By Ishan Sinha

A question I hear over and over in venture circles right now: is AI ushering in a real transformation? Or are we repeating the telecom bubble of the 1990s?

This summer, I picked up John Malone’s new autobiography, Born to Be Wired. Through Liberty Media, Malone backed iconic consumer brands like Formula 1 ^[3], SiriusXM ^[4], and Live Nation ^[5], and also held positions in Expedia ^[6] and QVC ^[7]—proving his eye for durable, consumer-facing businesses.

But his career-defining move came much earlier. During the telecom boom of the 1990s, he scaled and sold Tele-Communications Inc. (TCI)—then one of the largest cable providers in the U.S.—to AT&T for $48 billion  ^[9], in a deal that many now see as a parallel to today’s AI buildout.

I was expecting some good media anecdotes. What I got instead was a masterclass in how to think about infrastructure, consumer pull, and capital strategy—three themes that feel eerily relevant to what I think we’re seeing in AI today. I had personal reasons for picking it up: Malone grew up in Milford, Connecticut, just 10 minutes from where I did ^[1]. He also went to Yale, where I studied as well ^[2]. But the deeper resonance came from his approach to building in uncertain environments.

Malone’s story reminds us: just because something looks like a bubble doesn’t necessarily mean it is. And more importantly, even bubbles can leave behind infrastructure that powers generational companies—if those companies are built with the right mix of vision, discipline, and creativity. Here are the most important lessons I think founders can take from Malone’s story:

Bet on Infrastructure, But Pressure-Test the Demand

John Malone was a man who saw around corners. Long before “broadband” was a buzzword, he envisioned cable as the “information superhighway” ^[10], a vast network capable of delivering far more than just television.

But the high costs of set-top boxes (up to $7,000 each) and the complexity of early interactive TV meant business models didn’t deliver immediate ROI ^[14]. Yet Malone didn’t pull back. Instead, he stuck with his vision, betting that the infrastructure he was building would eventually become indispensable. He was right. The same fiber-optic pipes once dismissed as failures became the foundation for broadband internet and made today’s platforms like YouTube and Facebook possible ^[12].

The idea that killer applications can come years later is one our team talks about often when speaking with AI infrastructure companies with AI infrastructure companies. Right now, everyone seems to be focused on compute, GPUs, and model architecture. The capital going into this space is enormous, and the obvious question is: are we investing ahead of sustainable demand, or just chasing a moment?

Malone’s mentor, Monty Shapiro, had a mantra that stuck with him: “What if not?”  ^[13] What if the consumer use case never comes? What if the ROI isn’t immediate? That lens—optimism tempered with scenario planning—is something I believe founders should adopt today. Especially those building deep tech foundations.

The cautionary tale here is Nortel. In the 1990s, Nortel’s revenue ballooned in part because it was financing its customers’ purchases. It was a short-term win that unraveled when demand didn’t materialize. Today, as Nvidia sells record volumes of chips, I think it’s fair to ask: are we seeing similar dynamics? And if so, how do we make sure we’re building something that lasts beyond the hype?

Only Real Demand Justifies AI’s Infrastructure Spend

There’s a lot of dazzling AI content out there—viral Drake songs, anime-style art, hyperreal videos—but in my opinion many of these tools still feel like tech demos. What’s the consumer hook that keeps them coming back? That’s the question that I think matters.

Malone understood this intuitively. He didn’t just invest in cable infrastructure, he backed HBO, CNN, MTV, BET ^[18]. Because he knew what actually drove growth was content people cared about. He believed distribution only mattered if you had something worth distributing.

