Elon Musk's xAI may allow Cursor to use its Computing power, and what makes this important for three of the world's biggest Cloud companies -- Google, Amazon, Microsoft
Elon Musk ’s artificial intelligence company xAI is now preparing to share its massive computing infrastructure with coding startup Cursor , according to a report by Business Insider. The arrangement could mark an important moment in the AI arms race, effectively positions xAI as a cloud provider and challenging the dominance of Google, Amazon, and Microsoft in the lucrative cloud computing market. The report further adds that Cursor is planning to train its upcoming AI coding model, Composer 2.5 using tens of thousands of xAI’s graphic processing units (GPUs). These chips are the backbone of AI model training, and access to them has become of the most competitive aspects of the industry. By renting out its GPUs, xAI could generate new revenue streams while offsetting the enormous costs of building and operating its Colossus data centers.
Why this change matters for Google, Amazon and Microsoft
The BI report also mentions that this is important because it mirrors the business of the world’s largest cloud providers:
* Amazon Web Services (AWS), Microsoft Azure, and Google Cloud collectively own millions of GPUs and rent computing power to thousands of companies, generating billions in profits.
* Newer players like CoreWeave and Lambda have built entire businesses around supplying GPUs to AI developers.
If xAI begins offering its infrastructure to startups, it could chip away at the dominance of the “big three” cloud companies, while also reshaping the economics of AI development.
xAI’s Colossus project
Elon Musk has long argued that xAI’s competitive edge lies in sheer computing power. The company reportedly has around 200,000 Nvidia GPUs, with plans to expand to 1 million units. This scale rivals or even surpasses some hyperscalers.
However, internal challenges remain. Last week, xAI’s president Michael Nicolls admitted in a staff memo that the company’s Model FLOPs Utilization (MFU) — a measure of GPU efficiency — was “embarrassingly low” at about 11%, compared to industry averages of 35–45%. He set a target of 50% in the coming months.
Cursor’s rising profile
Cursor, valued at around $50 billion, has quickly become a major player in AI coding tools. Its Composer 2 model, released in March, was built on an open-source model from Chinese startup Moonshot AI and fine-tuned using Cursor’s developer data.
The partnership with xAI could give Cursor access to unprecedented computing resources, helping it compete with rivals like Anthropic and OpenAI, which are aggressively expanding their own coding assistants.
The collaboration also reflects growing ties between the two companies. In March, Cursor’s former product engineering leads Andrew Milich and Jason Ginsburg joined xAI, now overseeing product teams that report directly to Musk and xAI president Nicolls.
The cloud competition heats up
For Google, Amazon, and Microsoft, the prospect of xAI entering the cloud market is a potential disruption. These companies have relied on their cloud divisions as profit engines, renting computing power to startups and enterprises alike. If xAI begins offering GPU access at scale, it could:
* Intensify competition in cloud services.
* Lower costs for startups seeking alternatives to the big three.
* Shift power dynamics in the AI ecosystem, where access to compute is as critical as algorithms themselves.
Why this change matters for Google, Amazon and Microsoft
The BI report also mentions that this is important because it mirrors the business of the world’s largest cloud providers:
* Amazon Web Services (AWS), Microsoft Azure, and Google Cloud collectively own millions of GPUs and rent computing power to thousands of companies, generating billions in profits.
If xAI begins offering its infrastructure to startups, it could chip away at the dominance of the “big three” cloud companies, while also reshaping the economics of AI development.
xAI’s Colossus project
Elon Musk has long argued that xAI’s competitive edge lies in sheer computing power. The company reportedly has around 200,000 Nvidia GPUs, with plans to expand to 1 million units. This scale rivals or even surpasses some hyperscalers.
However, internal challenges remain. Last week, xAI’s president Michael Nicolls admitted in a staff memo that the company’s Model FLOPs Utilization (MFU) — a measure of GPU efficiency — was “embarrassingly low” at about 11%, compared to industry averages of 35–45%. He set a target of 50% in the coming months.
Cursor’s rising profile
Cursor, valued at around $50 billion, has quickly become a major player in AI coding tools. Its Composer 2 model, released in March, was built on an open-source model from Chinese startup Moonshot AI and fine-tuned using Cursor’s developer data.
The partnership with xAI could give Cursor access to unprecedented computing resources, helping it compete with rivals like Anthropic and OpenAI, which are aggressively expanding their own coding assistants.
The collaboration also reflects growing ties between the two companies. In March, Cursor’s former product engineering leads Andrew Milich and Jason Ginsburg joined xAI, now overseeing product teams that report directly to Musk and xAI president Nicolls.
The cloud competition heats up
For Google, Amazon, and Microsoft, the prospect of xAI entering the cloud market is a potential disruption. These companies have relied on their cloud divisions as profit engines, renting computing power to startups and enterprises alike. If xAI begins offering GPU access at scale, it could:
* Intensify competition in cloud services.
* Shift power dynamics in the AI ecosystem, where access to compute is as critical as algorithms themselves.
