Wedbush raises AMD (AMD.US) price target to $600: AI conference unleashes “end-to-end” ambition, 2027 profit forecast soars 20%

The Zhitong Finance App learned that after AMD successfully hosted the “Promising AI 2026” conference in San Francisco, Wall Street brokerage firm Wedbush Securities released a major research report on Monday, drastically raising the target price of AMD (AMD.US) from $450 to $600, maintaining a “outperforming the market” rating. More notably, Wedbush simultaneously drastically revised AMD's 2026 and 2027 performance forecasts — the 2027 revenue forecast jumped directly from US$77.9 billion to US$81.2 billion, and earnings per share were raised from US$12.33 to US$14.74.

Wedbush analyst Matt Bryson pointed out in the report that AMD has successfully repositioned itself as an “end-to-end computing business covering GPUs, CPUs, networks, software, and growing client and physical AI,” and showcased “an unusually strong lineup of cutting-edge laboratories and enterprise partners” at the conference. At the core of this strategic transformation is AMD's evolution from a single chip vendor to an AI infrastructure provider that can compete head-on with Nvidia (NVDA.US) at the system level.

Confidence behind raising the target price: EPS is expected to soar 20% in 2027

The magnitude and scope of the current increase by Wedbush have exceeded market expectations. Let's take a closer look:

2026 Q2 Forecast: Earnings per share increased from $1.61 to $1.64 billion, and revenue increased from $11.3 billion to $11.4 billion

Full year 2026: Earnings per share increased from $6.95 to $7.22, and revenue increased from $47.2 billion to $49.2 billion

Full year 2027: Earnings per share increased sharply from $12.33 to $14.74 (+19.5%), and revenue increased from $77.9 billion to $81.2 billion (+4.2%)

Bryson made it clear that the newly announced cooperation agreement with Microsoft (MSFT.US) and Anthropic gave Wedbush “significantly stronger confidence” in AMD data center AI chip/system revenue growth in the second half of 2026 and 2027. Additionally, communication with server vendors and supply chain participants indicates that AMD is better addressing supply chain challenges to meet customers' growing demand for data center computing.

Innovating AI 2026: A “full-stack revolution” from chip to system

AMD's 2026 conference, hosted by AMD in San Francisco from July 23 to 24, was the direct catalyst for this rating increase. AMD CEO Su Zifeng unveiled a matrix of seven new full-stack AI products at the conference, covering the three levels of hardware, network and software.

The leap from chip to cabinet level. The Helios rack-level AI platform is the core of the entire launch. Up to 72 Instinct MI455X AI accelerators can be deployed in a single rack, with 18 Venice EPYC CPUs based on Zen6 architecture, the FP4 peak computing power of 2.9 exaFlops, and is equipped with a 31TB HBM4 unified video memory pool. AMD claims that compared to Nvidia's Vera Rubin NVL72 solution, Helios' AI computing power is 15% higher, HBM capacity is 50% higher, horizontal expansion bandwidth is 50% higher, and can generate up to 30% more tokens per dollar. Su Zifeng made it clear that the first shipment of Helios will begin in September 2026, and will continue to be released in the fourth quarter and the first half of 2027.

Demand for Agentic AI-powered CPUs has exploded. AMD executives pointed out that as AI moves from model training to large-scale deployment of AI agents (Agentic AI), the role of CPUs is being redefined. AI agents require significant general-purpose computing power to complete inference, scheduling, and coordination tasks — which is the strength of x86 CPUs. Su Zifeng anticipates that in some scenarios, the ratio of CPU to GPU may evolve from the current 1:4 to 2:1 — that is, each GPU needs to be equipped with two CPUs. AMD has reiterated its goal of gaining more than 50% share of the server CPU market.

In her speech, Su Zifeng gave the latest industry forecast: the global AI computing market is expected to exceed 2 trillion US dollars in 2030, of which the AI accelerator market will reach 1.4 trillion US dollars and the data center CPU market will reach 220 billion US dollars. She stressed that the AI industry has moved from being driven by a single GPU computing power to a “full-stack collaborative computing power era of CPU+GPU+DPU+FPGA+ software ecosystem.”

