---
title: "From NVIDIA's Rubin leading the \"AI Factory Era\" to AMD's MI500 \"Thousandfold Roadmap\": The \"AI Bull Market Narrative\" continues to dominate the stock market"
type: "News"
locale: "zh-CN"
url: "https://longbridge.com/zh-CN/news/271622738.md"
description: "NVIDIA and AMD launched their respective AI computing platforms at the CES exhibition, NVIDIA's Vera Rubin and AMD's next-generation AI GPU. Both CEOs emphasized that AI computing resources remain scarce, alleviating market concerns about an AI bubble. Jensen Huang stated that the Vera Rubin platform has been fully put into production, while AMD previewed that the MI500 will achieve a thousandfold performance leap, demonstrating the continued growth in demand for AI computing power"
datetime: "2026-01-06T11:18:30.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/271622738.md)
  - [en](https://longbridge.com/en/news/271622738.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/271622738.md)
---

> 支持的语言: [English](https://longbridge.com/en/news/271622738.md) | [繁體中文](https://longbridge.com/zh-HK/news/271622738.md)


# From NVIDIA's Rubin leading the "AI Factory Era" to AMD's MI500 "Thousandfold Roadmap": The "AI Bull Market Narrative" continues to dominate the stock market

According to Zhitong Finance APP, NVIDIA's long-time biggest competitor in the AI data center and PC chip field, AMD (AMD.US), announced shortly after NVIDIA CEO Jensen Huang unveiled the next-generation AI GPU computing infrastructure platform named Vera Rubin at CES (Consumer Electronics Show), that it will launch a strong competitor to the Vera Rubin platform aimed at enterprise-level ultra-large-scale data centers—AMD's next-generation AI GPU computing platform. During this CES event, AMD CEO Lisa Su emphasized the exclusive energy efficiency and cost-performance advantages of its next-generation products for the data center market, aiming to further shake the monopolistic market position of NVIDIA (NVDA.US), known as the "superpower of AI chips," in the AI computing infrastructure market.

At the highly anticipated CES opening event in early 2026, in the face of global capital market concerns about the "impending burst of the AI bubble" and anxiety over the gradual failure of Moore's Law, **Jensen Huang presented a new cabinet-level AI GPU computing infrastructure platform named Vera Rubin, while AMD CEO Lisa Su brought the MI455X AI GPU, which boasts a tenfold performance improvement over the MI355X, and previewed the MI500 with a "maximum thousandfold performance leap," collectively sending a positive signal to global investors: the demand for AI computing continues to show exponential expansion, and the construction of AI infrastructure is still in the early to mid-stage.**

In his speech, Jensen Huang stated that the Vera Rubin computing platform has fully entered production, **with all six core chips of the next-generation AI computing infrastructure platform Vera Rubin having completed manufacturing and key testing processes, now entering full production phase, emphasizing that the global AI computing arms race has entered the "cabinet-level/platform-level AI factory era" (for example, the six-in-one Rubin platform emphasizes system synergy and cost/performance leap).** Under the physical limit where transistors have only increased by 1.6 times, NVIDIA has forcibly achieved a fivefold improvement in AI inference performance and a 3.5-fold improvement in training performance for Vera Rubin compared to the already powerful Blackwell through "extreme collaborative design." Jensen Huang stated that Microsoft's next-generation "AI super factory" is expected to deploy hundreds of thousands of chips based on the Vera Rubin platform.

![1767686776(1).png](https://imageproxy.pbkrs.com/https://img.zhitongcaijing.com/image/20260106/1767686778837365.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg)

In addition, in response to market concerns and dissatisfaction regarding the high acquisition costs of NVIDIA's AI computing infrastructure, **the next-generation AI computing platform Vera Rubin will significantly reduce the token generation cost for AI inference to just 1/10 of the Blackwell architecture AI GPU cluster, making the long-costly Agentic AI (i.e., AI agents with agency functions) commercially viable.**On the Vera Rubin platform, NVIDIA utilizes a memory storage platform for inference context built on the BlueField-4 DPU, adding 16TB of high-speed shared memory to each AI GPU device out of thin air, greatly addressing the long text "memory wall" issue.

