Author:律动BlockBeats
Why did NeoCloud experience the largest gain among this round of US tech stocks rebound?
One of the strongest sectors in this round of U.S. tech stock rebound is NeoCloud: CoreWeave, Nebius, and select AI infrastructure companies with power and data center resources.
Logically, capital is pricing an AI infrastructure equity instrument with multiple leverage: computing power capacity that has been contractually locked in and can be delivered quickly.
Once AI demand is revised upward, NeoCloud’s revenue expectations, financing capabilities, and shareholder equity value could all rise simultaneously. This gives it strong upside potential during the tech stock rebound phase; electricity, data centers, financing, and valuation elasticity together form this leverage.
The bottleneck for AI is shifting. Initially, the most scarce resource was GPUs, followed by HBM and high-speed networks; today, what customers truly lack is an end-to-end capability to go live: securing GPUs, sufficient power, completing data center construction, network interconnection, and delivering large-scale clusters within months.
NeoCloud is precisely positioned at this gap.
The funds are purchasing a "powered-up computing power facility"
NeoCloud's offerings typically include GPU clusters, networking, liquid cooling, data centers, power access, and operations 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 prefer to maintain greater flexibility and avoid relying entirely on a single hyperscaler.
Thus, NeoCloud, with its existing electricity and rapid deployment capabilities, has become an "accelerator" for AI infrastructure investment.
The market is willing to assign them higher valuations 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: Buying 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 more positive answer.
CoreWeave's Q2 revenue reached $2.575 billion, with a disclosed backlog (signed but unconfirmed expected revenue) of approximately $104 billion; Nebius's AI Cloud ARR (annualized recurring revenue) reached $3 billion, along with the disclosure of several 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 customers sign long-term capacity contracts → Some customers provide advance payments or minimum payment commitments → The company finds it easier to secure debt and equipment financing → New GPU, data center, and power capacity come online → Revenue and EBITDA (earnings before interest, taxes, depreciation, and amortization) grow → Financing capability and expansion capacity continue to improve
This shifts NeoCloud’s narrative from being a high Capex GPU lessor to an AI infrastructure operator expanding with order-backed growth.
Growth exhibits a clear 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 the expected differences across various stages.
Storage leaders benefit from strong AI demand, and the outlook for products such as HBM and DRAM remains robust. 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 in. Strong earnings may be offset by stock pressure if long-term guidance fails to be further upgraded.
The challenge for storage companies lies in their cyclical nature. The market trades 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, customer bases, and technological capabilities, and are the primary beneficiaries of AI investment. Their AI businesses are diluted by large revenue bases from advertising, enterprise software, e-commerce, and consumer services; additional AI capital expenditures take longer to translate into improved margins for the overall company. For capital seeking flexibility, a single large NeoCloud contract often has a greater marginal impact on revenue and valuation than an equivalent-sized order for any 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 the next round of funding and scaling. Capital tends to 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 understanding 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 upfront 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 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; 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 clouds, yet its proportion of fixed assets and debt on the balance sheet is higher. If a large contract simultaneously raises revenue expectations, utilization rates, and access to financing, the market’s reassessment of shareholder equity value can be sharply steep. The rapid stock price increase following earnings reports often results from the combined effect of upward revisions to profit expectations and valuation multiples.
Three-tier leverage creates a positive feedback loop during the upward phase:
Larger long-term contracts → Easier access to financing and increased capacity → Higher utilization and operating profits → Enhanced equity value and financing capability → More contracts and next-phase expansion opportunities
The same mechanism also amplifies downside risk. If customers delay, utilization declines, GPU or power delivery is delayed, or debt costs rise, fixed costs and financing obligations will compress shareholder returns. Therefore, the market’s high-elasticity pricing for NeoCloud reflects its high execution requirements.
Order visibility is at the core of this reassessment.
The most attractive aspect of NeoCloud is the visibility of revenue.
If a customer signs a take-or-pay contract, they are still obligated to make minimum payments even if actual usage fluctuates in the short term. For operators, this type of revenue is easier to predict; for creditors, these contracts also enhance the feasibility of asset financing.
Therefore, the market will continue to track several indicators:
· Term, enforceability, and client credit of signed contracts;
· 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 is energized, installed, and begins billing contributes to revenue and cash flow.
This is also a reassessment of power assets.
The most valuable insight about NeoCloud in the community is shifting the focus from the number of GPUs to Power (electrical capacity).
GPU supply will expand with purchases by NVIDIA, AMD, and cloud providers; however, the development of high-quality power capacity is slower. It involves the 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 can enter this mainline: they already possess some power resources, land, and infrastructure, and only need to shift their assets from mining loads to AI loads. Of course, a resource base does not equate to business success—ultimate success still depends on customers, financing, and delivery capabilities.
NeoCloud led the rally in tech stocks, driven by a realignment of the market’s focus on the AI infrastructure value chain.
The asset that capital currently values most is computing capacity that combines GPUs, electricity, data centers, and long-term customer contracts, and can deliver rapidly. It captures AI capital expenditures while possessing more contract-based characteristics than standalone chips and components, offering high growth elasticity along with a premium from infrastructure scarcity.
Next, whether NeoCloud can continue to outperform depends on a straightforward question: Can these massive orders be turned into powered-up clusters, confirmed revenue, and cash flow that covers capital costs on time?
Twitter: https://twitter.com/BitpushNewsCN
BitPush Telegram community: https://t.me/BitPushCommunity
BitPush TG subscription: https://t.me/bitpush