These trends are catalyzing the earnings for some of the key players powering the AI infrastructure. AMD has reported blockbuster Q1 2026 numbers:

• Revenue: $10.3 billion, up 38% YoY, driven by surging demand for CPUs and accelerators as customers race to build GW-scale AI infrastructure. This also helps AMD cross the $10 billion per quarter threshold.
  
  

AMD Revenue By Categories

• Margins: Gross margin up at 53% and operating margin up at 14%, but both down slightly sequentially.
• Cash Flow: Free cash flow almost tripled to $2.6 billion, providing a good foundation to scale further.
• Data Center:
  
  Full Stack Offering:
• • Data center revenue climbed to $5.8 billion, up 57% YoY and forming 56% of the total revenue.
  • Data center has become the core for AMD thanks to its full-stack portfolio covering compute (CPU, GPU, FPGA), networking (xPUs and open standards) and packaging expertise to ROCm software and rack-scale solutions.
  • After NVIDIA, AMD is the only credible full-stack merchant solution, making it a solid two-horse race.
  • NVIDIA data center revenues are still 7x AMD’s data center revenues, and AMD would like to close this gap with an uptick in rack-scale solutions.
    
    Agentic AI Accelerates CPU Demand:
  • With the rise of “Agentic AI”, which requires more balanced compute architectures, CPUs are back in high demand as the CPU-to-GPU ratio reduces.
  • Great traction for 5^th-Gen AMD EPYC CPUs from data centers to telco edge. They are expected to drive CPU revenues up 70% YoY in the coming quarter. Targeting Agentic AI workloads, the upcoming 6^th-Gen Zen 6 architecture-based AMD EPYC (Venice and Verano) is expected to be produced on TSMC’s advanced 2nm (N2) and push up revenues even further.
    
    

    

  • With Intel facing supply issues and unable to meet the insatiable demand for CPUs, AMD benefits as a solid alternative. As a result, it has climbed up to become an equally large player in the server CPU market, potentially surpassing Intel in terms of x86 server CPU market share.
  • The rise of Arm-based server CPUs also presents a solid alternative with a great outlook, but x86 still dominates, with AMD holding the fort.
  • The server CPU TAM is estimated to climb to the $110-$120 billion range per year by 2030, according to Counterpoint’s Server CPU Forecast.
  • On GPUs, despite earlier concerns about the complex stacked design of the MI455 causing downward shipment revisions, the supply chain is still targeting shipments to begin in mid-Q3.
  • The AMD MI5550 vs NVIDIA Rubin Ultra showdown will be the one to watch next year as the MI400 series gears up to challenge NVIDIA’s Rubin this year.
    
    

    

    
    Growing Customer Base and Strong Partnerships:
  • Meta has become a flagship partner for AMD and plans to deploy up to 6 GW of AMD Instinct GPUs, with the first GW to be powered by a custom AMD Instinct MI450-based GPU. Meta will also be a lead customer for 6^th-Gen AMD EPYC CPUs.
  • AWS, Google Cloud, Microsoft Azure and Tencent have announced new and expanded 5^th-Gen EPYC-powered cloud instances.
  • Beyond the known five-year 6GW partnerships with Meta and OpenAI (which involved AMD warrants), Anthropic has reportedly signed an agreement to use AMD's MI450 chips due to broader supply constraints.
  • Anthropic's rapidly growing annual recurring revenue (ARR), which has already crossed the double-digit billion-dollar mark, means the scale of this partnership could be substantial.
  • Downstream infrastructure readiness is important, supported by the likes of Wistron and Foxconn Industrial Internet, to cater to strong demand from gigawatt-scale customers like OpenAI, Meta and Anthropic through the next three years.
  • AMD and Samsung have announced a collaboration on HBM4 supplies for AMD Instinct MI455X GPUs and advanced DRAM solutions for 6^th-Gen AMD EPYC CPUs.
  • On the foundry side, the fight for the N2 and N3 capacities at TSMC is going to be a hot topic this year, apart from securing CoWoS capacity to help AMD scale without any bottlenecks.
  • AMD is betting big on Sovereign AI infrastructure. AMD and Tata Consultancy Services (TCS) are co-developing AMD Helios-based rack-scale AI infrastructure to accelerate enterprise AI deployments and Sovereign AI initiatives in India.
  • AMD is collaborating with South Korea’s leading NAVER Cloud and a promising startup, Upstage, to deploy AMD Instinct GPUs and EPYC CPUs across their AI infrastructure, advancing sovereign AI initiatives in South Korea.

• Client and Other Segments: PC Recovery Beckons, Embedded and GPU Slower
• • The Client and Gaming segment revenues climbed to $3.6 billion, up 23% YoY, mainly driven by a 25% increase in unit shipments for the Ryzen processors and robust demand for Radeon GPUs.
  • Embedded revenues grew to $873 million, up 6%, with steady demand for FPGAs and other solutions across end markets.

Key takeaways:

• Agentic AI tailwinds are driving demand for both CPUs and GPUs at AMD, alongside its Helios rack-scale AI solutions, which leverage its networking capabilities and an open standards approach across multiple applications, including sovereign deployments.
• AI Server CPU Catalyst: CPU demand continues to ramp up and AMD is poised to become the leading AI server CPU vendor this year. EPYC-powered cloud instances jumped nearly 50% YoY, expanding across AWS, Google Cloud, Microsoft Azure and Tencent.
• Anchor Customers: Meta being an anchor customer, AMD is leveraging the massive demand beyond striking deals with Anthropic and OpenAI, making it one of the biggest AI server compute suppliers after NVIDIA.
• Supply Chain Bottlenecks: These would be the only thing that can hold back AMD’s growth. Therefore, a prudent supply chain strategy and execution will be the key for front-end, backend and rack-scale manufacturing.