AMD CEO Lisa Su stated during her keynote speech at the CES opening event that since the rise of ChatGPT globally, the number of active users using AI has increased from 1 million to 1 billion, a milestone that took the internet decades to achieve. It is expected that by 2030, the number of active users using AI will reach 5 billion. **In order to make AI ubiquitous, the demand for AI computing power will continue to experience explosive growth in the coming years, and it is likely that the world's AI computing capacity will need to be increased by about 100 times.**

**At CES, AMD launched the MI440X and MI455X AI GPUs for enterprise-level large data centers and showcased the high-end Helios cabinet-level system. AMD even previewed that the performance of the MI500 series AI GPU products to be launched in 2027 will reach 1000 times that of the flagship products in 2023.** CEO Lisa Su appeared alongside OpenAI co-founder, emphasizing that the global demand for AI computing power is far from being met and aims to further break NVIDIA's absolute monopoly in the AI computing infrastructure market, competing for AI computing orders worth tens of billions or even hundreds of billions of dollars.

## From a single GPU to a rack-scale AI system project, global AI computing power continues to show a pattern where supply falls far short of demand

From these two nearly "same-day confrontations" at CES, it is clear that AI has entered the second stage, constrained by computing power, where the focus of industry competition has shifted from "single GPU parameters" to "rack-scale AI system engineering," which includes the overall synergy of computing (high-speed interconnection of GPU/CPU), interconnection (NVLink/high-performance Ethernet facilities), data processing (DPU/NIC), software stack, and security and trust capabilities.

On one hand, AMD is using the MI440X to penetrate a broader enterprise on-premises data center scenario, emphasizing the deployment of training/fine-tuning/inference in existing data centers in a compact form while achieving data retention and compliance; **on the other hand, it bundles the MI455X with the "Helios" cabinet system and the next-generation Venice CPU, leveraging the endorsement of OpenAI executives to strengthen the visibility of the demand for "the AI computing gap is still widening," and extends the product roadmap to the MI500 series in 2027, providing a long-term upward imagination space for a "maximum thousand-fold performance leap" compared to the MI300.**

More notably, **Jensen Huang directly defined "Vera Rubin" at CES as the next-generation AI computing infrastructure platform: through the extreme synergy of six core chips and network/security components, transforming the AI factory from "stacking GPUs" into a scalable, standardized rack-level AI system, fundamentally revolutionizing the basic concept of the "AI factory," thereby continuing to make significant improvements in the unit cost of training and inference (especially in MoE and long-context inference scenarios' token costs), and announcing that the Rubin AI computing platform has entered mass production.**It will gradually land through partners in the second half of 2026.\*\*

Jensen Huang previously revealed at the Washington GTC conference at the end of October that NVIDIA has up to $500 billion in AI GPU computing infrastructure orders for the entire calendar year of 2025 to 2026. These orders correspond to the company's Blackwell architecture, which is at the core of the current global wave of artificial intelligence deployment, as well as the Rubin architecture AI GPU computing clusters to be launched in 2026.

Some Wall Street analysts have stated that this competition of "roadmap density + systematic delivery" is essentially a supply-side response to the common demand for "endless AI computing power needs" from leading buyers such as Google, Microsoft, Meta, and OpenAI: **As the scale of models, inference links, and multi-modal/agentic AI workloads drive exponential growth in computing power consumption, the capital expenditure focus of tech giants tends to concentrate on AI computing infrastructure. Global investors continue to anchor the "AI bull market narrative" around the new iterations of NVIDIA and AMD's products and the delivery of AI computing clusters as one of the most certain investment narratives in the global stock market. This also means that investment themes closely related to AI training/inference, such as electricity, liquid cooling systems, and optical interconnect supply chains, will continue to rank among the hottest investment camps in the stock market alongside leaders in AI computing like NVIDIA, AMD, Broadcom, TSMC, and Micron.**

After Google launched the Gemini3 AI application ecosystem in late November, this cutting-edge AI application software quickly became popular worldwide, driving an instant surge in Google's AI computing power demand. **The Gemini3 series products brought an enormous AI token processing volume upon release, forcing Google to significantly reduce the free access volume for Gemini 3 Pro and Nano Banana Pro, and temporarily limit access for Pro subscribers. Coupled with recent trade export data from South Korea showing that demand for SK Hynix and Samsung Electronics' HBM storage systems and enterprise-level SSDs remains strong, this further validates Wall Street's assertion that "the AI boom is still in the early construction stage of supply not meeting demand for computing power infrastructure."**

## After the heavy release of MI440X and MI455X, is "King of the Hills" AMD's stock price stepping into a new round of growth?

It is understood that AMD is adding a new model—MI440X—to its existing AI GPU product line, primarily for smaller enterprise-level data centers, allowing customers to deploy hardware on a large scale locally and keep data within their own AI computing infrastructure. AMD CEO Lisa Su also highlighted AMD's latest flagship AI GPU product, MI455X, stating that the cabinet-level AI computing clusters based on this AI chip have achieved a leap in energy efficiency in terms of training/inference capabilities.

