Platformonomics TGIF #93: July 11, 2025

By

Logo featuring abstract mountain shapes in blue and gray.

Platformonomics TGIF is a weekly roll-up of links, comments on those links, and perhaps a little too much tugging on my favorite threads.

Optimism abounds about getting a (disproportionate) piece of the AI pie!

News

CAPEX Clues: The Music is Still Playing, For Now…

Text in a serif font stating 'Wall Street Proposes ‘Off Ramp’ for Data Center Investors'.

I’ve been skeptical about all the financial entities piling into data centers. There just aren’t that many customers beyond the hyperCAPEX companies, who prefer to own their data centers. But fear not, Wall Street’s brightest minds are hard at work looking for “other investors” (aka “greater fools“, specifically “insurance and retirement funds”) to take these investments off their hands before the music stops playing:

Goldman Sachs has floated one alternative: that they sell off their stakes to other investors once long-term tenants are locked in and buildings are “stabilized.”

There’s a ton of capital invested in this and people are quite excited about the returns. But we need to build the off ramp,” he said. “That’s where the smartest people in the industry are focused right now.”

There’s an important hitch, of course. To attract insurance and retirement funds, the industry would need to introduce more certainty and less risk into data center lease contracts. That includes nixing early termination clauses for tenants and extending leases to 17 to 20 years, from 10 to 15 years, according to the white paper. Whether those stricter terms become market standards remains to be seen.

Meanwhile, the New York Times had a nice piece on Amazon’s build-out in Indiana that gives a sense of what the financial entities (and their customers) are up against in terms of scale, vertical integration, and know-how. And, by NYT standards, the story is light on the pious degrowth ideology they peddle (“How dare anyone use energy! We should return to the golden age characterized by printing presses and delivery trucks!”).

Aerial view of Amazon's largest data center in Indiana, showcasing expansive infrastructure and numerous cooling units, with a caption highlighting its role in supporting artificial intelligence development.

Words are Cheap, CAPEX is Expensive: Oracle Edition

Text graphic announcing 'Oracle Strikes Big Cloud Deal That Could Double Cloud Revenue by 2028'.

Database vampire and incessant stock promoter Oracle wants you to know they’ve signed their deal with OpenAI, and it could bring in $30 billion in revenue in FY28. They said nothing about how much infrastructure they have to deploy to book that revenue. Given Oracle’s struggles to deploy CAPEX even at modest scale, that forecast seems optimistic. This warrants a post of its own about how much CAPEX this deal might require, how likely are they to find 4.5 gigawatts in the next two years, and the financial consequences for a company that already has negative free cash flow.

Big Bad Bankruptcy Bill Blame

Headline text stating that Trump's Big Beautiful Bill will accelerate an American energy crisis and could cost the U.S. the AI race.
Tweet discussing the potential political ramifications for data centers if a specific bill passes.

Data centers already get cast as the boogeyman for their electricity consumption (somehow EVs and heat pumps escape such scrutiny). Big Tech’s data centers are an easy target for degrowthers and Luddites, but if Trump’s OBBB materially impairs the growth of power generation as many expect, the political imperative for Trumpistas to divert blame will create yet another group of data center critics.

We should expect the electrical utilities to borrow the telco playbook and try to get Big Tech to pay twice, because “they have the money”. Why should utilities be held responsible for their failure to keep up with energy demand?

As a periodic reminder, energy-scarcity is incompatible with living in an advanced civilization. Yet here we are.

Never Take a Dependency on Elon Musk: xAI MechaHitler Edition

Text from an NBC News article discussing antisemitic remarks made by Grok, an AI associated with Elon Musk.

Thin-skinned fractional CEO (and former public sector volunteer) Elon Musk knows that frontier models are a scale game and he needs to maximize enterprise, ISV, and API consumption of Grok. There are a lot of ways to attack the enterprise market. Elon’s strategy seems to be making the key Nazi vertical his enterprise beachhead.

Is This Founder Mode? Meta AI Edition

Text headline stating 'Meta AI Researcher Warns of ‘Metastatic Cancer’ Afflicting Company Culture' displayed in a bold serif font.

“I have yet to meet someone in Meta-GenAI that truly enjoys being there. Someone that feels like they want to stay in Meta for a long time because it’s such a great place,” wrote Blankevoort, referring to the nearly 2,000-person group that develops Meta’s flagship AI model, Llama. “You’ll be hard pressed to find someone that really believes in our AI mission. To most, it’s not even clear what our mission is.”

Software Migration Alerts: Olo, itel

Header text stating 'Thoma Bravo Agrees $2 Billion Deal to Take Olo Private' in bold font.
Headline about Thoma Bravo's Nearmap acquiring insurance tech firm itel for over $1.3 billion.

When Thoma Bravo shows up, that software you rely on is washed up.

Market research: would you considering investing in a Platformonomics ETF that invests in software companies with private equity-owned competitors?

Go back

Your message has been sent

Choose one option
Warning

One response

  1. I promoted Gemini in “research” mode to look for glarying obvious signs in the market overlooked similar to what experts missed in 2008. Forgive this massive block of text, but curious what you think:

    AI’s Social Integration: Reshaping Behavior and Economic Expectations
    A. Consumer AI Adoption: The Utilitarian Shift and the Monetization Gap
    Artificial Intelligence has reached a consumer “tipping point” in 2025. More than half of American adults (61%) have used AI in the past six months, with nearly one in five relying on it daily. Globally, this translates to 1.7-1.8 billion users, with 500-600 million engaging daily. This is not mere experimentation; it’s habit formation at an unprecedented scale. Usage patterns reveal a utilitarian approach: consumers primarily adopt AI because it makes life “easier, faster, cheaper” for routine tasks. Adoption is highest among students (85%), employed adults (75%), higher-income households (74% for $100k+), and surprisingly, parents (79%), who use it for childcare, research, and note-taking. The “default tool dynamic” is prevalent, with consumers typically trying general AI assistants first and only seeking specialized alternatives if the default falls short. This widespread adoption coupled with a massive monetization gap suggests that consumers perceive AI primarily as a “free utility” or an embedded feature rather than a standalone, paid product. This “default tool dynamic” reinforces this perception.
    Despite this broad adoption, a significant “monetization gap” exists. The current consumer AI market is only $12 billion annually. However, if 1.8 billion users paid an average of $20 per month (a common subscription price), the market would be $432 billion per year. This indicates that only about 3% of users pay for premium services , a strikingly low conversion rate (e.g., ChatGPT converts only ~5% of weekly active users).

    So I see two ways to take this:
    1. Monetize immediately or short term death for many capex spenders
    2. The whole thing is just insanely overvalued in general

    Curious to your thoughts as always. Great and entertaining read as usual!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Get Updates By Email

Discover more from Platformonomics

Subscribe now to keep reading and get access to the full archive.

Continue reading