The agreement, which runs through 2028, is designed to speed up how ChatGPT and other OpenAI products respond to users, as competition gets stronger among the world’s largest technology firms to control the infrastructure behind advanced models. 

Cerebras will provide the capacity through its own cloud services, powered by its wafer-scale chips, with new data centres to be built or leased specifically for the contract.

Seven hundred and fifty megawatts is not incremental capacity, it shows industrial-level demand. 

OpenAI is no longer thinking only about training models but about inference, the everyday work of answering questions, reasoning through problems and serving millions of users at once. 

“Integrating Cerebras into our mix of compute solutions is all about making our AI respond much faster,” OpenAI stated. 

Talks between the two companies began last August after Cerebras showed that OpenAI’s open-source models could run more efficiently on its chips than on standard graphics processors. 

Months of negotiations followed. The result is a structure where Cerebras owns and operates the hardware, while OpenAI pays to access it as a service. Capacity will be added in stages over the next few years.

Cerebras is preparing to return to public markets after withdrawing its previous listing attempt in late 2024. It now plans to re-file for an initial public offering in the second quarter of 2026. This OpenAI contract helps address a long-standing concern among investors: dependence on a single customer. 

In 2024, UAE-based technology group G42 accounted for nearly 87% of Cerebras’ revenue. A multi-billion-dollar US client changes that picture.

The deal also fits neatly into OpenAI’s bigger strategy. The company is laying the foundation for a potential initial public offering that could value it at around $1 trillion. 

Its chief executive, Sam Altman, has publicly committed $1.4 trillion to building 30 gigawatts of computing capacity, enough to power roughly 25 million homes in the United States. 

Competition is high. Nvidia still tops the market for training chips, while Google and Microsoft are pushing their own custom hardware through cloud platforms. 

By turning to Cerebras, OpenAI is clearly trying to reduce reliance on a single supplier and gain an edge in speed, particularly for reasoning-heavy models.

But then, the spending spree is unsettling some observers, with warnings that the surge in valuations and capital commitments across the sector shows the excesses of the dot-com era. 

The counter-argument is equally strong, demand for inference compute is exploding as artificial intelligence moves from labs into daily use. 

The collaboration will see Broadcom develop and deploy OpenAI-designed processors beginning in the second half of 2026, with full rollout expected by 2029.

According to both companies, the project will deliver up to 10 gigawatts of custom chips, a scale that would consume roughly the same power as eight million U.S. homes or five times the energy output of the Hoover Dam. 

Shares of Broadcom rose by more than 10% after the announcement, as investors believe in the chipmaker’s impact in the AI hardware market.

“Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential,” Sam Altman, CEO of OpenAI, said in a statement.

While the financial terms of the deal remain undisclosed, analysts say OpenAI’s plan is capital-intensive. Gadjo Sevilla, an analyst at eMarketer, noted that “Financing such a large chip deal will likely require a combination of funding rounds, pre-orders, strategic investments, and support from Microsoft, as well as leveraging future revenue streams and potential credit facilities.”

Experts estimate that a single 1-gigawatt data centre could cost between $50 billion and $60 billion, with Nvidia products alone accounting for more than half of those expenses. This places the estimated cost of OpenAI’s 10GW buildout at a huge scale, stressing the financial commitment required to sustain its growing AI infrastructure.

The Broadcom partnership follows a string of recent hardware deals by OpenAI. Just last week, the company announced a 6-gigawatt chip supply agreement with AMD, which includes an option to acquire a stake in the chipmaker. 

Days earlier, Nvidia revealed plans to invest up to $100 billion in OpenAI and provide advanced data-centre systems with at least 10GW of capacity.

In joining the ranks of companies like Google, Amazon, and Meta, which already design custom silicon for their AI systems, OpenAI aims to gain a stronger hold on its computing backbone. 

However, creating high-performance chips from scratch remains a big technical challenge. Even Microsoft and Meta’s internal chip projects have struggled to match Nvidia’s performance in the AI accelerator space.

For Broadcom, the collaboration strengthens its standing as one of the biggest beneficiaries of the AI hardware boom. The company’s stock has surged nearly sixfold since the end of 2022, driven by rising demand for custom chips and networking solutions. 

Last month, Broadcom revealed a $10 billion order from an unnamed new client, though the company later clarified that OpenAI was not that customer.

The upcoming OpenAI chips will be scaled entirely on Broadcom’s Ethernet and networking infrastructure, directly challenging Nvidia’s InfiniBand system, which currently tops high-performance AI workloads.

If OpenAI meets its 2026 production target, it would be one of the fastest chip development turnarounds in the industry’s history, and a big moment for a company aiming to keep pace with its own growth.

The Japanese group, SoftBank will purchase Intel common stock at $23 per share, giving it close to a 2% stake and making it one of Intel’s top shareholders. 

The announcement, made after U.S. markets closed on Monday, sent Intel shares up more than 5% in after-hours trading, while SoftBank’s stock slipped over 5% in Tokyo.

Masayoshi Son, chairman and CEO of SoftBank, said, “Semiconductors are the foundation of every industry. For more than 50 years, Intel has been a trusted leader in innovation. This strategic investment reflects our belief that advanced semiconductor manufacturing and supply will further expand in the United States, with Intel playing a critical role.”

The investment lands at a challenging moment for Intel. Under new CEO Lip-Bu Tan, the company has been cutting costs and narrowing focus. It shut down its automotive chip division earlier this year, slashed between 15% and 20% of its foundry workforce, and doubled down on its client and data centre chip portfolio, where it still lags behind rivals Nvidia and AMD.

Tan, who has been navigating both corporate restructuring and political issues, welcomed SoftBank’s support. “We are very pleased to deepen our relationship with SoftBank, a company that’s at the forefront of so many areas of emerging technology and innovation and shares our commitment to advancing U.S. technology and manufacturing leadership. Masa and I have worked closely together for decades, and I appreciate the confidence he has placed in Intel with this investment.”

Beyond chips, SoftBank’s CEO Son has been positioning the group as a key player in the next wave of artificial intelligence infrastructure. Earlier this year, SoftBank acquired Foxconn’s Lordstown, Ohio factory, which will be repurposed to manufacture AI data centre equipment. 

The site forms part of the $500 billion Stargate project, a venture involving SoftBank, Oracle and OpenAI, aimed at building large-scale AI server capacity in the United States.

Meanwhile, U.S. politics are heavy over the semiconductor sector. The Trump administration has floated the idea of converting CHIPS Act funds into equity, potentially giving Washington as much as a 10% stake in Intel.

The White House has also threatened new tariffs on imported chips in a bid to strengthen domestic production and cut reliance on Asian giants like TSMC and Samsung.