Success brings great joy, while failure may lead to total loss.
Holding a massive amount of AI infrastructure-related orders is no longer enough to "protect" a company.
Oracle holds $500 billion in orders, yet its stock price has dropped 40% since its peak in September. Broadcom currently has about $73 billion in AI product orders backlog, and after the latest earnings report, its stock price turned from rising to falling.
CoreWave, dubbed "NVIDIA's favorite child," has quarterly revenues in the billions but managed to secure over $36 billion in orders from OpenAI and Meta within a week. Over the past month, the company's stock price has cumulatively dropped 17%.
While there are concerns about whether they have enough capability (money) to meet customer demands, there are also worries about whether the customers themselves are truly "reliable."
At the core of AI infrastructure are a few major players: Meta, Google's parent company Alphabet, Microsoft, Amazon, Apple, NVIDIA, along with star AI startups like OpenAI and Anthropic.
Star startups are still quite immature and rely heavily on external financing for infrastructure, which is a clear risk.
The giants should be like a stabilizing force—they are financially sound, with ample cash, and are filling the coming years with hundreds of billions in crazy infrastructure plans.
However, the returns from AI spending remain minimal, and whether nurturing new dreams with "old capital" will drag down the giants depends entirely on the timeliness of those dreams being realized.
Success brings great joy, while failure may lead to total loss.
01
Holding the "future" card, Oracle has experienced great joy and great sorrow in just a few months.
When joy arrived, Oracle's stock price soared 40% in a single day, and founder and CEO Larry Ellison briefly surpassed Musk to become the world's richest person.
At that time, Ellison exclaimed, "Artificial intelligence is everything!"
Artificial intelligence is indeed everything, and for Oracle, it was the reason for this wave of joy—OpenAI had reached a five-year, $300 billion computing power procurement agreement with Oracle, igniting Oracle's stock price.
However, just three months later, despite holding more orders, the "magic" had vanished.
Oracle's latest earnings report for the second quarter of fiscal year 2026 (covering September to November 2025) showed a 14% year-over-year revenue increase, with the company stating that its backlog of orders had reached $523 billion.
This figure increased by $68 billion compared to the previous fiscal quarter.
Upon the release of the earnings report, the stock price fell 11% on the same day, marking the largest single-day drop for the company since January. From the peak in September, Oracle's stock price has already dropped 40%.
Future orders, amid skepticism about the "AI bubble," have shifted from a beautiful hope to a heavy burden.
Oracle appears to be struggling—its earnings report shows a negative cash flow of $10 billion, with quarterly capital expenditures (CapEx) reaching $12 billion, nearly $3.7 billion higher than analysts' predictions.
Oracle's CFO revealed that the company's fiscal year spending has also been raised by $15 billion, reaching $50 billion.
The biggest fear in the market is: does Oracle have the capability to raise enough money to support such large-scale AI infrastructure?
Some analysts predict that Oracle will need to borrow $100 billion to complete its construction. In the second quarter, the company raised $18 billion in debt, marking one of the largest bond issuances on record for a tech company.
During the conference call, Oracle vigorously defended itself, clearly opposing the prediction of needing to borrow $100 billion, stating that the actual financing amount would be significantly less. The key lies in Oracle's approach of "customer-provided chips."
In other words, instead of Oracle buying chips and renting them to customers, customers bring their own chips, which is unprecedented in the cloud service industry.
Additionally, Oracle emphasized that some suppliers are willing to rent rather than sell chips to them, allowing Oracle to synchronize payments and receipts.
If Oracle's claims are true, it could significantly reduce its upfront investment and greatly increase its return rate.
However, for the market, the risk has not disappeared but has shifted: from Oracle to Oracle's customers. Customers like Meta or OpenAI are purchasing expensive GPUs and installing them in Oracle's data centers.
Whether Oracle's future of hundreds of billions can be realized certainly depends on its ability to "deliver," but it also depends on whether customers can "pay." About two-thirds of Oracle's nearly $500 billion in undelivered orders come from unprofitable OpenAI, with another known $20 billion from new agreements with Meta.
Similarly, Broadcom also holds a large number of orders but has received negative feedback from the market.
