The AI Capex Gamble, Oracle’s Wake-Up Call, and a Media Merger That Changes Everything

The most important earnings report of the week came from Oracle, and it sits right at the center of the AI debate. Oracle increased its annual capital expenditure budget from 35 billion dollars to 50 billion dollars. The market reacted swiftly and harshly. The stock dropped 11 percent after hours.

This reaction matters far beyond Oracle. By most estimates, US GDP growth in 2025 will be around 2 percent, and virtually all of that growth is coming from hyperscaler AI capital spending. Strip out AI-related capex and GDP growth is close to zero. That makes AI spending not just a sector story, but a macroeconomic one.

The stakes could not be higher. If the current approach of scaling ever larger models starts to lose effectiveness, hyperscalers may slow their spending. If that happens, the virtuous cycle that has driven chip demand, cloud growth, and equity valuations could reverse.

Oracle’s Results Expose the Financing Problem

Oracle’s earnings were mixed, but the details are what matter.

The company beat on earnings per share but missed on revenue. Its backlog rose again, reaching 523 billion dollars, up from 450 billion the prior quarter. That backlog growth was once enough to excite investors. This time, it was no longer a surprise.

What shocked the market was the capex increase. Unlike Microsoft or Google, Oracle cannot fund this level of spending entirely from internal cash flow. Its balance sheet has already absorbed a significant increase in debt. Investors are now asking a simple question: how does Oracle finance a 500 billion dollar plus backlog without meaningfully stressing its capital structure?

That concern explains why the stock first surged earlier this year on AI optimism and then corrected sharply as investors focused on financial reality.

Housing Confirms the Slowdown Thesis

Housing earnings this week reinforced the view that this part of the economy remains weak.

Toll Brothers reported revenue growth, but earnings per share declined as incentives continued to compress margins. Backlog fell 15 percent year over year, which is not encouraging for future earnings.

Home Depot’s investor day told a similar story. The company guided to 2026 earnings per share growth of zero to four percent, below expectations. Housing is not providing any offset to slowing growth elsewhere.

Netflix, Warner Brothers, and a Corporate Governance Failure

Fifteen years ago, a Time Warner CEO mocked Netflix as an irrelevant threat. Last week, Netflix announced the purchase of Warner Brothers at an enterprise value of 83 billion dollars.

Paramount responded with a hostile all-cash offer at a higher price. Paramount wants the entire company. Netflix is interested mainly in the studio and streaming assets and would spin or sell the rest. It is now unclear who will ultimately win.

Before discussing strategy, it is worth addressing a gross injustice. David Zaslav engineered the Discovery Warner merger in 2022. At the time, the stock traded around 24 dollars. Earlier this year it traded below 8 dollars. Fundamentals deteriorated sharply. Yet Zaslav earned roughly 150 million dollars in compensation during his tenure and stands to make another 550 million dollars from this sale.

This is a failure of corporate governance. Destroying shareholder value while receiving extraordinary compensation is a problem that extends well beyond this single case.

Can the Netflix Deal Even Happen?

If Netflix wins, regulatory risk is substantial. Netflix controls roughly 20 percent of the streaming market and HBO about 15 percent. Combined, that is 35 percent.

Opponents will argue that this level of concentration ends the streaming wars. Netflix will argue the market includes YouTube, TikTok, and X. That argument will be tested. Politics will matter as well, and the current administration appears sympathetic to certain bidders.

Even if approved, this is a high-risk transaction. Large mergers often fail or underperform. The successful ones are executed by companies with deep experience integrating acquisitions. Netflix does not have that history at this scale. Warner Brothers has a fragmented culture shaped by repeated ownership changes. Execution risk is real.

If I had to bet, I would still say the odds favor Paramount, especially now that a politically connected consortium is backing its bid.

The Bigger Question: Is AI Scaling Hitting a Wall?

This brings us back to the most important issue of all. Is AI spending justified?

Nvidia’s explosive revenue growth is simply the visible outcome of hyperscaler capex. The real question is whether that capex will generate adequate returns.

Ilia Sutskever, one of the founders of OpenAI, recently raised serious doubts. He argued that current large language models will not achieve artificial general intelligence and that the benefits of scaling larger models are diminishing. In his words, we may be leaving the age of scaling and returning to the age of research.

Other researchers have echoed this view, arguing that new methods are required to achieve the next leap forward.

If this view proves correct, hyperscalers may slow chip purchases. If that happens, the entire AI-driven growth cycle could unwind. Some estimates suggest big tech needs two trillion dollars in AI revenue by 2030 to justify current investment levels. That is an extraordinary hurdle.

Productivity Gains Are Real but Are They Enough?

A software engineer wrote to me describing how AI has improved productivity by roughly 5 to 30 percent, mainly through faster triage, debugging, and scaffolding code. These gains are meaningful. AI is clearly here to stay.

But productivity improvements of this magnitude, while valuable, do not obviously justify the scale of global AI infrastructure spending we are seeing today. Much more value creation will be required.

Final Thoughts

This is not my area of technical expertise, but as a long-term investor, I cannot ignore the risks. AI is already delivering real benefits. At the same time, the financial and macroeconomic stakes are enormous.

If AI continues to scale effectively, the current investment boom may be justified. If not, the consequences will extend far beyond tech stocks, touching GDP growth, employment, and market stability.

Thanks for reading this week’s wrap.
If you’d like to catch my interviews and market breakdowns, visit The Real Eisman Playbook or subscribe to the Weekly Wrap channel on YouTube.

This post is for informational purposes only and does not constitute investment advice. Please consult a licensed financial adviser before making investment decisions.