CoreWeave is significantly increasing its spending plans for 2026 as demand for artificial intelligence computing continues to grow at a rapid pace. The AI-focused cloud infrastructure company raised its capital expenditure forecast after reporting stronger-than-expected second-quarter results, highlighting how quickly companies are expanding data-centre capacity to meet the computing requirements of advanced AI models.
CoreWeave now expects to spend between $35 billion and $39 billion on capital expenditure in 2026, up from its earlier forecast of $31 billion to $35 billion. The increase reflects the company’s decision to accelerate investments in data centres, Nvidia GPUs and other infrastructure needed to provide AI computing services to customers.
The decision comes as CoreWeave’s business is experiencing extraordinary demand. The company reported second-quarter revenue of about $2.58 billion, more than double the $1.21 billion recorded in the same quarter a year earlier. Revenue also came in slightly ahead of Wall Street expectations, showing that the company’s rapid expansion is translating into substantial sales growth.
Perhaps the strongest indicator of demand is CoreWeave’s backlog. At the end of the second quarter, its revenue backlog had climbed to around $104.2 billion, compared with $99.4 billion at the end of the first quarter. The company also secured more than $25 billion in new customer commitments during the quarter.
A large backlog gives CoreWeave significant visibility into future revenue. However, the contracts also create pressure on the company to build the necessary computing infrastructure quickly. This is one of the main reasons capital spending has risen so sharply.
CoreWeave’s business model is different from that of traditional cloud providers. Instead of offering a broad range of general-purpose cloud services, the company focuses heavily on GPU-based computing infrastructure designed for AI training, inference and other demanding workloads. Its close relationship with Nvidia gives it access to advanced graphics processors that are increasingly difficult to secure in large quantities.
The rapid growth of generative AI has created a huge requirement for computing power. Companies developing large language models and other AI systems need thousands of GPUs working together inside specialised data centres. Training and operating these systems also requires high-speed networking, advanced cooling systems, large amounts of electricity and sophisticated data-centre management.
This is creating an infrastructure race across the technology industry. Large technology companies are spending billions of dollars on AI data centres, while specialised cloud providers such as CoreWeave are expanding capacity to serve customers that do not want to build all of that infrastructure themselves.
CoreWeave’s second-quarter capital expenditure reached approximately $9.4 billion, compared with $6.8 billion in the previous quarter. The company’s earnings presentation showed $9.35 billion in capital expenditure for the quarter, underlining the scale of the investment programme.
The spending is producing rapid physical expansion. CoreWeave added around 500 megawatts of computing capacity during the quarter and has outlined an ambitious longer-term plan to reach approximately 5 gigawatts of capacity by 2030.
However, the company’s rapid growth comes with significant financial costs. CoreWeave reported a net loss of roughly $626 million for the second quarter. The company is therefore spending enormous amounts of money today in the expectation that long-term AI demand will generate sufficient revenue and cash flow to justify the investment.
Debt and financing are consequently important issues for investors. Building AI data centres requires huge upfront investment, and CoreWeave has relied on financing arrangements to support its expansion. The company recently secured a new $2.6 billion loan facility, adding another source of funding for its infrastructure programme.
The company also has strong relationships with some of the biggest names in the AI industry. Its customer base includes major technology companies and AI developers, including Meta and Anthropic. These relationships are helping CoreWeave build a large pipeline of contracted demand.
The bigger question for investors is whether the current AI spending boom can continue long enough to support such enormous infrastructure investment. The demand outlook remains strong, but the economics of AI cloud computing are closely tied to GPU utilisation, electricity costs, financing expenses and the prices customers are willing to pay for computing capacity.
There are also risks from shortages of power, advanced chips and other data-centre components. Even when companies have the money to build new facilities, securing electricity connections, GPUs and specialised equipment can take time. These constraints could limit how quickly CoreWeave and its competitors can turn their investment plans into operational capacity.
For the wider technology industry, CoreWeave’s spending increase is another sign that AI is becoming a major driver of global infrastructure investment. Demand for GPUs is supporting semiconductor companies, while data-centre construction is creating opportunities for power providers, networking companies, cooling-system manufacturers and construction firms.
The company’s results also provide an important signal for Nvidia and the broader AI ecosystem. CoreWeave is one of the major customers using Nvidia’s advanced processors to provide cloud-based AI computing. As CoreWeave expands its capacity, demand for Nvidia’s chips and related infrastructure could remain strong.
CoreWeave shares reacted positively to the results, rising sharply in after-hours trading as investors focused on the company’s strong revenue growth and enormous backlog.
For now, CoreWeave is betting heavily on the idea that AI demand will continue to expand. Its decision to raise 2026 capital spending to as much as $39 billion shows the scale of that bet.
The company is moving from being a specialised GPU cloud provider toward becoming a major piece of the global AI infrastructure network. Whether the strategy ultimately delivers strong returns will depend on how efficiently CoreWeave can convert its massive backlog into revenue while controlling debt, operating costs and capital spending.
For the broader market, however, one message is already clear: the AI boom is no longer only about software and chips. It is driving a massive build-out of physical computing infrastructure, and companies such as CoreWeave are at the centre of that transformation.
