All articles tagged with #hyperscalers
Nvidia heads into earnings after the bell with lofty expectations tied to AI growth, while investors weigh its reliance on hyperscalers; broader markets see oil slide due to Hormuz talks, the Trump-Saudi nuclear agreement moving to Congress, Canada’s retaliatory tariffs on U.S. goods, and a closing 'And Finally' item reporting Dolly Parton’s death.
Nvidia’s earnings will test whether its AI-data-center growth can outpace dependence on a small group of hyperscalers (Amazon, Google, Microsoft, plus Meta and SpaceX). While ACIE revenue has been expanding faster, investors worry about continued concentration and the sustainability of hyperscaler spending, even as Nvidia ramps Vera Rubin shipments and pursues GPU-based asset financing to broaden its customer base. Analysts expect Q2 revenue around $92 billion with data-center sales representing the bulk, and full-year revenue near $396 billion as the company seeks diversification beyond its top customers.
Hyperscalers including Amazon, Microsoft, Google, Meta and Oracle plan to spend about $750B this year on AI infrastructure for data centers—the largest capital project in human history—driving a nationwide buildout that strains utilities, land, water, and skilled labor while sparking a broad, bipartisan political backlash at federal, state, and local levels; Meta’s 5 GW Hyperion campus and Nvidia-backed plans illustrate the scale and economic ripple effects across construction, energy, and local communities.
CNBC’s Jim Cramer argues rising political opposition to data-center projects could slow the buildout and pressure valuations of traditional vendors, but tighter rules may primarily benefit hyperscalers—Amazon, Alphabet, Microsoft and Meta—who can absorb compliance costs and outcompete smaller developers, potentially reducing competition for land, labor and electricity.
Morgan Stanley argues the AI boom is entering a more capital-intensive phase, with 2027 capex for key hyperscalers rising about 57% from 2026 as they expand data centers and power infrastructure. While returns on invested capital could stay above 25%, the upfront cash required creates a widening financing gap since revenue and free cash flow lag. Credit spreads have widened for weaker borrowers, though cash-rich players like Nvidia and Broadcom have more financing flexibility, and private capital and asset-backed financing will play a larger role in funding hardware and infrastructure. Investors should watch 2027 capex guidance, free-cash-flow revisions, and financing activity to gauge who can sustain aggressive AI investments.
Nvidia signs MOUs with six major asset managers to create compute financing platforms that could raise up to $500 billion to fund AI hardware, borrowing against GPUs and related infrastructure and potentially guaranteeing part of the deals. This moves AI capex off corporate books, easing hyperscalers’ spending while raising questions about GPU depreciation, credit risk, and the circularity of financing the AI boom.
Micron Technology (MU) jumped in premarket trading to clear the $1,000 level, its first time above since July 6, up about 3.4% and 240% year-to-date, as hyperscalers boost demand for high-bandwidth memory (HBM3E/HBM4) used in AI accelerators. Analysts cite improving demand signals and ongoing supply tightness through 2027, helping fuel the renewed rally.
SpaceX plans a massive expansion of AI compute, targeting 6–10 GW of capacity and potentially spending $300–$500 billion in 2027 on data centers, as it rents large-scale compute capacity to Anthropic and Google and as rivals like Microsoft and Alphabet race to secure chips and land. The boom hinges on favorable economics (high margins and rapid GPU payback), but analysts warn bottlenecks at TSMC and potential price bullwhips could temper the surge and alter ROI and funding dynamics for the broader AI infrastructure push.
Despite forecasts of roughly $1 trillion in AI-data-center spending, cash isn’t the bottleneck. The big limits are power availability, skilled labor, land, and permitting, plus regulatory pushback on new builds. BloombergNEF warns of a 19 GW power shortfall by 2035 if growth persists, and utilities may approve only a fraction of requested hookups. As a result, AI data-center deployment is likely to proceed more slowly and in fits and starts, even as demand remains robust and suppliers of GPUs, CPUs, memory, and power equipment stand to benefit. In a worst-case scenario, compute could be constrained and users shift to cheaper models, but the current trend is toward continued heavy investment despite the bottlenecks.
Ahead of Nvidia’s August 26 earnings, investors are pricing in another blockbuster quarter, driven by relentless demand for AI data-center chips from hyperscalers and the ramp of its Blackwell architecture. Goldman Sachs cautions that lofty expectations may limit upside, warning of a possible sell-the-news reaction if profitability signals or demand fail to meet high bars; consensus calls for roughly $92 billion in revenue (about 96% year-over-year) and about $2.08 in EPS, leaving Nvidia vulnerable to a downside surprise if any of these momentum drivers falter.
Cisco reported FY4Q net income of $3.86B on $17.25B in revenue, with adjusted EPS of $1.22, topping expectations. The quarter benefited from about $4B in AI infrastructure orders from hyperscalers, bringing its full-year AI orders to $9.3B. Management guided FY27 revenue of $72.2–$73.4B and adjusted EPS of $5.05–$5.11, with Q1 above consensus. Despite the strong results and upbeat outlook, CSCO edged lower in after-hours trading.
CoreWeave stock rose about 19% after Q2 revenue doubled to $2.6B and guided $3.4–3.6B for Q3, though profitability remains elusive amid rising expenses and debt; Nebius jumped 34% after a Q2 where revenue grew 514% to $575M with contracts quadrupling, underscoring robust AI‑compute demand from hyperscalers.
AI-hungry giants fueled a broad stock rally after solid earnings, with Microsoft and Amazon rising over 5% as Azure topped $100 billion in annual cloud revenue and AWS growth neared 37%. Alphabet, Meta, and Oracle also advanced on resilient or upbeat results, while investors weighed continued AI‑centered capex against returns. Amazon even pushed its market cap past $3 trillion, underscoring confidence in AI infrastructure demand despite ongoing spending questions.
Tech stock investors face a looming $1 trillion problem as hyperscalers accelerate AI-driven capex, with 2027 spending projected to top $1 trillion and capex set to outpace cash flow from operations from 2026 through 2028; Meta has raised its 2026 capex guidance and Alphabet signaled higher 2027 spend, while Microsoft remains the only major player with a favorable near-term free cash flow outlook.
Earnings show Amazon, Alphabet, and Microsoft turning AI infrastructure into revenue, with combined cloud backlog rising from about $800 billion to over $2.3 trillion in signed commitments. Capacity constraints kept GPU prices rising (Nvidia H100 rentals up ~63%), while Oracle trails on debt. The trio funds expansion with massive free cash flow, suggesting durable profits from AI—not just hype—making them potentially safer, long‑term AI investments.