**Lisa Su has also joined the "chorus" of executives in the U.S. AI technology sector (including her counterparts at NVIDIA), advocating that the unprecedented wave of global AI deployment will continue strongly, as the productivity transformation potential brought by AI and the demand for massive AI computing infrastructure are far from over. The so-called "AI bubble theory" is merely a concern of the financial market.**\*\*

"We do not have enough computing infrastructure to support all the transformations we are planning," said Lisa Su. "The speed and progress of AI technology innovation in the past few years have been incredible. We are just getting started."

AMD is widely regarded as the closest competitor to NVIDIA in the semiconductor segment for AI chips used to create and run artificial intelligence applications or AI agents. In recent years, AMD has built a brand new AI computing infrastructure business worth billions of dollars around its proprietary AI chips, driving significant revenue and profit growth in recent quarters. Wall Street institutional investors who have driven up its stock price generally hope that AMD can demonstrate more aggressive progress in capturing 80%-90% market share in the AI computing infrastructure space currently dominated by NVIDIA.

![1767686850(1).png](https://imageproxy.pbkrs.com/https://img.zhitongcaijing.com/image/20260106/1767686861253795.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg)

The AMD Helios AI cabinet system based on the MI455X AI GPU and the new Venice data center-level central processing unit design will be launched later this year. OpenAI co-founder Greg Brockman joined Lisa Su on stage at CES in Las Vegas to elaborate on their long-term partnership with AMD and discuss future plans for large-scale deployment of their AI computing systems. Both parties talked about their shared views and beliefs: future global economic growth will be closely linked to the availability of AI computing resources.

**The newly launched MI440X AI GPU will be compatible with compact computers in existing small data centers. Lisa Su also previewed the MI500 series data center processors set to launch in 2027. She stated that this series will offer up to 1000 times the performance of the MI300 series (which was first launched in 2023).**

AMD's stock price has surged over 80% since 2025, but most of that increase occurred in October. **The main catalyst for AMD's strong rally was the procurement agreement for a 1-gigawatt AI chip computing cluster with Saudi Arabia's "Sovereign AI System" Humain, as well as a significant AI computing infrastructure cooperation agreement worth billions of dollars with AI application leader OpenAI.** These large-scale collaborations not only validate the strong capabilities of AMD's AI computing infrastructure technology but also make Wall Street investment institutions like Citigroup more optimistic about its future financial prospects.

![1767686949(1).png](https://imageproxy.pbkrs.com/https://img.zhitongcaijing.com/image/20260106/1767686987233237.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg)Wall Street financial giant Citigroup hailed AMD as the "King of the Hill," stating, "Considering the higher level of earnings per share (EPS) growth for the calendar year 2027, AMD is the stock with the strongest buying momentum in the market. Our research feedback indicates that recent analyst day activities have helped solidify its revenue/profit margin targets and a future EPS target of up to $20." The team led by Citigroup analyst Christopher Danely wrote in a report to clients and reiterated a target price of $260.

The average target price compiled by TIPRANKS from Wall Street analysts shows that analysts expect an average target price for AMD of up to $282.33, **implying a potential upside of at least 28% within the next 12 months, with the highest target price being the bullish target of $377 given by the well-known investment firm Raymond James in its initial coverage of AMD stock.**

![1767687041(1).png](https://imageproxy.pbkrs.com/https://img.zhitongcaijing.com/image/20260106/1767687043196578.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg)

AMD CEO Lisa Su provided a very optimistic outlook for the artificial intelligence (AI) computing chip market in early November, expecting AMD's performance to show a stronger growth trajectory over the next five years. Lisa Su announced AMD's financial fundamentals targets for the next three to five years, stating that AMD aims to capture a "double-digit" share in the data center AI chip market, with annual revenue from AMD's data center chips expected to reach $100 billion within five years (compared to about $16 billion currently), and profits expected to more than double by 2030.

## "AI Bull Market Narrative" Will Continue to Dominate Global Stock Market Upward Trajectory in 2026

Morgan Stanley, another Wall Street financial giant, released a latest research report indicating that under the unprecedented AI infrastructure boom, the "long-term bull market logic" for chip stocks remains intact. In 2026, chip stocks centered around AI chips and memory chips are likely to be one of the best-performing sectors in the U.S. stock market, while the AI data center optical interconnect industry chain may grow into a stronger new generation of technological force.

