds supply, and whether it's energy or data centers, most signed contracts are with companies that have very low leverage. That's not to say there won't be problems, but the situation has changed.

Jon Gray 13:00

Now let's fast forward 84 years. On January 22, 1984, I was about to turn 14, at Elm Place Junior High in Highland Park, Illinois, watching Joe Theismann's Washington Redskins take on Marcus Allen's Los Angeles Raiders in the Super Bowl. The game wasn't much to watch—final score 38-9. But the highlight was two commercials.

Jon Gray 14:10

The first was Steve Jobs' famous 1984 ad: "On January 24th, Apple Computer will introduce Macintosh. And you'll see why 1984 won't be like '1984'."

That ad was incredibly prescient—it was about democratizing computing. You shouldn't have to use a Cobalt operating system, or wrestle with MS-DOS and arcane commands. You should just be able to click a mouse and have powerful computing at your fingertips.

Jon Gray 14:50

If you think about the metaphor for today, we're now talking about access to intelligence for everyone on the planet—even if you're in the smallest village, you can access the world's most brilliant minds, without being in Cambridge, UK or Cambridge, Massachusetts. That's a very powerful message.

Jon Gray 15:09

But I think the most relevant topic for today's discussion is the question posed by Clara Peller during that same Super Bowl. This 81-year-old woman looked at a huge, fluffy bun and asked a very important question: Where's the beef?

That's exactly our problem—we're investing trillions in capex. What's this all about? Does it really make sense, or is it just a bunch of circular financing where people will eventually lose all this money?

Jon Gray 15:49

That's the question we're trying to answer. I want to come back to our current situation and why I'm confident in all these investments. It all starts with what's happening inside our own firm: AI use cases are expanding rapidly, almost everywhere.

Take process improvement, for example. The IGI diamond grading system in India, and Phoenix Tower, our mobile tower company partner, which now processes leases five times faster. They invested $4 million in developing new AI processes, generating $4.5 million annually in returns—an annual ROI on software engineering of over 100%. Almost every company we work with is benefiting. Enverus has a new code-fixing process, getting 18x returns on their investments in pilot companies—this is why investment keeps growing.

In customer service, Tricon, a rental housing company, can process applications 90% faster. In content creation, advertising, design, visual effects—it's everywhere.

Jon Gray 17:06

But it's not just about productivity. Our portfolio company Chamberlain, a garage door opener company, launched their face-recognition MyQ Secure View 3-in-1 smart lock. A few years ago, they realized they could use AI visual technology to expand into front doors and garage doors for Amazon packages. Today, they have $40.5 million in annualized revenue, and the CEO expects that to reach $500 million in five years. We'll see more and more examples like this over time.

In legal and compliance, marketing review efficiency has improved by 50%. In software development, my personal favorite is the CIO and portfolio intelligence agents—we're seeing 99% time reduction.

Now, these benefits are starting to show up in the data. The U.S. annual productivity growth rate has averaged 1.5% over the past decade, but over the last two and a half years, it's hit 2.6%. Hyperscalers are at the forefront of this trend, with their revenue per employee having grown 65% over the past three and a half years, and margins are rising too. EBITDA margins for S&P 500 companies and our portfolio companies have expanded by 500 to 700 basis points over the past four years.

This is driving very strong earnings growth. The average growth rate over the past 12 months is 15%, and even excluding one-time items, it's 32%. There's no doubt that as this technology spreads through the economy, it will have a massive impact.

Jon Gray 19:42

And now, everything is about to change. Currently, most of this technology is still in our offices. But as it moves off the desktop and into real-world applications like robotics, autonomous vehicles, and national security, the demand for computing power will only increase, and productivity will keep improving.

Take autonomous driving company Waymo. Over the past two and a half years, its miles driven have grown 250-fold, while its serious accident rate has dropped 94%. This trend will continue and have profound implications.

Jon Gray 20:32

Regarding AI's impact on our lives, challenges are inevitable, and society will undergo transformation. But it's also meaningful to talk about some truly positive impacts.

One is job growth. At our data center company QTS, blue-collar jobs are booming. In less than two years, the number of construction workers on site has tripled. Despite AI increasing productivity, our overall headcount is still growing. Over the past decade, the number of new business applications has doubled, and it's much easier to start a company today.

