NVDA has increased over 10 times in three years: Why the AI infrastructure cycle continues to drive NVIDIA's rise

Since 2026, NVDA has once again become one of the most dominant core assets in the Nasdaq market. As AI data center expansion continues to advance, the Blackwell cycle is fully launched, and the demand for AI Agents and reasoning re-accelerates, global technology funds have refocused on the AI mainline after a period of high-level volatility, with Nvidia remaining at the core of the entire AI rally and U.S. stock risk appetite.

Over the past three years, one of the biggest changes in the U.S. stock market is not the AI concept itself, but the global capital beginning to re-understand the significance of "AI infrastructure."

If in 2023 the market was trading on the hot topic of generative AI, by 2026, what the market is truly trading is: how much capital expenditure global tech companies are willing to invest in AI computing power over the next few years. NVDA has risen from its late 2022 lows to its current highs, with a total increase of over tenfold, fundamentally not just because of GPU sales growth, but because Nvidia is gradually becoming the most core beneficiary asset of the global AI infrastructure cycle.

The so-called "AI infrastructure cycle" in the current market essentially refers to the ongoing increase in data center, AI computing power, inference network, and GPU cluster investments by global tech companies, with Nvidia occupying the most central position in the entire industry chain.

From the current weekly chart, NVDA’s long-term trend has not been truly broken. Even after experiencing multiple rounds of high-level volatility and a Nasdaq tech stock correction over the past year, funds still flow back into Nvidia when AI sentiment heats up again. Compared to traditional semiconductor companies, NVDA is increasingly resembling a kind of "AI liquidity anchor asset."

Why has NVDA risen over tenfold in the past three years

The real starting point of this round of NVDA’s rise actually stems from a fundamental change in the AI market structure.

Before the explosion of generative AI, the valuation logic for the semiconductor industry was more centered around consumer electronics, PC cycles, and traditional cloud computing demand. But after ChatGPT was launched, the global tech industry quickly entered an AI arms race stage, with giants like Microsoft, Meta, Google, and Amazon continuously increasing their AI data center budgets, and GPUs rapidly becoming the most scarce and critical resource in the entire AI industry chain.

The market gradually realized that the core of large model competition is not "whose model parameters are larger," but "who can sustain enough computing power support."

This directly changed Nvidia’s market positioning.

In the past, the market was more willing to see Nvidia as a high-growth chip company, but now, more and more institutions are starting to see it as an infrastructure gateway for the AI era. The long-term advantages of GPUs, CUDA, AI network systems, and data center ecosystems mean Nvidia can benefit not only from AI training demand but also from the growth of inference markets in the coming years.

This change is distinctly different from traditional tech stocks.

Ordinary tech companies rely more on product cycles and user growth, whereas NVDA’s current rise logic is increasingly aligned with infrastructure assets. The market’s focus is no longer just on quarterly profits but on whether global AI capital expenditure can continue to expand over the next few years.

Because of this, Nvidia has become one of the strongest trend assets in the Nasdaq market over the past three years.

Why do AI data center expansion and the Blackwell cycle continue to raise market expectations

Since 2025, the market’s focus on AI has clearly shifted.

Previously, the market paid more attention to large model capabilities and generative AI concepts, but now, funds are more concerned about how long AI data centers can continue to expand and whether AI computing power demand will enter a long-term growth phase.

More and more tech companies are realizing that the core of AI competition is gradually shifting from "model release" to "infrastructure reserves."

Whoever owns larger GPU clusters, more AI data centers, and more stable inference capabilities will be able to take the lead in the next AI competition.

The Blackwell cycle is precisely under this context that market expectations are being reinforced.

Compared to the Hopper architecture, Blackwell further improves inference efficiency, training performance, and energy consumption control, while better adapting to AI Agents and long-term inference needs. As more AI products enter commercialization, market expectations for inference-side GPUs are also continuously rising.

Previously, the market was worried that AI GPU demand might only be a temporary burst, but current trends show that AI data center expansion is increasingly aligning with long-term infrastructure development logic.

Microsoft, Meta, and Amazon continue to increase AI CapEx, and more countries are promoting Sovereign AI, with local AI data center construction demands rising steadily. AI computing power is even gradually becoming a new global strategic resource.

The market has therefore revised upward Nvidia’s long-term growth potential.

Because once AI competition enters the infrastructure stage, Nvidia’s importance is further elevated.

How AI Agents and inference demand are changing Nvidia’s growth logic

One of the most obvious changes in the AI market in 2026 is that AI Agents are once again entering the center of market attention.

From OpenAI to various automation AI platforms, more and more companies are pushing for Agent commercialization. Compared to the previous focus on chatbots and generative AI, AI Agents emphasize continuous reasoning, autonomous execution, and long-term operation capabilities, which are significantly increasing the demand for inference-side GPUs.

In the past, the market mainly traded "training demand," but now, funds are refocusing on "inference demand."

Training markets are more like one-time capital investments, while inference markets are closer to long-term continuous consumption. As AI Agents, large model searches, automated offices, AI programming, and robotics gradually commercialize, GPUs are no longer just training tools but are becoming the infrastructure of the entire digital economy.

Nvidia’s growth logic is thus changing.

