One of the strongest directions in this round of U.S. tech stock rebound comes from NeoCloud: CoreWeave, Nebius, and certain AI infrastructure companies with access to power and data center resources.

Logically, capital is pricing an AI infrastructure equity instrument with multiple leverage layers: computing capacity that has been contractually locked in and can be delivered quickly.

Once AI demand is revised upward, NeoCloud’s revenue expectations, fundraising capacity, and shareholder equity value could all rise simultaneously. This gives it strong upside potential during a tech stock rebound; electricity, data center capacity, fundraising, and valuation flexibility together form this leverage.

The bottleneck for AI is shifting. Initially, the most scarce resource was GPUs; then it became HBM and high-speed networks. Today, what customers truly lack is the full capability to go live: securing GPUs, having sufficient power, completing data center construction, establishing network connectivity, and delivering large-scale clusters within months.

NeoCloud is precisely positioned at this gap.

The funds are being used to purchase an "already powered-up computing power facility."

NeoCloud’s offerings typically include GPU clusters, networking, liquid cooling, data centers, power access, and operational services. Customers purchase a large-scale computing capacity that can run AI training and inference directly.

This is crucial. While GPUs can be procured, power capacity, land, substations, data center permits, and network access cannot be replicated in the short term. Large cloud providers, despite having capital and customers, are also constrained by construction timelines; some AI companies seek greater flexibility and prefer not to rely entirely on a single hyperscaler.

Thus, NeoCloud, with its ready-to-use electricity and rapid deployment capabilities, has become an “accelerator” for AI infrastructure investment.

The market is willing to assign them higher valuations, primarily because these resources have two key characteristics:

· Scarcity: Limited available electricity and deliverable data center capacity;

· Contractible: Customers are willing to sign multi-year capacity contracts with minimum commitments.

When scarce resources can be locked in through long-term contracts, the market reclassifies them from ordinary IT service revenue as cash-flow assets with infrastructure characteristics.

The earnings report changed the market's perception of the business model.

Previously, the main question from the market about NeoCloud was straightforward: Purchasing GPUs and building data centers require massive CapEx—will the company fall into a cycle of continuous fundraising and continuous cash burn?

The recent earnings report provided a somewhat positive answer.

CoreWeave's Q2 revenue reached $2.575 billion, with a disclosed backlog of approximately $104 billion in signed but unearned revenue; Nebius's AI Cloud ARR (annualized recurring revenue) reached $3 billion, alongside the disclosure of multiple large, long-term contracts. The market is not only focused on quarterly revenue, but more importantly on the complete business闭环 that these figures reveal:

AI customer signs long-term capacity contract
Some customers provide advance payments or minimum payment commitments.
Companies find it easier to obtain debt and equipment financing.
→ GPU, data center, and power capacity now available
→ Growth in revenue and EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization)
Funding capacity and scalability continue to improve

This shifts NeoCloud’s narrative from a “high CapEx GPU lessor” to an “AI infrastructure operator expanding with order-backed growth.”

Growth exhibits a distinct flywheel effect as long as orders, financing, and delivery can be continuously aligned.

Why didn't the funds prioritize storage and the three major clouds?

The choice of funds reflects expected differences across various stages.

Leading memory manufacturers are benefiting from strong AI demand, with high demand for products such as HBM and DRAM. However, the market is beginning to worry about supply ramp-up, elevated prices, peak profit margins, and whether earlier optimistic expectations have already been fully priced into stock valuations. Strong earnings may be offset by stock pressure if long-term guidance does not continue to be raised.

The challenge for storage companies lies in their cyclical nature. The market trades based on the projected price, shipment volume, and gross margin trajectory over the next several quarters; strong current performance struggles to sustain valuation expansion when supply may catch up with demand and average selling prices could decline. HBM/DRAM, NAND/SSD, and HDD each belong to different sub-cycles, and the stock performance of all storage companies cannot be attributed to a single cause.

The three major clouds—Microsoft Azure, Amazon AWS, and Google Cloud—have stronger cash flows, larger customer bases, and greater technological capabilities, and are the primary beneficiaries of AI investment. Their AI businesses are diluted by massive revenue bases from advertising, enterprise software, e-commerce, and consumer services; additional AI capital expenditures take longer to translate into improved profit margins across the entire conglomerate. For capital seeking flexibility, a single large NeoCloud contract often has a greater marginal impact on revenue and valuation than an equivalent-sized order would on the overall valuation of the three major clouds.

