Oracle’s AI bet exposes the cost of chasing compute

Oracle’s Cloud Growth; Debate Around AI Risks

Oracle’s growth shows how AI infrastructure demand can boost revenue while intensifying debt, cash-flow, valuation, and community risks.

3 key takeaways

1Oracle’s AI infrastructure push is producing strong results but demands heavy spending, leverage, and continued OpenAI-related demand.
2Warnings about AI extend beyond existential risk to Chinese misuse, weapons research, data-center opposition, and weak coordination.
3Microsoft’s planned capacity expansion underscores both the scale of AI demand and the financial exposure created by the compute race.

Don't miss

Lior Sousa frames AI safety coordination through familiar automotive standards, linking technical safeguards to the practical politics of building data centers.

The brief

Oracle’s strong results mask a costly transition into AI infrastructure, where capital spending, leverage, negative free cash flow, and dependence on OpenAI demand raise the stakes.

Analyst Stefan Slavinsky lowered Oracle’s price target despite its performance, pointing to software multiple compression, investor expectations, capital needs, and leverage before turning to Adobe’s slowing growth.

The AI risk debate moves from existential warnings to practical threats: Chinese model misuse, alleged weapons and biological applications, frontier-lab coordination, and opposition to new data centers.

Eclipse CEO Lior Sousa compares AI safety standards with seat belts and airbags, arguing that collaboration, industrial investment, and clearer local benefits can make the buildout more durable.

Microsoft’s plan to more than triple data-center capacity shows how compute shortages are constraining growth while intensifying competition and the risk of an AI debt bubble.

What was said on this episode

13 statements · 5 positive · 5 negative · 3 neutral

Brady Fordon Oracle CorporationPositive3:17
Oracle delivered the data-center capacity it promised in the quarter.
“overall... it was a very clean quarter. It pretty much showed that they are delivering the capacity they promised.”
Listen at 3:17
Brady Fordon Oracle CorporationNeutral3:59
Oracle is transitioning from traditional software toward AI data-center infrastructure.
“they're trying to transition from a traditional software business to an AI data center for a developer.”
Listen at 3:59
Brady Fordon Oracle backlog conversionNeutral5:15
Oracle’s key challenge is converting backlog into revenue quickly.
“the big question is how quick can you turn that RPO into revenue”
Listen at 5:15
Stefan Slavinskyon Oracle CorporationPositive6:15
Oracle earnings will grow 25–30% annually over the next couple of years.
“This is a company that's going to be growing earnings at 25%, 30% over the coming couple of years.”
Listen at 6:15
Stefan Slavinskyon Oracle equityPositive6:40
Oracle equity should perform if growth, margins, and cash flows materialize.
“then the equity should perform.”
Listen at 6:40
Stefan Slavinskyon Oracle free cash flowPositive9:01
Oracle should return toward positive free cash flow within a few years.
“For Oracle, there's light at the end of the tunnel. We should be inflecting back towards positive free cash flow in the next couple of years.”
Listen at 9:01
Stefan Slavinskyon Oracle customer prepaymentsPositive10:08
Customer prepayments reduce Oracle’s need for external financing.
“Oracle needs to spend less of their own capital. need to raise less capital from the market in order to finance that.”
Listen at 10:08
Stefan Slavinskyon Adobe backlogNegative10:46
Adobe’s backlog declined sequentially in the third quarter.
“the RPO, the backlog, actually declined quarter over quarter, which is very rare for Q3.”
Listen at 10:46
Stefan Slavinskyon Adobe net new ARRNegative10:53
Adobe’s net new ARR fell nearly 40% year over year.
“The net new ARR was down 40% almost year over year.”
Listen at 10:53
Stefan Slavinskyon Adobe core creative appsNegative10:58
Adobe’s core creative applications slowed to approximately 8% growth.
“The core creative apps slowed to around 8% growth.”
Listen at 10:58
Stefan Slavinskyon Adobe freemium strategyNegative11:27
Adobe revenue growth will slow as it expands freemium offerings.
“revenue growth is going to slow here as they really push into freemium.”
Listen at 11:27
Brady Fordon Hyperscaler data-center supply constraintsNegative49:04
Data-center supply constraints limit hyperscalers’ cloud revenue growth.
“if it weren't for the supply constraints as data centers, they'd be able to juice their cloud revenue much more.”
Listen at 49:04
Brady Fordon Agentic AI workloadsNeutral50:11
Agentic AI workloads increasingly run on CPUs rather than GPUs.
“in the agentic era, a lot of what you think of as AI workloads is being done off CPUs”
Listen at 50:11

Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.