Oracle booked more than $30B of AI cloud contracts in one quarter.
It also had free cash flow of -$5.4B.
On the same day, the FT said the 30-year US bond yield hit 5.35 per cent, its highest level since 2007.
AI demand is real.
AI spending is huge.
Rates are not making that spending easier.
Oracle Shows The AI Cloud Trade
Oracle’s Q1 was strong on demand and heavy on spending.
Revenue was $19.3B, above the $19.14B estimate.
Adjusted EPS was $1.92, above the $1.74 estimate.
RPO was $664B, up $209B year over year.
Cloud Infrastructure was $7.4B, up 121% year over year.
Cloud Applications was $4.2B, up 10% year over year.
Non-GAAP operating margin was 42%.
Then comes the spending.
CapEx was $28.5B.
Free cash flow was -$5.4B.
Oracle also booked more than $30B of AI cloud contracts in Q1, and said there is no incremental impact on its plans to raise capital.
Shares climbed 7% in after-market trading.
So demand is not the doubt.
The question is whether the cash going out today becomes enough profitable cloud revenue later.
The FT said Oracle previously told investors it would invest $70bn in the coming fiscal year to finance data center construction, up from $55.7bn in the year ended May 31.
Oracle also completed a $20bn share sale as part of a $50bn package of measures to help finance data centers.
S&P downgraded Oracle’s credit in July to one notch above junk status, citing reliance on a few customers and an uncertain path to profitability in its AI data center business.
That is the cleanest version of today’s AI question.
The demand is visible.
The spending is visible too.
Oil Is Making Inflation Harder To Kill
August PPI rose 0.4% month over month and 5.4% year over year.
The source tied that move to energy costs hitting businesses, especially oil and diesel.
CPI is due tomorrow.
The expected numbers are:
- 3.4% headline CPI
- 2.4% core inflation
A hotter CPI would strengthen the case for a Fed hike next week. A cooler CPI could calm markets after the PPI reaction.
Traders now put roughly 70% odds on a Fed rate hike next week, up from the mid-60s before PPI.
Oil surged above $100, with Brent briefly above $105 to $107 in some reports, as the U.S.-Iran conflict disrupted shipping through the Strait of Hormuz and Red Sea.
Higher oil makes inflation harder to kill. Higher inflation makes rate cuts harder.
George Gammon warned that crack spreads have “absolutely skyrocketed,” refined product exports such as gas and diesel are down over 6 million barrels a day, and gas prices in the United States will go up by at least 50% if things stay the same.
That is a serious risk if energy keeps feeding into inflation.
Bonds Are Pushing Back
The FT said the 30-year US bond yield jumped as much as 0.06 percentage points to 5.35 per cent, its highest level since 2007.
The 10-year yield is approaching 5 per cent.
Europe is tightening too.
The European Central Bank raised rates by 25 basis points, its second hike this year.
The market is pricing in a 90% chance of a third hike this year.
Europe joins Korea and Japan, which have also hiked this year.
Stanley Druckenmiller told a private Wall Street audience that US borrowing costs are still “a little low,” rate cuts “are no longer needed,” and Fed committee members who say fed funds rates are restrictive are “just ridiculous.”
His point was simple: asset prices, the capex boom, and the war for capital do not look like a world that needs lower rates.
The FT also said the gap between benchmark borrowing costs in the US and China has reached its widest level ever, threatening to accelerate a shift in capital flows between the world’s two biggest economies.
Stocks did not like the setup:
- S&P 500 fell 0.58%
- Nasdaq fell 0.65%
- Dow fell 0.60%
The Buildout Keeps Getting Bigger
Jensen Huang said the AI infrastructure buildout is “not tens of billions, but tens of trillions.”
He also said neoclouds are becoming critical because “they secure land, power, and shell” after hyperscalers have exhausted much of that capacity.
The buildout is bigger than chips.
Microsoft plans to 3x data center capacity, according to Kalshi.
