Investing.com -- Advanced Micro Devices shares jumped roughly 7% on Tuesday after management outlined an aggressive artificial-intelligence growth strategy at the Citi Global TMT Conference, anchored by a new $2 trillion total addressable market estimate by 2030.
At the core of AMD's thesis is a pivotal shift in the AI landscape. Computing demand is rapidly transitioning from training large AI models to executing them—a process known as inference. According to Citi's summary of the event, AMD management noted that "In the past twelve months, Inference has become the majority driver of AI computing," with workloads shifting from basic chatbots toward independent, agentic AI systems.
To capture that expanding market, AMD's data center business is projected to more than double next year. The company's next-generation AI accelerator, the MI450 GPU, will launch this quarter, with production ramping up in the fourth quarter and into 2027. AMD's traditional computing segments are also surging to support this infrastructure; the server CPU business is expected to expand more than 80% year-over-year in the second half of this year and over 70% next year.
The clearest sign of mounting demand comes from AMD's complex, rack-scale Helios systems. The company counts Meta Platforms and two unnamed AI labs as its anchor clients, all of which have provided demand forecasts larger than their initial purchase agreements. With Helios volume expectations for 2027 already outpacing initial projections, AMD is now seeing demand emerge from new "neo-cloud" providers.
While data center GPUs currently yield profit margins below AMD's corporate average, the company's broader portfolio is offsetting the dilution. Strong double-digit growth in its high-margin server and embedded segments during the second and third quarters has kept overall profitability intact.
Looking ahead, AMD is already securing its future hardware pipeline. Management expressed confidence in the durability of its demand, noting ongoing technical engagements with its three leading customers for next-generation MI500 and MI600 chips. Additionally, AMD disclosed a partnership with Cerebras and hinted at unannounced internal plans to further target the low-latency inference market.
(0)Comments