CoreWeave backlog reaches $104B, Meta expands contracts to $21B; all four major AI labs sign deals.

The cloud landscape is rapidly shifting. For years, traditional hyperscale providers like Amazon’s (AMZN) AWS, Microsoft’s (MSFT) Azure, and Alphabet’s (GOOGL) Google Cloud have dominated the market. Now, a new wave of “neocloud” providers is flipping the script by prioritizing AI-first infrastructure. This transition is accelerating faster than anticipated, fueled by a surge in AI demand that hyperscalers are struggling to meet independently.

Given infrastructure’s pivotal role in the AI boom, neocloud providers such as CoreWeave (CRWV), Nebius (NBIS), and Lambda present investors with a novel, high-growth avenue to capitalize on artificial intelligence. These firms are effectively capturing the segment of cloud computing that hyperscalers are not optimized to handle efficiently.

**What Is a Neocloud?**

A neocloud is a highly specialized computing network engineered exclusively to run the most demanding AI workloads using top-tier graphics processing units (GPUs).

While hyperscale clouds like AWS, Azure, and Google Cloud operate as general-purpose platforms, neocloud providers are “masters of one.” Built around a singular objective—delivering dedicated, highly optimized GPU computing power—they do not aim to function as broad cloud services. In short, hyperscalers prioritize flexibility, whereas neoclouds prioritize raw AI compute capacity. This specialization translates to fewer bottlenecks, higher efficiency, and infrastructure specifically tuned for modern AI models.

CoreWeave has emerged as a key architect reshaping the AI infrastructure landscape. In September, the company secured a $14.2 billion agreement with Meta Platforms (META) to supply the tech giant with AI cloud infrastructure. The market responded swiftly, with CRWV shares surging 12% following the announcement.

Niche technologies typically require external validation to gain traction. Landing a major contract with a household name signals industry confidence and validates the underlying technology, while simultaneously de-risking the investment for prospective clients and shareholders. The Meta partnership delivered precisely that validation. It not only legitimized the neocloud model but also proved that specialized providers can directly compete with hyperscalers like AWS, Google Cloud, and Azure in targeted AI infrastructure segments.

The deal further underscored the intense demand for AI-specific infrastructure and Big Tech’s growing preference for partnering with specialized vendors. Such collaborations enable major tech firms to launch new services, accelerate technological adoption, and deliver more customized solutions.

Notably, the Meta agreement was not CoreWeave’s inaugural major partnership. In March, OpenAI entered into a similar arrangement, committing $22.4 billion to secure high-performance cloud computing capacity for training its advanced AI models. Microsoft remains another core client and strategic partner. Microsoft leverages CoreWeave’s AI infrastructure to bolster its own data center operations, while CoreWeave provides Microsoft with critical GPU access essential to its AI ecosystem. Collectively, these alliances forge a powerful validation loop: expanding client bases enhance corporate credibility, which in turn positions CoreWeave to secure increasingly lucrative contracts.

On a macro level, the CoreWeave-Meta deal marks a strategic pivot for enterprise cloud procurement. It signals that AWS, Azure, and Google Cloud are no longer the exclusive options, as organizations actively seek out alternative providers capable of delivering secure, specialized AI infrastructure. For investors, these developments serve as early confirmation of the institutional trust major tech firms place in CoreWeave—a confidence factor that has catalyzed explosive growth in neocloud demand.

**CoreWeave: A Massive Neocloud Backlog Means Less Risk**

CoreWeave has decisively achieved its initial objective of validating both its proprietary offerings and the neocloud sector as a whole, a milestone cemented by its Meta and OpenAI agreements. These contracts have generated a substantial order backlog valued at approximately $55.6 billion as of the end of Q3 2025—nearly double the size of the previous quarter’s backlog.

Such a backlog is uncommon for a firm of CoreWeave’s age, granting the company exceptional visibility into future revenue streams. This dynamic benefits CoreWeave in several key ways. First, it guarantees a robust pipeline of near-term revenue, significantly bolstering the company’s bottom line. Second, it provides investors with the long-term revenue visibility they consistently favor. Additionally, a contracted backlog serves as a stabilizing buffer against short-term market volatility, as demand is already locked in.

