Chips Roar Back: Is the Semiconductor Rally Just Getting Started?
AI's New Narrative Is Here: Why NeoCloud Stocks Are Being Repriced
AI infrastructure stocks staged a sharp rebound after the latest earnings from hyperscale cloud providers (CSPs).
$Amazon (AMZN.US)$ rose 4.6%, while $Microsoft (MSFT.US)$ and $Alphabet-A (GOOGL.US)$ each gained around 5%. The rally quickly spread across the AI infrastructure ecosystem, with $CoreWeave (CRWV.US)$ jumping nearly 20%, $NEBIUS (NBIS.US)$ up about 12%, and $IREN Ltd (IREN.US)$ gaining roughly 8%.
This wasn't simply a relief rally after a selloff. Instead, the market is beginning to reprice AI infrastructure around a new investment thesis: AI spending is finally producing measurable returns.
The Market Is Moving Beyond the Old AI Story
For much of the past year, investors focused on one concern:
AI companies were spending heavily on GPUs and data centers, but when would those investments generate enough revenue to justify the massive capital outlays?
That created the "old narrative":
More GPUs → Higher CapEx → Higher depreciation → Lower free cash flow → Uncertain AI monetization
The latest cloud earnings suggest that this cycle may finally be changing.
Instead, investors are starting to see a more complete commercial loop:
Higher token usage → Faster enterprise AI adoption → Stronger cloud revenue and margins → Higher AI CapEx → More demand for GPUs, networking, data centers, and cloud infrastructure.
In other words, AI demand is no longer being measured solely by infrastructure spending—it's increasingly being validated by revenue growth.
Why NeoCloud Stocks Are Rallying
NeoCloud providers such as CoreWeave and Nebius sit between AI model developers and hyperscale cloud providers, offering dedicated GPU capacity for AI workloads.
Their recent rally reflects rising confidence that AI compute demand is both real and sustainable.
As hyperscalers continue expanding AI infrastructure, investors become more willing to assume:
– Higher GPU utilization
– Stronger cloud leasing demand
– Longer customer contracts
– Better pricing power
– Faster revenue growth
Simply put, the stronger the CSP earnings, the stronger the investment case for NeoCloud providers.
Morgan Stanley: Open Models Could Expand the AI Market Even Further
Another major catalyst behind the recent rerating is the rapid progress of open-weight AI models.
According to Morgan Stanley, better open models could accelerate AI adoption through Jevons Paradox—as AI becomes cheaper, overall usage increases rather than declines.
The bank outlines three possible long-term scenarios.
Scenario 1: Closed Models Win
Large frontier AI labs continue to dominate because enterprises prioritize accuracy, security, and liability protection.
Primary beneficiaries include:
– Security software: $Palo Alto Networks (PANW.US)$ , $CrowdStrike (CRWD.US)$ , $Zscaler (ZS.US)$ , $Okta (OKTA.US)$
$Microsoft (MSFT.US)$ and NeoCloud providers such as $CoreWeave (CRWV.US)$ and $NEBIUS (NBIS.US)$ remain secondary beneficiaries under this scenario.
Scenario 2: A Hybrid AI Ecosystem
Morgan Stanley considers this the middle path.
Closed models handle complex, mission-critical tasks, while open models serve cost-sensitive, high-volume inference workloads.
As AI workloads become more diverse, software that manages, monitors, and secures AI infrastructure becomes increasingly valuable.
Key beneficiaries include:
– Cloud providers: Amazon, Alphabet, Microsoft
– Security software: $Palo Alto Networks (PANW.US)$ , $CrowdStrike (CRWD.US)$ , $Fortinet (FTNT.US)$ , $Zscaler (ZS.US)$ , $Okta (OKTA.US)$
– Model providers: $MINIMAX-W (00100.HK)$ , $Z.AI (02513.HK)$ , $Alibaba (BABA.US)$ , $Tencent (TCEHY.US)$
Scenario 3: Open Models Win
If open-weight models reach frontier-level performance while becoming dramatically cheaper, AI deployment could become far more decentralized.
Instead of relying primarily on hyperscale cloud platforms, more inference workloads could move to enterprise data centers and edge devices.
Major beneficiaries include:
– Cloud providers: Microsoft
– Model providers: $MINIMAX-W (00100.HK)$ , $Z.AI (02513.HK)$ , $Alibaba (BABA.US)$ , $Tencent (TCEHY.US)$
– Security software: $Palo Alto Networks (PANW.US)$ , $CrowdStrike (CRWD.US)$ , $Fortinet (FTNT.US)$ , $Zscaler (ZS.US)$ , $Okta (OKTA.US)$
– On-prem infrastructure: $Dell Technologies (DELL.US)$ , $Hewlett Packard Enterprise (HPE.US)$ , $NetApp (NTAP.US)$
The Common Winner Across Every Scenario
While the competitive landscape may evolve, one conclusion remains remarkably consistent across Morgan Stanley's framework.
AI compute demand continues to grow in every scenario.
Whether AI is dominated by closed models, open models, or a hybrid ecosystem, more inference means more computing power, networking capacity, and electricity.
That is why Nvidia appears as a primary beneficiary in every scenario, while companies supplying on-site power infrastructure—including Bloom Energy, INIO, Solaris Energy Infrastructure, WMB, and Liberty Energy —also remain long-term winners.
The relative winners among hyperscalers may change depending on market share, but the overall infrastructure buildout continues.
Bottom Line
The latest CSP earnings suggest that AI investing is entering a new phase.
The market is no longer focused solely on how much companies are spending on AI, but on whether AI is generating measurable revenue and cash flow.
At the same time, the rapid rise of open-weight models could expand AI adoption even further, reinforcing long-term demand for compute infrastructure.
For investors, the question is shifting from "Will AI demand continue?" to "Which companies are best positioned to capture the next wave of AI monetization?"
Disclaimer: Moomoo Technologies Inc. is providing this content for information and educational use only.Read more
Comments
to post a comment
56
70