NeoCloud sits between the two: it has a lower revenue base, a pure AI exposure, rapid order growth, and each new long-term contract directly supports its next round of funding and scaling. Investors easily view it as a highly elastic AI infrastructure play.

The current market trading logic can be summarized as:

NeoCloud is essentially an AI infrastructure lever.

Understanding NeoCloud’s leadership hinges on grasping its leverage effect. Buying shares in such companies is essentially holding an equity asset highly sensitive to AI computing demand, the price of deliverable capacity, and the financing environment. This leverage encompasses three layers of meaning.

First is operating leverage. The initial investments in GPUs, data centers, power connectivity, networking, and operations are high, and many costs become relatively fixed after capacity comes online. As utilization of already-enabled clusters increases and pricing per unit of capacity improves, additional revenue can be converted into profit at a rapid rate, resulting in clear marginal improvements in profitability.

Second is financing leverage. Long-term contracts, take-or-pay commitments, and customer prepayments enhance the project’s appeal to lenders and equipment financiers. This allows the company to use a portion of its equity capital to leverage larger investments in GPUs, data centers, and power infrastructure; once the new capacity begins generating revenue, it can support the next phase of development.

Third is equity leverage. NeoCloud’s revenue base and market capitalization are typically smaller than those of the three major cloud providers, yet its proportion of fixed assets and debt on the balance sheet is higher. When a large contract simultaneously boosts revenue expectations, utilization rates, and access to financing, the market’s reassessment of shareholder equity value can be sharply upward. The rapid stock price increase following earnings reports often results from the combined effect of upward revisions in profit expectations and valuation multiples.

Three layers of leverage created a positive feedback loop during the upward phase:

Larger long-term contracts
→ Easier access to financing and increased capacity
→ Higher utilization and operating profit
→ Increased equity value and financing capability
→ Earn more contracts and next-round expansion opportunities

The same mechanism also amplifies downside risk. If customers delay payments, utilization declines, GPU or power deliveries are delayed, or debt costs rise, fixed costs and financing obligations will compress shareholder returns. As a result, the market’s high valuation of NeoCloud reflects its high execution requirements.

Order visibility is at the core of this revaluation.

The most attractive aspect of NeoCloud is the visibility of its revenue.

If a customer signs a take-or-pay contract, they remain obligated to make minimum payments even if actual usage fluctuates in the short term. For operators, this type of revenue is easier to forecast; for creditors, such contracts enhance the feasibility of asset financing.

Therefore, the market will continue to track several indicators:

· Contract duration, enforceability, and client creditworthiness;

· The difference between enabled MW and contracted MW;

· Revenue per MW and Capex per MW;

· The proportion and payment schedule of the customer's advance payment;

· Utilization rate, renewal rate, and customer concentration;

Debt interest rates, debt maturity, and subsequent financing capacity.

Among these, "activated capacity" is particularly critical. The contracted MW represents demand, but only the MW that has been energized, installed, and is being billed contributes to revenue and cash flow.

This is also a reassessment of power assets.

The most valuable insight about NeoCloud in the community is to shift the focus from the number of GPUs to Power (electrical capacity).

GPU supply will expand with purchases from NVIDIA, AMD, and cloud providers; however, the development of high-quality power capacity progresses more slowly. It involves the power grid, substations, land, permits, data center construction, and regional network conditions.

Whoever can secure sufficient electricity sooner can convert their GPUs into sellable computing power sooner.

This is also why some companies that transitioned from Bitcoin mining operations have entered this main storyline: they already possess partial power resources, land, and infrastructure, and only need to shift their assets from mining loads to AI loads. Of course, a strong resource base does not guarantee business success—ultimate success still depends on customers, financing, and delivery capabilities.

NeoCloud led the rally in tech stocks, driven by a market realignment of the AI infrastructure value chain.

The asset that capital currently values most is computing capacity that integrates GPUs, electricity, data centers, and long-term customer contracts, and can deliver quickly. It captures AI capital expenditures while offering stronger contract-based characteristics than standalone chips and components, combining high growth potential with a premium from infrastructure scarcity.

Next, whether NeoCloud can continue to outperform depends on a simple question: Can these massive orders be turned into powered-up clusters, confirmed revenue, and cash flow that covers capital costs on time?