The “biggest leap forward” in the software ecosystem. AMD has released the ROCM.ai global AI software platform — called by executives “the biggest software leap since the early days of the strategy.” Through collaboration with OpenAI on Codex and Anthropic on Claude, developers' access to the AMD platform will improve significantly over the next few months.

Customer Matrix: Microsoft, Anthropic, Meta, OpenAI collective platform

AMD's customer lineup is luxurious. Su Zifeng personally confirmed that Anthropic, OpenAI, and Meta have decided to deploy the Helios system. The partnership with Anthropic is particularly critical — AMD will begin delivering the first 1 gigawatt of Mi450 accelerator computing power in the first half of 2027. In addition to the previously announced Microsoft Azure deployment, AMD has promised to increase the total scale of AI accelerators deployed to 14 gigawatts.

Anthropic: A strategic partnership was announced on July 22. Anthropic will deploy up to 2 gigawatts of AMD MI450 series GPUs in AMD Helios rack-scale solutions. The first gigawatt deployment will begin in the first half of 2027. AMD also promised to invest up to $5 billion in strategic equity in Anthropic. AMD will use Anthropic's Claude model to optimize Instinct GPU workloads and accelerate ROCM development.

Microsoft: Azure data centers will deploy AMD Helios rack-scale systems, and Helios is expected to ship to customers in the second half of 2026.

Meta, OpenAI, and Oracle have all confirmed plans to deploy the Helios system. Su Zifeng revealed that Helios was not independently completed by AMD and then handed over to customers, but was co-designed with leading companies such as OpenAI, Meta, and Anthropic. From chips and software to data centers and supply chains, customers entered the product definition process earlier.

Supply chain beneficiaries: the ripple effect from servers to storage

Wedbush pointed out that AMD's announcement at the conference will have a positive impact on a number of supply chain partners:

Cerebras (CBRS.US): Its wafer-level engine for inference will work with AMD's Helios rack

Ultramicrocomputer (SMCI.US) and Dell (DELL.US) are expected to see profit growth

Memory vendors Samsung Electronics and Micron (MU.US) will continue to benefit as AMD's main memory suppliers

Market's “poor expectations”: Stock prices fall back, Wall Street is optimistic

Despite Wedbush's sharp increase in target price, AMD shares fell more than 5% on July 27 to close at around $494.95. This seemingly contradictory market reaction just reveals the current valuation dilemma facing AI chip stocks — after experiencing a sharp rise in the first half of 2026, the market has fully priced any “in line with expectations” favorable news, and further increases in stock prices require continuous verification of performance beyond expectations.

AMD is transforming from a “catcher” to a “definer.” The mass production of the Helios rack-level system, the expansion of Venice CPUs, the advancement of the MI400 series accelerators, and the improvement of the ROCM.AI software ecosystem form a complete “end-to-end” AI computing power puzzle. More importantly, the CPU demand revaluation brought about by Agentic AI is fundamentally changing the rules of the game — when each GPU may require two CPUs, AMD's leadership in the x86 server CPU market will turn into an unprecedented growth lever.

Wedbush's target price of $600 means there is still about 21% upside from Monday's closing price. AMD plans to announce financial results for the second quarter of 2026 on August 4. At that time, the market will receive the latest data to test the growth momentum of its AI business — and Wedbush's raised expectations will be an important yardstick for investors to measure the success of this financial report.

Wall Street's collective follow-up: AMD has become a “must”

Wedbush isn't the only agency that raised AMD's price target.

Futurum Equities gave Wall Street's highest target price of $800, and AMD's GPU revenue is expected to reach 38 billion US dollars next year.

UBS raised the target price from $700 to $730, maintaining the buy rating. Analysts specifically pointed out that the gross margin of AMD server CPUs is about 10 percentage points higher than the company's overall gross margin — the most profitable product line also happens to be the fastest growing product line.

Bank of America raised the target price to $620, believing that AMD has “successfully transformed from a commercial GPU vendor to a full-stack, cabinet-level accelerator system supplier.”

Goldman Sachs reaffirmed its buy rating, with a 12-month target price of $640. Jefferies maintains a buy rating with a target price of $640. Benchmark raised the target price from $485 to $685.

According to S&P Global's survey of 51 analysts, the AMD consensus rating is “Strong Buy,” with an average target price of $575.49.