The market is no longer solely focused on GPU sales volume but is beginning to pay attention to:

• Whether AI inference demand will grow long-term;
• Whether AI Agents will continue to expand;
• Whether enterprise AI will enter large-scale deployment;
• Whether competition in AI cloud services will intensify further.

These variables jointly determine NVDA’s long-term valuation space.

Meanwhile, Nvidia is also gradually shifting from a chip company to an AI platform company. CUDA, AI network systems, supercomputing, and inference ecosystems are forming an increasingly complete AI infrastructure closed loop.

This is also why the market assigns a valuation to NVDA far exceeding that of traditional semiconductor companies.

Because what is being traded now is not just GPUs, but the entire AI infrastructure ecosystem.

Why global tech capital expenditure continues to flow into NVDA

The most core support for NVDA currently still comes from the fact that global tech capital expenditure has not shown a clear slowdown.

In the past, during the mobile internet era, tech companies competed mainly on user scale and advertising revenue; in the AI era, the focus has gradually shifted to data centers, computing power reserves, and model capabilities.

AI computing power has thus begun to become a new strategic resource.

Microsoft’s partnership with OpenAI, Meta’s continuous expansion of AI data centers, Google’s strengthening of the Gemini infrastructure, and Amazon’s ongoing increase in AI cloud investments all indicate that large tech companies have entered a new round of AI arms race.

This competitive logic is very similar to the previous cloud computing cycle.

The difference is that AI data centers depend far more on GPUs than traditional cloud computing, and Nvidia is at the most central position in the entire AI computing power system.

The so-called "macro trend of tech stocks" in the current market essentially refers to the valuation of large tech companies increasingly influenced by interest rates, Federal Reserve policies, dollar liquidity, capital expenditure, and risk appetite, with NVDA already being one of the most core beneficiaries of this macro capital flow.

Because of this, even during high-level volatility, the market remains willing to continue betting on NVDA.

This is because the funds truly believe that it’s not just the short-term AI hype, but that the AI infrastructure expansion cycle will continue for years to come.

Why funds are flowing back into the AI mainline after NVDA’s high-level consolidation

Looking at the trend, NVDA has actually entered multiple high-level consolidation phases over the past year.

High valuations, discussions of AI bubbles, and Nasdaq tech stock volatility once suppressed market risk appetite. But whenever the market returns to the AI mainline, funds still prioritize flowing into Nvidia.

The reason is very straightforward.

While the pace of AI application commercialization may still have uncertainties, the expansion of AI data centers, GPU procurement, and AI computing power competition are real and ongoing.

Therefore, during periods of risk appetite recovery, funds prefer the most certain, most liquid, and most industry-central AI assets, and NVDA precisely meets these conditions.

At the same time, the U.S. stock AI mainline is also being re-strengthened.

After AI Agents, robotics, autonomous driving, edge AI, and inference demand heat up again, the market begins to re-establish expectations for long-term AI growth, and Nvidia, covering GPUs, data centers, AI networks, and inference ecosystems, remains at the most core position in the entire industry chain.

This is also why NVDA can still reach new highs after a period of high-level consolidation.

Is NVDA shifting from a tech stock to a core asset of global AI infrastructure

Looking back at the past three years of market changes, it’s clear that Nvidia’s market positioning has undergone a significant transformation.

In 2023, the market was trading on AI hotspots.

In 2024, the focus shifted to AI data centers.

By 2026, the market is beginning to trade on the era of global AI infrastructure.

This shift indicates that Nvidia is gradually shedding the traditional tech stock logic.

Ordinary tech companies rely more on product cycles, while infrastructure assets depend more on long-term capital expenditure and industry expansion cycles. Today, when discussing Nvidia, the focus is increasingly on whether the global AI computing power demand will continue to grow over the next few years, rather than just quarterly GPU shipments.

Of course, risks still exist.

If future AI capital expenditure slows significantly or AI commercialization yields lower-than-expected returns, high-valuation AI assets could still experience sharp volatility. But from the current market structure, global funds remain willing to continue betting on the long-term expansion of AI, and Nvidia remains one of the most core beneficiaries of this AI infrastructure cycle.

Nvidia is no longer just an AI chip company.

It is gradually becoming a core infrastructure asset of the global AI era.

FAQ

Why has NVDA risen over 10 times in three years?

NVDA’s over tenfold increase in three years is mainly due to AI data center expansion, surging GPU demand, and continuous increases in AI CapEx by global tech companies.

Why does Blackwell continue to influence NVDA’s market expectations?

Blackwell is Nvidia’s new generation AI GPU architecture, with improved inference and training performance, reinforcing market expectations for long-term AI infrastructure expansion.

Why is AI Agent beneficial for NVDA?

AI Agents will continuously increase inference computing power demand, and NVDA remains one of the most central suppliers of AI inference GPUs globally.

What is the biggest risk for NVDA currently?

The biggest risk for NVDA is a slowdown in future AI capital expenditure and market overestimation of AI’s long-term growth, leading to valuation volatility.

Why do global tech companies continue to increase AI CapEx?

Microsoft, Meta, Google, and Amazon are competing around AI data centers and model capabilities, making AI computing power investment a strategic-level capital expenditure.

Has NVDA shifted from a tech stock to a core AI infrastructure asset?

NVDA is gradually shifting from a traditional tech stock to a core global AI infrastructure asset, as the market increasingly focuses on its long-term position in AI data centers and inference ecosystems.