Another critical development is improved portfolio diversification. During fiscal year (FY) 2024, Microsoft accounted for 62% of CoreWeave’s total revenue, with Microsoft and a single other customer collectively representing 77%. Historically, this heavy concentration posed a notable risk for investors. With Meta and OpenAI now integrated into its client base, no single customer exceeds 35% of CoreWeave’s total business. This structural shift toward diversification prompted investors to aggressively buy CRWV shares following the Meta announcement, as it substantially mitigated concentration risk.

While CoreWeave’s massive backlog and projected future payouts represent highly favorable conditions for the company and its shareholders, the path to sustained profitability remains challenging.

**CoreWeave Sees Explosive Growth but New Revenue Headwinds**

CoreWeave is experiencing exceptional revenue growth. Third-quarter 2025 revenue reached $1.36 billion, marking a 134% year-over-year (YOY) increase. During its Q3 earnings call, management guided for full-year 2025 revenue to fall between $5.05 billion and $5.15 billion, representing 164% to 170% YOY growth. These figures command attention, yet they must be weighed against a $110.1 million net loss recorded in the same quarter.

Year-to-date through Q3 2025, cumulative net losses stand at $716 million. However, the Q3 loss represents a marked improvement over the $360 million loss in Q3 2024, as well as the $315 million loss in Q1 2025 and the $291 million loss in Q2 2025. Several factors contextualize these losses. Part of the shortfall stems from data center construction delays, supply chain constraints, and existing debt obligations. More significantly, a substantial portion reflects non-cash depreciation charges tied to aggressive capital expenditure (capex) investments aimed at fueling expansion. CoreWeave’s depreciation and amortization expense alone reached $630.5 million in Q3 2025, a sharp increase from $254 million in the prior-year period. Total capex for the quarter hit $1.9 billion.

**CoreWeave’s High-Risk, High-Reward Capex Strategy**

The adage “You have to spend money to make money” perfectly encapsulates CoreWeave’s current operational posture. The company projects 2025 capex to range between $12 billion and $14 billion, against forecasted revenue of $5.05 billion to $5.15 billion. Consequently, the projected capex-to-revenue ratio for fiscal year 2025 sits between 2.3x and 2.8x, meaning CoreWeave is allocating roughly $2.35 to $2.77 in capital expenditures for every dollar of revenue generated. This heavy investment cycle explains the current lack of profitability, with 2026 poised to intensify the pressure. Management expects capex to more than double next year, reaching approximately $30 billion.

This aggressive spending profile introduces substantial investment risk. Should AI demand decelerate prematurely, CoreWeave could face severe financial strain. Nevertheless, the financial structure follows a deliberate growth thesis. CoreWeave’s trajectory hinges on sustained expansion, which justifies the heavy allocation toward data center development and high-end GPU cluster acquisitions. Management maintains firm conviction that these upfront costs will yield substantial long-term returns. While inherently high-risk, the strategy carries commensurately high reward potential. The $55 billion-plus backlog strongly suggests the approach will ultimately succeed, contingent upon AI cloud service and infrastructure demand remaining robust.

**Is CoreWeave Overvalued?**

Analyst sentiment regarding CoreWeave’s valuation remains split. Skeptics cite a steep price-to-sales (P/S) ratio of 8.8x as of early December, indicating the market is pricing in roughly $8.80 for every dollar of revenue. By contrast, the S&P 500 Index peaked at a 3.32x P/S ratio in September 2025. This premium reflects CoreWeave’s status as an unprofitable, high-growth enterprise, with investors explicitly pricing anticipated future expansion into the current share price.

Conversely, bullish analysts argue the stock remains undervalued, pointing to its formidable near-term growth trajectory and expansive long-term addressable market. Its valuation is indeed high, but proponents contend it accurately reflects the company’s unprecedented growth runway and long-term strategic positioning.