Track market-moving stories before they get noisy

Key Takeaways AMD launched Versal Premium Gen 2 MoP SoCs with up to 32GB LPDDR5X memory in one package.The devices deliver up to 288GB/s bandwidth and cut board space requirements by as much as 60%.Versal Premium Gen 2 MoP devices are expected to begin sampling by the end of 2026. Advanced Micro Devices (AMD - Free Report) recently introduced Versal Premium Gen 2 Memory on Package (MoP) adaptive system-on-chips (SoCs), integrating up to 32GB of LPDDR5X memory into a single package to deliver up to 288GB/s bandwidth while reducing board space requirements by as much as 60%. The new architecture eliminates the complexity of board-level memory design, enabling compact, high-performance systems for AI, networking, aerospace and defense, test and measurement, and professional video applications.

The new devices support PCIe 6.0, CXL 3.1 and LPDDR5X speeds of up to 9,000Mb/s, allowing seamless pairing with AMD EPYC processors and access to CXL memory expansion for data-intensive workloads. The launch of Versal Premium Gen 2 Memory on Package (MoP) adaptive SoCs strengthens AMD’s position in the fast-growing AI infrastructure market by addressing one of the biggest challenges in edge AI, networking, aerospace and defense systems, delivering higher memory bandwidth in compact, power-efficient designs. The latest solution expands AMD’s opportunities beyond hyperscale AI servers into embedded AI, telecom, defense and industrial markets.

The new MoP devices further complement AMD’s EPYC server processors through native PCIe 6.0 and CXL 3.1 connectivity, enabling customers to build scalable, memory-intensive AI platforms. This reinforces AMD’s strategy of providing a broad compute portfolio that delivers the best performance and total cost of ownership across different workloads. This brings a competitive advantage to AMD against the likes of NVIDIA (NVDA - Free Report) and Broadcom ((AVGO - Free Report) ). The EPYC processor is playing a significant role in driving AMD’s data center momentum. CEO Lisa Su noted that first-quarter data center revenues surged 57% year over year, fueled by strong EPYC and Instinct sales, while server CPU revenues climbed more than 50%.

AMD expects server CPU revenues to grow more than 70% in the second quarter, supported by rising adoption of EPYC processors. AMD is on track to launch sixth-gen EPYC Venice later in 2026, with more customers validating platforms than prior generations. AMD now aims to accelerate customer time-to-market by providing a pre-validated in-package memory interface compatible with existing Vivado and Vitis design tools. Versal Premium Gen 2 MoP devices are expected to begin sampling by the end of 2026, while standard Versal Premium Series Gen 2 devices are already shipping.

Tough Competition Hurts AMD’s ProspectsAMD’s prospects suffer from stiff competition. NVIDIA and Broadcom are major competitors in the Data Center space.

NVIDIA is at the center of AI computing, with its products widely used across data centers, gaming and autonomous vehicles. The company’s newer Hopper 200 and Blackwell GPU platforms are being adopted quickly as customers work to grow their AI infrastructure. Data Center revenues reached $75.2 billion in the first quarter of fiscal 2027, up 92% from a year ago and up 21% sequentially, driven by the ramp-up of Blackwell 300 products and demand for InfiniBand, Spectrum-X Ethernet and NVLink solutions. NVIDIA remains AMD's primary rival in GPU-accelerated supercomputing.

Broadcom is benefiting from strong demand for its networking products and custom AI accelerators. In the second quarter of fiscal 2026, AI semiconductor revenues reached a record $10.8 billion, up 143% year over year and above management’s outlook. Broadcom expects AI semiconductor revenues to reach $16 billion in the third quarter of fiscal 2026, up more than 200% year over year. For fiscal 2026, management expects AI semiconductor revenues of $56 billion, up approximately 180% from fiscal 2025. Broadcom also reiterated that AI semiconductor revenues are expected to exceed $100 billion in fiscal 2027.

AMD’s Share Price Performance, Valuation & EstimatesAMD shares have jumped 142.7% year to date, outperforming the broader Zacks Computer and Technology sector’s growth of 15%.

AMD Stock’s Price Performance
Image Source: Zacks Investment Research

AMD stock is overvalued, with a forward 12-month price/sales of 15.88X compared with the broader sector’s 6.49X. AMD has a Value Score of F.

AMD Valuation
Image Source: Zacks Investment Research

The Zacks Consensus Estimate for second-quarter 2026 earnings is pegged at $1.60 per share, unchanged over the past 30 days, suggesting 233.3% year-over-year growth.

AMD currently carries a Zacks Rank #3 (Hold). You can see the complete list of today’s Zacks #1 Rank (Strong Buy) stocks here.

Advanced Micro Devices NASDAQ:AMD stock has surged more than 130% this year, but Wall Street is still chasing the stock higher.

In June alone, Barclays, UBS, Mizuho and Bernstein all raised their price targets on the chipmaker as analysts are no longer treating AMD as just a second-place GPU challenger to Nvidia.

They are increasingly arguing that CPUs are becoming an AI story and the driver is agentic AI, or AI systems that do more than answer one prompt.

Barclays was one of the first major firms to put a bigger number on the CPU opportunity.

Analyst Tom O’Malley raised his AMD price target to $665 from $500 and kept an Overweight rating.

His core argument was that “CPU-to-GPU ratios are narrowing as CPU demand reaches new levels in the rapidly expanding world of agentic AI,” adding that AMD is “best positioned to benefit from this transition”.

The CPU-to-GPU ratio simply means how many central processors are needed for every graphics processor inside AI systems.

Early AI spending was dominated by GPUs because training large models required enormous parallel computing power.

Agentic AI changes the mix because it needs more coordination, routing and software execution around those GPUs.

O’Malley’s model sees the standalone server CPU market approaching $200 billion by 2030.

UBS pushed the argument even further.

Analyst Timothy Arcuri raised his AMD target to $670 from $455 and kept a Buy rating.

That now stands above Barclays’ $665 call and makes UBS one of the most bullish voices on the stock.

The firm said it was “incrementally more constructive” on AMD as standalone CPU racks gain traction.

In plain English, UBS thinks customers are starting to buy CPU-heavy systems for AI workloads that do not rely only on GPU clusters.

That matters because AMD’s CPU business has often been overshadowed by its Instinct GPU ramp.

Investors are watching whether AMD can become a credible second source to Nvidia in AI accelerators. UBS is saying another part of the story may be hiding in plain sight: server CPUs.

Arcuri lifted his 2030 AMD server CPU revenue forecast to $50 billion from $41 billion.

Mizuho and Bernstein added a second layer to the bull case: scarcity.

Mizuho raised its AMD target to $615 from $515 and kept an Outperform rating, citing strong demand linked to agentic AI.

The firm also flagged that CPU and memory suppliers could remain supply-constrained into 2027.

That turns the story from pure demand into a supply-side argument.

If companies need more CPUs for AI workloads, and supply remains tight, pricing and revenue assumptions may have room to move higher.

Bernstein also raised its AMD target, lifting it to $600 from $525 while maintaining an Outperform rating.

The firm increased its 2030 server CPU market estimate to $223 billion from $137 billion, reflecting a much larger opportunity tied to agentic AI.

The caveat is valuation, as AMD’s average Wall Street price target still sits below where the stock recently traded, which means shares have already run ahead of broad consensus.

The next real tests are AMD’s Advancing AI event in July and Q2 earnings in early August.