Broadcom also released a new earnings report, showing that for the fourth quarter of fiscal year 2025 ending November 2, it exceeded expectations in both core revenue and profit, with AI semiconductor-related revenue growing 74% year-over-year.
During the conference call, Broadcom CEO Hock Tan stated that the company currently has about $73 billion in AI product orders backlog, which will be delivered over the next six quarters. He emphasized that this is the "minimum value," and the backlog is expected to expand further as new orders continue to pour in.
However, Broadcom refused to provide clear guidance on AI revenue for the entire year of 2026, citing uncertainty in customer deployment rhythms, which may lead to fluctuations between quarters.
After the earnings report was released, Broadcom's stock price initially rose about 3%, but then turned to decline, with an after-hours drop of over 4%.
Compared to Oracle's dramatic ups and downs, Broadcom's situation can only be seen as a minor bump, but the underlying market sentiment is similar—people are no longer optimistic about the "future" of the booming AI infrastructure.
Broadcom's customers are also relatively concentrated, with its AI-related orders mainly coming from OpenAI, Anthropic, Google's parent company Alphabet, and Meta.
02
The onion of AI infrastructure, when peeled to the core, reveals a few familiar companies—the "Seven Sisters" of the U.S. stock market and OpenAI, Anthropic.
Also gaining attention this year is the AI cloud infrastructure startup CoreWave, which went public in March and is the largest tech startup IPO since 2021. Its stock price subsequently more than doubled, even surpassing the "seven tech giants."
Its customer concentration is also extremely high, primarily relying on orders from Microsoft, OpenAI, NVIDIA, and Meta.
Just this past Monday (December 9), CoreWave issued $2 billion in convertible bonds, while its total debt had already reached $14 billion as of the end of September. Market concerns have intensified, and its stock price has dropped 17% over the past month.
The point remains: the market has developed deep-seated doubts about the AI industry as a whole, not only regarding whether these AI infrastructure-related companies can provide services as planned but also whether the big clients making crazy trades can truly fulfill their bills.
The complex cycle of transactions among all parties has formed a tight and opaque web, making everything even less clear.
If we look at the types of customers, startups like OpenAI and Anthropic have sparked early concerns.
The reason is simple: neither has a stable revenue-generating capability, at least not enough for their expanding infrastructure plans; they need to rely on external financing, and the uncertainty is evident.
The giants, on the other hand, act more like a barometer and safety net in the game.
The giants spend hundreds of billions annually on capital expenditures, a significant portion of which is used to expand data centers. Their combined capital expenditures in 2026 will exceed four times the total spending of the U.S. listed energy sector on drilling exploratory wells, extracting oil and gas, transporting gasoline to gas stations, and operating large chemical plants. Just Amazon alone has capital expenditures that exceed the total of the entire U.S. energy sector.
Compared to immature startups, the giants are clearly financially robust; they are financially sound and have ample cash flow. At least for now, their spending has not exceeded their capacity.
For instance, Microsoft, Google, and Amazon together will spend over $600 billion from 2023 to this year, with expected revenues of $750 billion.
If we look at their recent earnings reports, we find that their performances are quite strong, with "exceeding expectations" becoming a norm, suggesting that there is no need for concern—in other words, they can afford to invest heavily in AI infrastructure.
However, upon closer inspection, none of them have fundamentally changed their revenue structure; while AI has begun to generate returns, its contribution to overall revenue often remains a supporting role, even as it occupies the spotlight in spending.
For example, Microsoft, in a report from TheCUBE Research at the end of July regarding its quarterly earnings, estimated that AI services contributed about 19% to Azure cloud growth, exceeding $3 billion, but this accounted for less than one-tenth of Microsoft's total revenue.
More than half of Google's revenue still comes from advertising and search, while Amazon's e-commerce and advertising still account for over 70% of its revenue.
In other words, the giants are using mature businesses to nurture the future of AI.
The question is, how long can they sustain this?
03
The giants have already begun to spark a "debt frenzy."
In September, Meta issued $30 billion in bonds. Alphabet recently announced plans to issue about $17.5 billion in bonds in the U.S. market and approximately $3.5 billion in Europe.
