The question isn't whether AI infrastructure spending is massive — it's whether anyone's going to make money fast enough to pay for it all.
The Summary
- Sandra Rivera from VSORA and Apollo Global's Jim Zelter both call AI capex spending "unprecedented," but Rivera asks the harder question: can revenue grow fast enough to justify the infrastructure build?
- Apollo is betting yes, partnering with Nvidia and the New York Yankees while positioning for a "higher for a while" rate environment that makes capital more expensive
- The semiconductor supply chain and physical infrastructure are now the bottleneck, not the vision
The Signal
Sandra Rivera, VSORA Chair, laid out the central tension in AI infrastructure: demand and revenue need to scale fast enough to justify the tens of billions being poured into data centers, chips, and power infrastructure. It's a timing problem disguised as a technology problem. The money is flowing. The question is whether the business models arrive before the patience runs out.
Jim Zelter, President of Apollo Global Management, sees the scale as "unprecedented" but doesn't flinch at the price tag. Apollo is deep in the game, partnering with Nvidia on infrastructure financing and even the New York Yankees on what appears to be stadium and real estate plays tied to the broader AI capital wave. When private equity giants start bundling sports franchises with semiconductor companies, you're watching capital find creative ways to ride the same wave.
"Rates are going to be higher for a while" — which means the cost of all this infrastructure spending just went up.
Here's what matters for Web4 builders:
- The infrastructure layer is now a race against cost of capital. Higher rates mean projects need to prove ROI faster.
- Partnerships like Apollo-Nvidia signal that AI infrastructure is becoming an asset class, not just a tech play. Financial engineering meets silicon engineering.
- The semiconductor outlook isn't just about chips. It's about power grids, cooling systems, data center real estate, and the entire physical layer that agents need to run at scale.
The Implication
If you're building on AI infrastructure, you're now competing with the cost of capital itself. The window for "build it and they'll come" is closing. Revenue models need to be clear, fast, and defensible. The good news: institutional money is pouring in. The bad news: institutional money has expectations and timelines.
Watch where the financing deals go next. If Apollo is partnering with Nvidia and sports franchises, other private equity shops are mapping similar plays. The companies that control the physical layer of Web4 are about to get very well capitalized and very impatient.