Google plans to invest €13B in Finland AI infrastructure over the next 2 years, according to Bloomberg.
Nebius said AI infrastructure demand is running well ahead of supply, visibility stretches to 24+ months, and customers are asking for 2028 capacity.
Amazon is selling its first-ever sterling bonds in a four-part UK deal after selling more than $92B in bonds this year.
The bottlenecks are land, power, shells, financing, and time.
Where AI Is Being Wired In
Palantir named Nebius its preferred sovereign AI infrastructure partner.
The plan is to integrate Nebius compute and inference endpoints directly inside Palantir’s enterprise perimeter, so eligible commercial customers can run open AI models, adapt them with proprietary company data, and keep control over compute, models, and data.
Palantir and Nvidia also announced a partnership to “bring sovereign AI to critical supply chains, starting with NVIDIA’s own operations.”
That stack combines Nvidia’s open Nemotron models with Palantir’s Foundry, AIP, and Ontology.
The goal is real-time visibility into constraints and faster materials allocation decisions across millions of parts and thousands of suppliers.
Jensen Huang also said:
“Cybersecurity is likely the next major growth area for artificial intelligence.”
One source listed CrowdStrike, Palo Alto Networks, Fortinet, Cloudflare, SentinelOne, Zscaler, Rubrik, and Okta.
The logic attributed to Jensen was simple: AI is already good at writing code, and cybersecurity is continuous code inspection.
Find weaknesses.
Attack them through red teaming.
Patch them through blue teaming.
Apple’s new A20 Pro in the iPhone 18 Pro has 50% more memory bandwidth.
The source said Apple is using Micron LPDDR5X mobile DRAM with 12GB of RAM.
Apple confirmed today that A20 Pro has its widest memory interface ever in an iPhone.
The source’s interpretation was that more AI on the iPhone means more compute memory demand.
Adobe also had a beat and raise.
Q3’26 revenue was $6.8B, above the $6.69B estimate, up 13% year over year.
Adjusted EPS was $6.13, above the $6.09 estimate, up 15% year over year.
ARR was $27.5B, AI-first ARR growth was 150%+, and monthly active users were 1 billion+.
Adobe raised full-year revenue and EPS guidance.
Shares were also reported down 1.7% in another brief.
Good numbers are not always enough if expectations are higher.
AI Safety Moves Onto The Watchlist
An Anthropic researcher quit and warned that AI could pose serious risks.
The broader recap identified him as Jacob Coxon, a 27-year-old former OpenAI employee.
He said:
“We’re on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control already.”
OpenAI called for mandatory federal AI safety rules for the small number of labs building the most advanced AI systems.
The proposed rules include independent testing, stronger cybersecurity standards, and incident reporting.
AI is not just a revenue and capex story. It is becoming a regulation story too.
Meta’s Settlement Has Scale
Meta settled with a coalition of 52 attorneys general for up to $18 billion before Mark Zuckerberg testified under oath.
The speaker said Meta will book a legal charge of about 10 billion dollars in the third quarter.
About 12.2 billion is unconditional, and the other 5 billion or so is paid if TikTok and YouTube sign similar deals.
The speaker also said Meta made 18.8 billion dollars in operating income last quarter and takes in about 550 million dollars a day.
The settlement is huge in absolute dollars.
Meta’s operating income is also huge.
Positioning And Insider Signals
David Tepper’s Appaloosa LP was described as highly concentrated.
The source said:
- Top 5: about 57.7%
- Top 10: about 81.5%
- Top 20: about 98.2%
The source identified the biggest theme as “AI + SEMIS + BIG TECH + POWER/ENERGY.”
The next five listed positions were:
- $EWY at 6.55%
- $META at 5.09%
- $VST at 4.70%
- $NVDA at 4.08%
- $NRG at 3.44%
Uber also had insider-buying attention.
One source said Uber CEO Dara Khosrowshahi bought about $10M of stock at an average price of $70.96.
