AI Capital Expenditures Surge Amidst Market Pressures and Competition

In August 2026, the narrative surrounding artificial intelligence (AI) shifted dramatically from mere demand to the financial implications of that demand. Major tech companies, including Amazon, Google, Microsoft, and Meta, collectively spent $165 billion on capital expenditures in the second quarter, marking an 87% increase year-over-year. This surge in spending highlighted a growing urgency to build AI infrastructure, but it also raised critical questions about financing and the sustainability of these investments.

The financial landscape for AI is evolving. Nvidia, a leader in the sector, is collaborating with major investment firms like Apollo Global and Goldman Sachs on a staggering $500 billion funding package. This includes a commitment of up to $105 billion to support an OpenAI data center in Ohio. Similarly, Broadcom is exploring a financing package nearing $100 billion. This shift indicates that AI infrastructure is no longer just about securing orders; it is increasingly tied to complex financing arrangements, raising concerns about who will fund the buildout and how returns will be managed.

Memory supply has emerged as a significant bottleneck in AI infrastructure. Micron reported that its data center customers are requesting 50% more memory than it can supply, leading to long-term contracts that differ from traditional spot-price cycles. This scarcity is expected to drive up prices, with Nvidia indicating that memory costs could increase server prices by over 15%. The ability of companies to pass these costs onto customers will be crucial in determining their profitability amidst rising expenses.

Nvidia's recent financial performance exemplifies the robust demand for AI. The company reported $96.2 billion in revenue, with data center revenue soaring 117% year-over-year. However, the competitive landscape is also intensifying. OpenAI’s new Jalapeño inference chip, developed with Broadcom, promises significant performance advantages over Nvidia’s offerings. This emerging competition suggests that the future of AI spending may not solely depend on Nvidia's GPUs, challenging the company's market dominance.

The neocloud segment, represented by firms like CoreWeave and Nebius, is also experiencing rapid growth. CoreWeave's Q2 revenue reached $2.6 billion, with a backlog of $104 billion, while Nebius reported a staggering 514% increase in AI cloud revenue. However, these companies face the challenge of converting their substantial contracts and power commitments into actual cash flow, especially as financing costs rise.

In the broader software market, fears that AI would undermine traditional software companies have not materialized uniformly. Firms like Palantir and CrowdStrike reported significant revenue growth, while others, such as HubSpot and Datadog, faced declines. This divergence indicates that AI is reshaping the software landscape, with companies that can effectively integrate AI into their offerings standing to gain a competitive edge.

Despite strong earnings across the S&P 500, concerns linger about the sustainability of growth in the face of rising interest rates and inflation. The 30-year Treasury yield recently climbed to 5.34%, raising the cost of capital for tech investments. This environment complicates the financing landscape for AI, as companies must now navigate higher borrowing costs while attempting to scale their operations.

The recent struggles of Shein, which is seeking a valuation significantly lower than its previous peak, serve as a cautionary tale for the tech sector. Public investors are becoming more discerning, demanding clearer paths to profitability amid slowing growth. This trend could signal a broader shift in investor sentiment, impacting how AI companies raise capital and manage expectations.

Looking ahead, several critical questions remain. Can memory suppliers maintain margins as new capacity comes online? Will neocloud companies successfully convert their backlog into free cash flow? Can Nvidia sustain its pricing power in the face of rising memory costs and increasing competition from custom silicon? As the cost of capital rises, the viability of various AI business models will come under scrutiny, determining which companies can thrive in this evolving landscape.