OpenAI Reportedly Plans to Buy Google’s AI Chips as Competition With Nvidia Intensifies

OpenAI, the company behind ChatGPT, reportedly plans to purchase artificial intelligence chips developed by Google.

The potential agreement could deepen cooperation between two major companies that compete in the AI industry.

According to a Reuters report published on October 28, 2026, citing an earlier report by The Information, OpenAI is considering purchasing Google’s Tensor Processing Units, commonly known as TPUs.

These specialised computer chips are designed to handle the enormous calculations required by artificial intelligence systems.

OpenAI currently relies heavily on Nvidia technology for AI computing.

A deal involving Google’s chips could give OpenAI another option for powering its growing AI infrastructure.

However, neither OpenAI nor Google has publicly confirmed the reported purchase.

So why would OpenAI buy AI chips from a competitor, and what makes these processors so important?

What Are AI Chips?

Artificial intelligence systems need substantial computing power to function.

For example, when someone asks ChatGPT a question, powerful computers process the request and generate a response.

Training advanced AI models can require even more computing resources.

These calculations are performed by specialised computer processors.

Traditional computers use central processing units, or CPUs, to perform many general computing tasks.

Modern AI systems often rely on graphics processing units, or GPUs, because they can perform large numbers of calculations simultaneously.

Other processors, including Google’s TPUs, are designed specifically for certain types of AI computation.

These chips form an important part of the infrastructure supporting modern artificial intelligence.

What Are Google’s Tensor Processing Units?

Tensor Processing Units are specialised AI chips developed by Google.

They are designed to accelerate mathematical operations commonly used in machine learning.

Machine learning is a branch of artificial intelligence in which computer systems learn patterns from data.

For example, a machine learning model might analyse thousands of photographs to learn how to recognise different objects.

Training and running such models involves large numbers of mathematical calculations.

Google’s TPUs are designed to perform these operations efficiently.

The company has used TPUs within its own infrastructure and has expanded their availability to cloud customers.

Businesses can use supported TPU services for tasks involving AI model training and inference.

Inference is the process of using a trained AI model to generate predictions or responses.

Why Does OpenAI Want Google’s AI Chips?

According to the Reuters report, the proposed purchase could help OpenAI expand its computing options.

The company already uses substantial computing resources to train and operate its AI models.

As more people and businesses use AI tools, demand for computing infrastructure continues to grow.

Buying or using chips from multiple suppliers could potentially help OpenAI manage costs, improve availability, and reduce dependence on any single provider.

However, OpenAI has not publicly explained the reasons for the reported purchase.

The proposed deal therefore should not be presented as a confirmed decision to replace Nvidia.

What Is Nvidia’s Role in Artificial Intelligence?

Nvidia is one of the most important companies supplying processors used in modern AI systems.

Its GPUs are widely used for training and running machine learning models.

Training a large AI model involves processing enormous amounts of data and updating model parameters through repeated calculations.

GPUs are well suited to many of these tasks because they can handle large numbers of calculations in parallel.

Nvidia also provides software and networking technologies that support AI computing.

This combination has helped the company establish a strong position in the AI infrastructure market.

However, competitors are developing alternative processors.

Google’s TPUs are one example.

How Are Google’s TPUs Different From Nvidia GPUs?

Both GPUs and TPUs can perform calculations used in artificial intelligence, but they have different designs and software ecosystems.

Nvidia GPUs

GPUs are designed for parallel computing and support many different workloads.

They are widely used in AI research, scientific computing, graphics processing, and other applications.

Nvidia provides CUDA, a software platform that allows developers to use compatible GPUs for general-purpose computing.

Google TPUs

TPUs are specialised accelerators developed for machine learning workloads.

They are designed to perform certain mathematical operations efficiently.

Google provides software tools and cloud services that allow developers to use these processors.

Neither type of chip is automatically better for every task.

Performance and cost depend on the AI model, software, hardware configuration, and workload.

Has OpenAI Used Google’s AI Chips Before?

Yes, according to previous reporting.

Reuters reported in June 2025 that OpenAI had begun renting Google’s AI chips.

The latest report suggests OpenAI could expand that relationship by purchasing TPUs directly.

This would represent a different arrangement from renting access to processors through cloud infrastructure.

However, the financial terms, purchase quantities, and delivery arrangements have not been disclosed.

It is also unclear whether the companies have reached a final agreement.

Aren’t OpenAI and Google Competitors?

Yes.

OpenAI develops ChatGPT and other AI models, while Google develops Gemini and its own artificial intelligence products.

