#分享美股交易赢英伟达股票 Nvidia's Next Wave Could Be Even Larger

#分享美股交易赢英伟达股票 Nvidia's Next Wave Could Be Even Bigger

Nvidia (NVDA) has built an ecosystem around networking, software, inference, storage, and CPUs, enabling it to capture value at different points in the AI technology stack. Additionally, bottlenecks related to power, memory, networking, and photonics continue to expand Nvidia's TAM. With Rubin, Vera, and inference demands still in early stages, future EPS could significantly exceed current expectations.

Financial performance shows Nvidia's growth accelerating
This quarter again proves that Nvidia's growth is far different from any other semiconductor company. Even with a roughly $4.5 billion impact from 🇨🇳H20 restrictions, Nvidia still achieved $44.1 billion in Q1 revenue. However, excluding the H20 restrictions, the company's revenue could have exceeded $49 billion. Considering Rubin and Vera will launch in the coming quarters, and with rising AI infrastructure demand, Nvidia may be on track to surpass $200 billion in annual revenue. This growth is impressive on multiple fronts. Data center revenue is about $75.3 billion, with roughly half, $37.9 billion, coming from hyperscalers, and the remaining $37.4 billion from AI cloud, industrial, and enterprise customers. Nvidia has achieved diversification of demand sources, which is a significant milestone, as the company is not solely reliant on hyperscaler growth.
Another key milestone is the importance reflected by the attach rate in networking, currently around 90%. First, profit margins are outstanding. The company has achieved a 75% margin while growing at such an incredible pace, which is rare among tech companies. But even more noteworthy is that Nvidia has successfully improved operating leverage while making large investments in R&D and infrastructure. As a result, compared to the $35 billion profit margin in the previous period, Nvidia achieved a record-high profit margin of $49 billion in FY2027 Q1. In other words, Nvidia's revenue run rate is about $200 billion annually. Nvidia is entering the role of a top global networking company
Most investors tend to see Nvidia only as a GPU semiconductor manufacturer. However, networking is rapidly becoming crucial, making Nvidia one of the most important companies in the field. Quarterly revenue approaches $15 billion, growing at 200% annually. Today, Nvidia offers more than just computing power; it delivers a comprehensive architecture that enables efficient AI infrastructure operation.
This distinction is critical when evaluating modern AI clusters. Nvidia's GPUs are the main source of performance. However, as AI infrastructure scales to tens of thousands or even hundreds of thousands of GPUs, networks often become the bottleneck. Here, Nvidia's triple fabric architecture plays a role. NVLink facilitates communication between GPUs within a single rack, while Spectrum-X ensures efficient Ethernet-based networking. These tools help build AI infrastructure with fewer bottlenecks. Strategically, this approach is vital for Nvidia. Increasingly, hyperscalers are purchasing complete networks rather than individual components. As Nvidia's network attach rate continues to rise toward 90%, it is becoming the preferred network provider in AI architectures. This creates a powerful ecosystem effect, as each additional layer of integration further enhances Nvidia's platform stickiness. But what excites me most is Nvidia's strong push into co-packaged optics. Investing about $4 billion in Lumentum and Coherent indicates the company understands that optical interconnects could become the next bottleneck in AI infrastructure. Once AI scales to hundreds of thousands of GPUs, power consumption and low latency will become critical issues. Co-packaged optics address these by integrating optical interconnects directly into switch chips. This strategy has the potential to be a game-changer for Nvidia over the next decade. CUDA has made Nvidia a software architecture leader, and networking has become a core part of AI infrastructure. Now, co-packaged optics could become the nervous system of AI infrastructure.
While many focus on innovations like Blackwell and Rubin, they often overlook a key fact: Nvidia is expanding control over all critical elements needed to build efficient AI infrastructure. That’s why Nvidia’s networking business deserves more attention. Nvidia’s valuation reflects market expectations for continued AI infrastructure prosperity
Considering Nvidia's role as a foundational infrastructure provider in AI, its current $212 stock price remains relatively reasonable. Analysts expect Nvidia’s revenue to grow from $391 billion in FY2027 to over $660 billion in FY2029, with EPS rising from $8.94 to $15.64, leading to a lower future P/E ratio.
Based on current forecasts, growth is expected to slow, with FY2029 only growing 20%, compared to 81% in FY2027. However, if inference monetization progresses smoothly, and AI factory projects, sovereign AI, and CPUs perform as expected, Nvidia could surpass analyst estimates and reach its target stock price. This is because Nvidia has gone beyond GPUs, now participating in all other components including networking, software, orchestration, storage, and CPUs. Therefore, Nvidia’s stock can be primarily viewed as dependent on market expectations for sustained AI infrastructure spending. Investors are willing to pay a premium for the leadership position and product ecosystem stickiness. The key question is whether this growth can continue; otherwise, the assumptions supporting Nvidia’s current valuation will need to be revised.
Reassessment risks
The first major risk factor is demand from hyperscalers and enterprise customers. Nvidia’s growth outlook heavily depends on cloud providers, AI-native companies, and enterprise clients increasing their AI-related infrastructure investments. If demand does not accelerate and budget allocations are delayed, inference adoption may fall short of expectations, impacting profit performance. Another risk is execution risk. The company's growth now depends on deploying Rubin, Vera, and networking products. Any delays in production, logistics issues, or slower adoption could threaten its growth prospects. This is critical because hyperscaler AI infrastructure plans are predicated on Nvidia’s growth expectations.

Final conclusion
In the AI hardware space, segments where Nvidia does not dominate are few. Yet, despite an impressive track record, even considering Nvidia’s competitive advantages and market strength, the stock remains undervalued compared to peers. I believe Nvidia’s long-term profitability is significantly undervalued at the current share price. $NVDA