Bank of America stated in its research report that **the global AI arms race is still in the "early to mid-stage,"** and despite the recent sharp downward fluctuations in popular chip stocks like NVIDIA and Broadcom, investors should continue to focus on industry leaders. One of the largest asset management giants globally, Vanguard, recently pointed out in a research report that **the AI investment cycle may have only completed 30%-40% of its final cycle peak.**

The Gemini3 series products, once launched, brought an enormous AI token processing capacity, further validating Wall Street's claim that "the AI boom is still in the early acceleration phase of supply-demand imbalance in computing infrastructure." **According to the latest semiconductor industry investment outlook for 2026 released by two major Wall Street giants, Bank of America and Morgan Stanley, the investment theme of the "AI bull market narrative" will continue to dominate the bullish trend of the global stock market in 2026—such as "AI chips."The storage chips and optical interconnect technology closely associated with the expansion of AI training/inference systems can be regarded as the main investment line in the chip sector favored by both parties.**

Strategists from the international banking giant UBS also expect that **the AI investment boom led by AI chip giants like NVIDIA and the strong profit growth will support the bullish trajectory of the U.S. stock market in 2026.** "We have noticed that the forward price-to-earnings ratio has only slightly increased since the beginning of the year, which further reinforces the fact that strong earnings growth, rather than market concerns about a 'valuation bubble,' is driving the market upward," the strategists wrote in a recent research report.

According to Wall Street giants Morgan Stanley, Citigroup, Loop Capital, and Wedbush, the global investment wave in artificial intelligence infrastructure centered around AI chip computing hardware is far from over; it is merely at the beginning. Driven by an unprecedented "storm of demand for AI inference computing power," this round of AI infrastructure investment, expected to last until 2030, could reach a scale of $3 trillion to $4 trillion.

“**We believe that 2026 is the midpoint of an 8 to 10-year process of upgrading traditional IT infrastructure to accommodate accelerated and AI workloads.**” “A greater scrutiny of AI investment returns and the cash flows of hyperscale cloud service providers may lead to stock price volatility, but this will be offset by newer/faster LLM developers and AI factories serving enterprise and sovereign clients. **We predict that semiconductor sales will approach the first $1 trillion in 2026, achieving approximately 30% growth, while wafer fab equipment sales will see double-digit year-on-year growth,” analysts from Bank of America stated in their research report. Bank of America indicated that the chip sector, which is leading the U.S. stock market into this super bull market, is expected to continue its bullish trend in 2026, also emphasizing that NVIDIA and Broadcom are the most worthwhile chip stocks to hold long-term in 2026.**

### 相关股票

- [XL2CSOPNVDA (07788.HK)](https://longbridge.com/zh-CN/quote/07788.HK.md)
- [GraniteShares 2x Long NVDA Daily ETF (NVDL.US)](https://longbridge.com/zh-CN/quote/NVDL.US.md)
- [YieldMax NVDA Option Income Strategy ETF (NVDY.US)](https://longbridge.com/zh-CN/quote/NVDY.US.md)
- [NVIDIA (NVDA.US)](https://longbridge.com/zh-CN/quote/NVDA.US.md)
- [XI2CSOPNVDA (07388.HK)](https://longbridge.com/zh-CN/quote/07388.HK.md)
- [Direxion Daily NVDA Bear 1X ETF (NVDD.US)](https://longbridge.com/zh-CN/quote/NVDD.US.md)
- [T-Rex 2X Long NVIDIA Daily Target ETF (NVDX.US)](https://longbridge.com/zh-CN/quote/NVDX.US.md)
- [T-Rex 2X Inverse NVIDIA Daily Target ETF (NVDQ.US)](https://longbridge.com/zh-CN/quote/NVDQ.US.md)
- [Direxion Semicon Bull 3X (SOXL.US)](https://longbridge.com/zh-CN/quote/SOXL.US.md)

## 相关资讯与研究

- [Fractile Seeks $200 Million as Nvidia Faces Inference Pressure](https://longbridge.com/zh-CN/news/281215985.md)
- [Nvidia’s $2 Billion Investment in Marvell: What Kind of Company Is Marvell?](https://longbridge.com/zh-CN/news/281372238.md)
- [Nvidia Stumbled in the First Quarter: Here are My Top 3 Nvidia Predictions for the Second Quarter](https://longbridge.com/zh-CN/news/281620093.md)
- ['Buy AMD Stock,' Says Top Analyst as Strong Data Center Demand Drives Upgrade](https://longbridge.com/zh-CN/news/281601702.md)
- [Nvidia (NVDA) H100 Prices Surge 40% as New GPUs Fail to Meet Insatiable Demand](https://longbridge.com/zh-CN/news/281588202.md)