Jon Gray 21:10

The most exciting thing about this technology is in science. A study published in Nature showed that due to AI-discovered drugs, the success rate for Phase 1 clinical trials has significantly improved. AI is certainly better in chemistry; biology still needs improvement, but it's promising. A recent McKinsey study showed AI optimization can shorten a trial's duration by 40%. Our portfolio company Vara is doing the same work in this field. The most exciting thing for all of us: how AI will improve our health in the long term.

Jon Gray 22:02

Given all this is happening, where do we go from here? We have three fundamental beliefs.

First, these use cases will proliferate, which will dramatically boost productivity—not just in healthcare and legal, but in the real world. As productivity increases, people will use AI more, driving exponential growth in intelligence demand. We believe this will continue. The real bottleneck is in the physical world, in chips, electricity, and data centers.

Jon Gray 22:45

Why these challenges? I'm standing right now at the site of a data center, an AI factory, facing a series of real-world challenges.

First is the entitlement issue. Communities have concerns. Unfortunately, there's a lot of misinformation—building a data center uses almost no water, you can bring benefits to the community, and use new energy to lower electricity costs. But concerns still slow projects down, like pauses we've seen in New York.

Second is the difficulty of power construction, like waiting until 2031 to get a turbine from GE Vernova.

On chip supply, hyperscalers have increased their capacity ceilings nine-fold in five years, while chip companies' capacity hasn't even doubled. Chip companies are in a cyclical industry; they've hesitated, leading to severe chip shortages.

Another issue truly relevant to us is that these AI factories, encompassing power, data centers, and the chips themselves, require enormous capital—$55 billion per gigawatt of power. Blackstone Group LP has the unique scale to understand the situation and mobilize the necessary resources.

Now, let's look back at what kept us up most at night last year: underestimating the risk of technologically disruptive innovation.

Last year, this showed up in the stock market—shares of professional services, software, and information service companies fell. In many cases, their underlying fundamentals were sound, but uncertainty increased because market participants worried about these companies' future prospects, compressing valuation multiples. This also happened in private markets, with software deals down 66%, and larger deals falling even further.

But not all companies will face the same outcome. The six retailers facing Amazon's impact 25 years ago is a perfect example: Kmart, Sears, and Toys R Us were eventually left behind, while Walmart, Costco, and TJ Maxx survived. They had unique value propositions and successfully weathered the storm.

We'll see similar dynamics in white-collar businesses. Some companies have critical systems of record, and their management teams shifted focus from seats to outcomes, ultimately achieving better performance. During this period, some companies' multiples may be lower due to uncertainty.

Jon Gray 26:23

There are also winners in this world unrelated to tech—let's call them scarce assets. We invested in a cricket team in India—with 1.4 billion people, cricket is their favorite sport, and there are only 10 teams. AI can't replace cricket. We also have a great coffee chain, 7 Brew, famous for its Blonde Lady coffee. People still need and will love these products.

Jon Gray 26:55

Beachfront property is probably one of the best investment areas. No one loves real estate, and multiples are low, but these experiences won't disappear, and these assets are irreplaceable. Or infrastructure like Rome's airport, which is equally crucial.

So, what are the biggest risks?

First is cybersecurity. If critical infrastructure in financial institutions has a problem, it could easily trigger political instability, not just in the U.S. but globally.

Second, if regulation slows the technological frontier, it won't change the inevitable diffusion of technology, because today's technology is already so powerful. Continued technological evolution is a very important risk. Think space-based data centers, or edge computing on your phone. These things could happen. But given the demand curve, it's a bit like electricity—you need nuclear, natural gas, and renewables working in parallel. Ground-based data centers remain very important right now.

Third is geopolitical tension. Fourth is the issue of inflated valuations. Those just-starting companies with no revenue are valued at $10 billion, and some defense tech companies have astronomically high valuations. These need close attention.

This is what we're discussing every day when we make investments, and it's one reason we're focused on the bottom of the tech stack—compute—because that's where we have the most conviction.

So, what's the real "core"? It's chips, data centers, and electricity. And above that, the real core is the return on investment. We will continue to see increasing use cases and sustained demand growth, and that justifies this massive investment.

Finally, let's close with our "North Star" slide—we always say "it's all about returns." Since Pete and Steve founded this firm 40 years ago, we've been committed to creating value for our clients. Steve reminds us of this every day. That's why we write memos, why we read them on weekends, and why we spend so much time thinking about where the world is heading.

Gratitude is an enduring value. We are deeply grateful to all of you. I hope you enjoy the rest of your day. Thank you again.