Both companies compete for users, developers, and business customers.

However, competition does not prevent technology companies from working together in other areas.

A company might compete with another business in software while purchasing hardware or infrastructure services from it.

For example, a smartphone manufacturer might buy components from another company that also produces competing smartphones.

The technology industry frequently involves these overlapping relationships.

OpenAI and Google also announced plans earlier in October to make OpenAI models available through Google Cloud.

The reported chip purchase could deepen their existing commercial relationship.

Is OpenAI Developing Its Own AI Chips?

Yes.

OpenAI has also been developing custom AI hardware with Broadcom.

According to Reuters, OpenAI announced its first custom AI chip in June 2026 and planned to begin deploying it by the end of the year.

Developing custom chips could eventually give the company greater control over the hardware used to run its AI systems.

However, designing and manufacturing advanced processors is a complex and expensive process.

A company can therefore develop its own chips while continuing to purchase hardware from other suppliers.

Why Are Technology Companies Spending So Much on AI Infrastructure?

Training and operating advanced AI systems requires data centres, computer processors, networking equipment, electricity, and cooling systems.

A data centre is a facility that houses computers and equipment used to process and store information.

As demand for AI services increases, technology companies are building and expanding these facilities.

Reuters reports that Amazon, Microsoft, and Google’s parent company Alphabet have plans that could amount to as much as $600 billion in combined AI infrastructure spending during 2026.

This figure represents planned or projected expenditure, not necessarily money already spent.

The scale of investment illustrates how important computing infrastructure has become in the AI industry.

However, companies still need to demonstrate that their investments will generate sufficient financial returns.

Could More Competition Make AI Cheaper?

Potentially.

If technology companies have access to more chip suppliers, they may gain additional options when building AI infrastructure.

Greater competition could encourage improvements in pricing, performance, and energy efficiency.

Lower infrastructure costs could eventually help reduce the cost of some AI services.

However, this outcome is not guaranteed.

The price users pay for AI depends on many factors, including development expenses, computing costs, demand, and business decisions.

Therefore, a potential chip agreement does not mean ChatGPT subscriptions or API prices will automatically fall.

What Does This Mean for African AI Developers?

Many African developers use cloud services provided by international technology companies.

They may not purchase powerful AI processors directly, but they can access AI computing resources through cloud platforms.

Increased competition among chip suppliers could eventually influence the cost and availability of those services.

For example, a developer in Kenya creating an AI-powered educational application might benefit if cloud computing becomes more affordable.

However, the impact will depend on service pricing, regional availability, internet infrastructure, and application requirements.

The reported OpenAI-Google agreement has not yet produced any confirmed price changes for African developers.

QueryFun Take: The AI Race Is Also a Competition Over Hardware

Most conversations about artificial intelligence focus on chatbots, image generators, and new AI models.

However, these systems depend on physical computing infrastructure.

Processors, data centres, electricity, and networking equipment determine how much AI computing capacity companies can provide.

The reported OpenAI-Google chip purchase highlights the growing importance of this infrastructure.

It also shows that companies competing in AI software may still cooperate in hardware and cloud computing.

For students interested in technology, understanding GPUs, TPUs, machine learning, and cloud computing provides useful insight into how AI works behind the scenes.

The key point is that the proposed purchase remains unconfirmed.

Readers should distinguish between reported business plans and completed agreements.

Frequently Asked Questions

Is OpenAI buying Google’s AI chips?

Reuters reports that OpenAI plans to purchase Google’s TPUs, citing The Information. Neither company has publicly confirmed a completed purchase.

What is a TPU?

A Tensor Processing Unit is a specialised processor developed by Google to accelerate machine learning calculations.

Does OpenAI currently use Nvidia GPUs?

Yes. Nvidia processors form an important part of the computing infrastructure used to train and run OpenAI models.

Will OpenAI stop using Nvidia?

There is no confirmed announcement that OpenAI intends to stop using Nvidia hardware.

Why would OpenAI buy chips from Google?

Possible reasons include expanding computing capacity, managing costs, and diversifying suppliers. OpenAI has not publicly confirmed its reasoning for the reported purchase.

Will Google’s AI chips make ChatGPT cheaper?

There is no confirmed price reduction associated with the reported purchase. Any future impact would depend on several commercial and technical factors.

Is OpenAI developing its own AI hardware?

Yes. OpenAI has announced work on custom AI chips in collaboration with Broadcom.

Primary report: Reuters: OpenAI Plans to Buy Google’s AI Chips, The Information Reports