It said this was his first open-market purchase since May 2022.
Another source claimed the CEO bought $10M at $70, the CFO bought $1.6M at $71, the COO bought $5M at $75, Trump bought over $5M earlier this year, and Nancy Pelosi is holding calls.
Shay Boloor added the right caution:
“there are many reasons why CEOs buy their own stock”
He also wrote that CEOs “can be politically motivated or just used t change the narrative around a falling stock.”
And:
“CEOs are salesmen first.”
A claimed Trump portfolio breakdown listed:
- $DJT at 38%
- $NVDA at 8%
- $APPL at 8%
- $AVGO at 7%
- $MSFT at 7%
- $META at 6%
- $ORCL at 6%
- $GS at 5%
- $BAC at 5%
- $AMZN at 4%
Joseph Brooks wrote:
“A 38% position in DJT makes this portfolio far more concentrated than I expected.”
Four Names Off The Main AI Axis
Kopin was described by the host as a 40-year-old MIT spinout.
It makes micro displays used in US Army thermal rifle sights, F-35 pilot helmets, drone pilot goggles, and other applications.
The host said the stock had more than doubled in the last year, but was still down more than 30% from its all-time high at the time of recording, despite 51% revenue growth last quarter and guidance for GAAP profitability in Q4.
Murray said the US government wants to buy 3 million drones.
He described a drones-to-headsets ratio of about 1 to 20, which could mean 120 to 240,000 displays for Kopin moving forward.
Xiaomi came up as an Asia, smartphone, AI, and EV story.
The speaker said the stock is down 50%, Xiaomi is investing 40 billion renminbi, approximately 6 billion US dollars, and memory price increases have hit smartphone devices.
His AI view was blunt: there will be a price war, and “the return on investments will likely be terrible.”
EssilorLuxottica was described as a stable business with growing revenues, but the stock is down more than 50% over the last year.
The speaker said the P/E ratio went from 60 a year ago to 30 now.
His stated level was:
“Wake me up at teens P/E ratios.”
Salesforce came up in a SaaS value discussion.
Phil said Salesforce is “the corporate record,” and that record is “fundamental to the entire operation of the entire corporation.”
He said the price was “super super good,” while warning that companies are “almost always cheap for a reason.”
The Sceptics’ Corner
George Gammon said many people, including himself, think the market is in a “massive AI bubble” and is due for another crash.
His proposed rule is to sell when the 200-day moving average starts to go down and price is below it, and buy when price goes above the 200-day moving average and the moving average flattens out or goes back up.
Wall Street Millennial, citing Jay Ritter data, said the average first-year return for SPACs from 2012 through 2025 is -46%, the average 3-year return is -58%, and the average SPAC from the 2021 cohort declined by 73% after 3 years.
The speaker also said 148 SPACs went public in the first 8 months of 2026, the largest number since 2021.
ClearValueTax said he closed a copper position this week for a 21% gain and is investing in Gigastar again.
He said Gigastar is based in Chicago, is capitalizing on the influencer economy, has first mover advantage, is valued around $58 million, and that he likes the “riskreward.”
Can The System Handle The Load?
Solana’s co-founder Anatoli Yakavveno described Solana as an engineering response to blockchain bottlenecks.
He said Ethereum fees from CryptoKitties reached hundreds of dollars and Bitcoin fees reached 70 bucks a transaction.
He said optimizations that seemed obvious to him could make the system “a thousand times faster.”
Proof of history was compared to taking a number at a deli counter.
The speaker replied, “Pretty good analogy,” and added:
“You’re never waiting for somebody else to order your sandwich.”
That comes back to the same investing question: can the system handle the load?
In AI, demand is not the doubt.
The harder question is financing the buildout while rates go up, not down.
Oracle’s demand is not the doubt. The doubt, in S&P’s words, is the path to profitability, and Druckenmiller’s argument is that the capex boom and the war for capital are themselves reasons rates do not need to fall.