As I read the book, I saw parallels to some of the ways our portfolio companies are playing in the AI sphere:

• Fever uses AI to match consumers with live experiences they’ll love—and does so in a way that feels invisible, seamless, and consumer-first. This is a direct application of Malone’s consumer-driven philosophy to the experiential economy.
• GlobalComix applies AI translation to make global comics accessible to more readers. It’s not about the AI—it’s about unlocking demand. This reactive approach, exemplified by their partnership with Marvel, demonstrates how AI can unlock and expand the reach of beloved content.
• Range Media Partners puts talent at the center, and sees AI as a tool to amplify creators, not replace them. That mindset is critical. Whether it’s a blockbuster film or a compelling series, the core appeal of a well-told story remains paramount, and AI can serve as an invaluable co-pilot in its creation and distribution, not a replacement for its essence.

In each of these cases, I believe the tech only matters because it solves a real problem. If you’re a founder in AI media or consumer, the bar isn’t novelty, it’s staying power. I think you should ask yourself: would a user still care about this six months from now?

Modern Consumer Startups Need Modern Financing Models

I used to structure equity derivatives at Goldman Sachs. Back then, John Malone’s name came up constantly. Not just because of what he built, but how he financed it. He made cable work in a capital-intensive environment by inventing new ways to deploy it: spin-offs ^[21], tracking stocks, tax-advantaged swaps ^[22]. EBITDA as a metric was practically weaponized ^[20].

Consumer companies today are running into their own capital constraints, but the dynamics are different than they were for John Malone. Most aren’t spending on infrastructure; they’re spending on customer acquisition. And equity financing isn’t always the right tool, especially when founders want to protect ownership.

That’s why we’re excited about the rise of UA financing and revenue-based funding.  As John Malone did with EBITDA, we’re excited to see the consumer startup industry rethinking growth investment strategies.  When the payback math works, I think there’s no reason not to at least consider financing with non-dilutive capital.

Our portfolio company Ladder is a great example. They recently raised UA financing from General Catalyst that allows them to scale without giving up equity. That’s a Malone-style move—structuring capital to match the business model.

I believe founders should be thinking this way: how do you fund growth without mortgaging the future? I think that’s what creative capital allocation is really about.

Build for the Long Game, Grounded in Demand

John Malone’s story is more than a telecom history lesson. It’s a roadmap. For AI founders, I believe the path forward isn’t just about building fast. It’s about building right. That means:

• Being patient enough to bet on infrastructure that won’t pay off right away
• Staying grounded in what consumers actually want
• Finding smart, flexible ways to finance growth

If you’re building in AI infrastructure, media, or consumer experiences and wrestling with the same “What If Not?” questions, let’s talk: ishan@p72.vc.

Footnotes

1. John Malone, Born to Be Wired: Lessons from a Lifetime Transforming Television, Wiring America for the Internet, and Growing Formula One, Discovery, SiriusXM, and the Atlanta Braves (Simon & Schuster, 2025), 17. [↩]
2. Malone, Born to Be Wired, 29. [↩]
3. Malone, Born to Be Wired, 329, 334-335. [↩]
4. Malone, Born to Be Wired, 291-292, 296-299. [↩]
5. Malone, Born to Be Wired, 280-281, 288-289. [↩]
6. Malone, Born to Be Wired, 272, 275, 283. [↩]
7. Malone, Born to Be Wired, 93, 103, 151, 275, 277. [↩]
8. Malone, Born to Be Wired, 12, 59-60, 243, 341. [↩]
9. Malone, Born to Be Wired, 241. [↩]
10. Malone, Born to Be Wired, 139. [↩]
11. Malone, Born to Be Wired, 71, 139, 147. [↩]
12. Malone, Born to Be Wired, 16, 71, 179, 386-387. [↩]
13. Malone, Born to Be Wired, 14, 43. [↩]
14. Malone, Born to Be Wired, 65, 139-140, 158. [↩]
15. Malone, Born to Be Wired, 51. [↩]
16. Malone, Born to Be Wired, 99, 152. [↩]
17. Malone, Born to Be Wired, 99, 112, 115. [↩]
18. Malone, Born to Be Wired, 59-60. [↩]
19. Malone, Born to Be Wired, 120-123. [↩]
20. Malone, Born to Be Wired, 244, 339-341. [↩]
21. Malone, Born to Be Wired, 66, 77, 79. [↩]