Data from Bank of America shows that in just September and October, large tech companies focused on artificial intelligence issued $75 billion in U.S. investment-grade bonds, more than double the average annual issuance of $32 billion for the industry from 2015 to 2024.
These companies' revenue growth should currently support their spending, but to keep pace with the rapid developments in artificial intelligence, they will ultimately need more debt.
The Wall Street Journal sharply pointed out in an analysis: AI is making the giants weaker.
As of the end of the third quarter of this year, Microsoft's cash and short-term investments accounted for about 16% of total assets, down from about 43% in 2020. Alphabet and Amazon's cash reserves have also significantly decreased.
Alphabet and Amazon's free cash flow is expected to be lower than last year. Although Microsoft's free cash flow has shown some growth compared to the same period last year over the past four quarters, its disclosed capital expenditures do not include long-term lease expenses for data centers and computing equipment. If these expenditures are included, its free cash flow would also decline.
This trend seems destined to continue.
Analysts estimate that if Microsoft includes lease expenses next year, it is expected to spend about $159 billion; Amazon is expected to spend about $145 billion; and Alphabet is expected to invest $112 billion. If these predictions hold true, these companies will cumulatively invest $1 trillion over four years, most of which will be directed towards artificial intelligence.
Overall, these changes—decreasing cash balances, declining cash flow, and increasing debt—are fundamentally altering the business models of tech companies.
The tech industry is increasingly resembling sectors like semiconductor manufacturing, where hundreds of billions are invested in the construction of cutting-edge factories that take years to build, but the returns take even longer.
Deploying hundreds of billions in thousands of massive data centers presents clear and significant challenges from an execution standpoint in AI infrastructure.
Data centers consume enormous amounts of electricity—GPUs require substantial power for computation—and the current power grid cannot meet the surging demand. Additionally, cooling is also an issue. GPUs operate at high temperatures and require large amounts of freshwater to keep the equipment running. Some communities have already begun to oppose the construction of data centers, fearing it will impact their water supply.
This year, NVIDIA announced a new agreement with OpenAI worth up to $100 billion, with OpenAI planning to deploy 10 gigawatts of NVIDIA systems. However, NVIDIA's CFO recently admitted that this plan is still in the letter of intent stage and has not yet been formally signed.
On one hand, this casts a shadow over the "credibility" of the bustling AI infrastructure deals, and on the other hand, it hints at future uncertainties.
The reasons for the delay in signing the agreement have not been disclosed, but the "risk factors" section in the documents submitted by NVIDIA to the SEC can serve as a reference.
In the document, NVIDIA warns that if customers reduce demand, delay financing, or change direction, the company may face risks such as "inventory surplus," "penalties for order cancellations," or "write-downs and impairments of inventory."
Furthermore, the availability of "data center capacity, electricity, and capital" is crucial for the deployment of AI systems. The document states that building electrical infrastructure is a "process that will take years" and will face "regulatory, technical, and construction challenges."
Even if AI infrastructure progresses smoothly in the end, it is not the "end point" of success.
AI infrastructure ultimately serves AI demand; if the infrastructure is in place but market demand fails to materialize, then insufficient utilization of the infrastructure will lead to significant losses.
Of course, not everyone is frowning with concern; supporters believe this is a gamble worth taking, as AI demand will grow exponentially rather than linearly.
Analyst Azeem Azhar calculated that direct revenue from AI services has increased nearly ninefold in the past two years.
In other words, if this growth rate continues, it is only a matter of time before AI companies start generating record profits.
"I think those who are fixated on the specific financing methods of these investments are thinking outdated. Everyone assumes this technology will develop at a linear pace. But AI is an exponentially growing technology. It’s a completely different model," Azhar said.
But the question remains: will the moment when AI begins to explosively generate "profits" arrive, and when will it come?
Ultimately, whether AI infrastructure will drag down the giants is a race between AI market demand and AI infrastructure. If it keeps pace, AI infrastructure is "worth it"; if it falls behind, the massive data centers will end up like "ghost towns." That would be the best proof of the giants' incorrect bets on AI and would lead to catastrophic consequences.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
