Key Takeaways
- Arm’s price target jumps to $500 from $300, with Outperform rating reaffirmed
- AMD (AMD) price target increases to $600 from $525, Outperform rating continues
- Intel (INTC) target moves to $100 from $65, Market-Perform rating unchanged
- Server CPU market forecast expanded to $223B by 2030, up from previous $137B estimate
- Emerging agentic AI applications pushing CPU-to-GPU ratios closer to 1:1 from traditional 1:4 configurations
Investment bank Bernstein has significantly increased its price targets across three leading semiconductor companies — Arm, AMD, and Intel — citing explosive growth in server CPU demand fueled by the emergence of agentic AI technologies.
Advanced Micro Devices, Inc., AMD
Agentic AI Transforms Data Center Architecture
Lead analyst David Dai and his Bernstein research team observe that artificial intelligence applications are evolving beyond simple conversational chatbots toward sophisticated agentic AI platforms — autonomous systems capable of planning and executing multifaceted tasks independently. This evolution dramatically increases the computational load on central processing units, not merely graphics processors.
According to the analysts, traditional AI infrastructure deployed CPU-to-GPU ratios ranging from 1:4 to 1:8. However, agentic AI workloads are pushing these ratios toward parity at 1:1 or even higher. This architectural shift fundamentally alters the demand landscape for CPU manufacturers.
Reflecting this transformation, Bernstein has substantially revised its server CPU total addressable market projection upward to $223 billion by decade’s end. The new forecast represents a significant increase from the prior $137 billion estimate and shows approximately six-fold growth from the 2025 TAM of $37 billion.
The firm’s baseline projection assumes cumulative AI data center capital expenditure of $3.5 trillion, with balanced CPU-to-GPU deployment for inference operations.
Arm Positioned as Primary Winner
Bernstein elevated Arm’s price objective to $500 from $300 while maintaining its Outperform rating. The research firm identifies Arm as the “structural beneficiary” positioned to capitalize most significantly on the CPU resurgence.
Arm’s chip design architecture offers compelling advantages for agentic AI applications, particularly regarding power efficiency metrics. Beyond its traditional intellectual property licensing business model, the company has expanded into developing proprietary CPU designs.
Bernstein’s updated financial model now projects Arm will achieve $22 billion in annual revenue by 2030 — substantially exceeding the company’s own $15 billion guidance. Separately, SoftBank, which maintains majority ownership of Arm, received an upgraded price target to ¥11,200 from ¥8,200.
For AMD, the price target climbed to $600 from $525. With shares trading near $507 on June 16, Bernstein’s new target implies approximately 18% upside potential.
This marks the second upward adjustment within a two-month period. Bernstein had previously elevated AMD from Market Perform to Outperform in May 2026, establishing a $525 target that itself represented a substantial jump from an earlier $265 objective.
Both AMD and Intel Benefit from Revised Outlook
The investment firm forecasts AMD will generate approximately $14.60 in earnings per share during fiscal 2027, with potential to reach $20 by 2028 should AI adoption trends continue accelerating. Several other Wall Street analysts have established price targets for AMD ranging between $625 and $665.
For Intel, Bernstein raised its price target to $100 from $65 while keeping its Market-Perform rating intact. The firm noted it refreshed its Intel financial model to incorporate the strengthening server CPU market environment, with estimate adjustments proving “more materially” significant than those for AMD.
Bernstein additionally increased its price target on Chinese CPU developer Hygon to 450 yuan from 280 yuan, projecting the company will secure more than 35% market share of China’s x86 server CPU sector by 2030.
While AMD’s hardware solutions previously enjoyed popularity within cryptocurrency mining communities, the investment thesis has decisively pivoted toward AI-driven server infrastructure demand.


