Massive orders shock the market! Oracle, Broadcom, and even Alphabet - C have all been "blown away" by OpenAI

OpenAI's capital expenditure tsunami is reshaping the landscape of Silicon Valley, pushing tech giants like Oracle, Broadcom, and even Google to the forefront of growth with astonishing orders.

Overnight, Oracle's stock price surged more than 27% in after-hours trading, marking its best single-day performance since the dot-com bubble. Although quarterly revenue fell short of expectations, a massive cloud infrastructure contract from OpenAI caused its "remaining performance obligations" (RPO) to skyrocket to $455 billion, igniting the market's fervent expectations for AI-driven growth.

Recently, Broadcom disclosed that a "mysterious client" signed a custom AI chip order worth tens of billions of dollars, prompting a "significant" upward revision of its future AI revenue outlook. According to reports from the Financial Times and other media, this mysterious buyer is none other than OpenAI. Meanwhile, Google Cloud has also reached a computing power cooperation agreement with OpenAI, breaking Microsoft's previous exclusive supply arrangement.

From cloud computing to custom chips, OpenAI is pulling the entire tech supply chain into a high-speed "arms race" as "the most capital-intensive startup in history."

Oracle's "Stunning" Order

Oracle's earnings conference revealed a "stunning" order.

The financial report showed that the company's "remaining performance obligations" (i.e., contracted revenue not yet delivered) surged 359% year-on-year to $455 billion by the end of the first fiscal quarter. This figure far exceeds market expectations, primarily driven by an unprecedented cooperation agreement signed with OpenAI.

According to media reports, the agreement involves providing OpenAI with 4.5 gigawatts of data center capacity, valued at approximately $30 billion annually. This cooperation may involve the "Stargate" AI infrastructure plan, which includes Oracle, OpenAI, and SoftBank. To meet OpenAI's enormous computing demands, Oracle has planned to build new data centers in multiple locations across the United States.

Supported by strong demand, Oracle CEO Safra Catz raised the capital expenditure forecast for this fiscal year to $35 billion and expects the cloud infrastructure (OCI) business to grow by 77% this fiscal year, far exceeding previous expectations.

She stated that the company signed four contracts worth billions of dollars with three clients in the first quarter and anticipates that the total RPO may exceed $500 billion in the coming months. The market's enthusiastic response indicates that investors view Oracle as another cloud computing giant that will directly benefit from the AI wave, following Microsoft.

Broadcom's "Mysterious Client" and Self-Developed Chip Ambitions

In the chip sector, OpenAI has also stirred wavesBroadcom revealed in its recent earnings call that the company has secured its fourth major custom AI chip client and signed a production order worth $10 billion. Broadcom CEO Hock Tan stated that this new client has brought "immediate and quite substantial demand," and is expected to "ship very strongly" starting next year.

According to media reports, this unnamed "mysterious client" is OpenAI. This move signifies that OpenAI is following in the footsteps of giants like Google and Amazon by collaborating with Broadcom to design and mass-produce self-developed AI chips, aiming to break through computing power bottlenecks and reduce reliance on market leader NVIDIA. The custom chip, referred to as "XPU," is expected to help Broadcom capture a larger share of the AI infrastructure market.

For OpenAI, developing its own chips is a fundamental measure to satisfy its insatiable demand for computing power. Its CEO Sam Altman has stated that, considering the training needs of stronger models like GPT-5, the company is prioritizing computing power issues. Through collaboration with Broadcom, OpenAI aims to build a more controllable and efficient computing power foundation from the ground up.

Google Enters the Fray, AI Computing Power Competition Heats Up

In this battle for computing power, Google has also not remained on the sidelines.

Google Cloud CEO Thomas Kurian recently revealed at a conference that the department has unfulfilled commitments worth up to $106 billion in existing customer contracts, with at least $58 billion expected to convert into actual revenue over the next two years.

Additionally, media reports indicate that OpenAI has reached a collaboration with Google Cloud, which will provide computing power support for its AI model training. This partnership breaks Microsoft's exclusive cloud supplier status for OpenAI.

Analysts at Morgan Stanley believe that this move could not only serve as a catalyst for accelerating growth in Google Cloud's business but also reflects Google's internal confidence in the long-term competitiveness of its core search business—it is not afraid to arm a potential competitor.

For reference, earlier this year, OpenAI signed a computing power agreement worth billions of dollars with cloud service provider CoreWeave. The continuous expansion of suppliers highlights OpenAI's urgency in acquiring massive computing power.

The Most "Burning Money" Startup in History? OpenAI's $100 Billion Gamble

Supporting this series of massive orders is OpenAI's astonishing capital consumption plan.

Wall Street Insights previously mentioned that according to The Information, based on the company's latest financial forecasts disclosed to shareholders, from this year to 2029, OpenAI's cumulative cash consumption is expected to reach as high as $115 billion, a figure that has surged by about $80 billion compared to expectations six months ago.**

The massive funds are mainly used in four areas: investing nearly $100 billion in building servers and other infrastructure; paying for AI model training costs that far exceed expectations; covering the inference costs of running AI models; and paying high stock compensation to retain top experts in the fierce talent war.

Despite the astonishing rate of "burning money," the capital market remains enthusiastic. Investment institutions, including SoftBank, are still actively buying its equity, pushing OpenAI's latest valuation to $500 billion, nearly double what it was six months ago.

At the same time, driven by the strong growth of ChatGPT, OpenAI has also raised its revenue forecast for 2030 to $200 billion, showing extreme optimism about its commercialization prospects.

New "Play" in Finance Under the AI Arms Race

OpenAI's frenzied investment is a microcosm of the entire tech industry's AI "arms race." The high cost of this competition has led even cash-rich tech giants to seek new financial solutions.

An article from Wall Street Journal previously pointed out that to fund AI infrastructure construction without compromising balance sheet health, tech companies are collaborating with Wall Street, using complex financial instruments such as joint ventures, standby guarantee agreements, and syndicated loans to "externalize" some of the costs and risks.

For example, Meta has moved its data center project off-balance sheet by forming a joint venture with an investment firm; while Oracle has chosen to become a tenant of ultra-large data centers financed by a bank syndicate.

However, the frenzy of capital has also brought new risks.

UBS analysts warn that the influx of private credit may increase the risk of market overheating. At the same time, the concentration of lease contracts among a few tech giants has raised concerns about excessive credit concentration.

Rating agencies Moody's and S&P have warned Oracle, pointing out that its leverage is too high as it enters the AI construction phase, and if not improved, its credit rating may face a downgrade risk. This AI-driven feast is testing Silicon Valley's innovation capabilities, as well as its financial discipline