What is the story about?
Meta is doubling down on its artificial intelligence ambitions with massive investments in infrastructure, yet a key part of its strategy relies on outsourcing critical capacity to external providers.
The company’s latest multibillion-dollar agreement with Nebius Group underlines this contradiction, as it races to build an AI empire while still depending on third-party cloud systems.
The deal comes at a time when Big Tech is funnelling unprecedented sums into AI, with Meta alone expected to spend hundreds of billions of dollars in the coming years.
Last month, Meta declared that its capital expenditure could almost double to $135bn in 2026 as its AI spending plans get increasingly unhinged. Meta CEO Mark Zuckerberg said, “We are now seeing a major AI acceleration. I expect 2026 to be a year where this wave accelerates even further."
Shares of Nebius Group surged sharply after the company announced that Meta had agreed to spend up to $27 billion over the next five years to secure AI infrastructure capacity. The Amsterdam-listed firm saw its stock jump around 16 per cent, while Meta shares also edged higher.
Under the agreement, Nebius will provide Meta with $12 billion worth of dedicated computing capacity starting in early 2027. In addition, the social media giant has committed to purchasing up to $15 billion more in capacity that Nebius is developing for other customers.
The partnership marks one of Meta’s largest infrastructure deals to date, reflecting the scale of investment required to support advanced AI systems. It also builds on an existing relationship between the two companies, following a separate $3 billion agreement signed last year.
The announcement follows closely on the heels of Nvidia’s decision to invest $2 billion in Nebius as part of a broader push to expand AI data centre infrastructure. Nebius said this backing would help it deploy more than 5 gigawatts of Nvidia-powered systems by 2030.
Nebius is part of a growing class of “neocloud” providers specialising in AI workloads, competing with established giants such as Google and Amazon. These newer players are attracting support from chipmakers like Nvidia as demand for high-performance computing continues to surge.
Despite the external partnerships, Nebius CEO Arkady Volozh indicated that the company is scaling aggressively on its own. “We are pleased to expand our significant partnership with Meta… We will continue to deliver,” he said, noting that the company expects capital expenditure of between $16 billion and $20 billion this year.
While Meta is committing enormous resources to AI infrastructure, it is also reportedly considering deep cost-cutting measures.
A few days ago, reports claim that the company is planning layoffs that could affect 20 per cent or more of its workforce , as it looks to offset the rising costs of its AI push.
No final decision or timeline has been confirmed, but the potential scale of the cuts highlights the financial strain of building and maintaining cutting-edge AI systems. These investments include not just data centres, but also specialised hardware such as GPUs and custom-designed chips.
Meta has made AI its central strategic priority, signing major agreements with chipmakers like Nvidia and Advanced Micro Devices while also developing its own in-house silicon. Zuckerberg has previously outlined plans to spend up to $600 billion on infrastructure projects by 2028, funded largely through advertising revenue and external financing.
Across the industry, spending is accelerating rapidly. Major technology firms are expected to collectively invest around $650 billion in 2026 alone on AI-related infrastructure.
The contrast is stark: even as Meta builds out its own capabilities, it continues to rely on external partners like Nebius to meet immediate demand. This dual approach underscores the sheer scale of the AI race, where speed often takes precedence over self-sufficiency, and where even the biggest players cannot go it alone.
The company’s latest multibillion-dollar agreement with Nebius Group underlines this contradiction, as it races to build an AI empire while still depending on third-party cloud systems.
The deal comes at a time when Big Tech is funnelling unprecedented sums into AI, with Meta alone expected to spend hundreds of billions of dollars in the coming years.
Last month, Meta declared that its capital expenditure could almost double to $135bn in 2026 as its AI spending plans get increasingly unhinged. Meta CEO Mark Zuckerberg said, “We are now seeing a major AI acceleration. I expect 2026 to be a year where this wave accelerates even further."
The deal
Shares of Nebius Group surged sharply after the company announced that Meta had agreed to spend up to $27 billion over the next five years to secure AI infrastructure capacity. The Amsterdam-listed firm saw its stock jump around 16 per cent, while Meta shares also edged higher.
Under the agreement, Nebius will provide Meta with $12 billion worth of dedicated computing capacity starting in early 2027. In addition, the social media giant has committed to purchasing up to $15 billion more in capacity that Nebius is developing for other customers.
The partnership marks one of Meta’s largest infrastructure deals to date, reflecting the scale of investment required to support advanced AI systems. It also builds on an existing relationship between the two companies, following a separate $3 billion agreement signed last year.
The announcement follows closely on the heels of Nvidia’s decision to invest $2 billion in Nebius as part of a broader push to expand AI data centre infrastructure. Nebius said this backing would help it deploy more than 5 gigawatts of Nvidia-powered systems by 2030.
Nebius is part of a growing class of “neocloud” providers specialising in AI workloads, competing with established giants such as Google and Amazon. These newer players are attracting support from chipmakers like Nvidia as demand for high-performance computing continues to surge.
Despite the external partnerships, Nebius CEO Arkady Volozh indicated that the company is scaling aggressively on its own. “We are pleased to expand our significant partnership with Meta… We will continue to deliver,” he said, noting that the company expects capital expenditure of between $16 billion and $20 billion this year.
Meta layoffs 20 per cent of its workforce
While Meta is committing enormous resources to AI infrastructure, it is also reportedly considering deep cost-cutting measures.
A few days ago, reports claim that the company is planning layoffs that could affect 20 per cent or more of its workforce , as it looks to offset the rising costs of its AI push.
No final decision or timeline has been confirmed, but the potential scale of the cuts highlights the financial strain of building and maintaining cutting-edge AI systems. These investments include not just data centres, but also specialised hardware such as GPUs and custom-designed chips.
Meta has made AI its central strategic priority, signing major agreements with chipmakers like Nvidia and Advanced Micro Devices while also developing its own in-house silicon. Zuckerberg has previously outlined plans to spend up to $600 billion on infrastructure projects by 2028, funded largely through advertising revenue and external financing.
Across the industry, spending is accelerating rapidly. Major technology firms are expected to collectively invest around $650 billion in 2026 alone on AI-related infrastructure.
The contrast is stark: even as Meta builds out its own capabilities, it continues to rely on external partners like Nebius to meet immediate demand. This dual approach underscores the sheer scale of the AI race, where speed often takes precedence over self-sufficiency, and where even the biggest players cannot go it alone.
