Bank of America Securities has sharply raised its long-term outlook for the market for server central processing units (CPUs), predicting that the next wave of artificial intelligence spending will be about running AI systems, not just building them. In a client note, analyst Vivek Arya lifted his 2030 forecast for server CPUs to more than $240 billion, up from roughly $210 billion, versus an estimated $64 billion this year.
The upgrade centers on what the bank calls “agentic AI” — software that can autonomously coordinate multiple AI tasks. Instead of just training massive models, the next phase of AI adoption will involve “inference,” the process where AI generates answers in real time, and the orchestration layer that manages those tasks. BofA argues these workloads will require more CPU sockets per server and higher-priced chips, even if some parts of the AI buildout cool later in the decade.
What is a server CPU and why does it matter?
A server CPU is the main processor inside a data center server, handling the core computing tasks that run websites, databases, and increasingly, AI applications. While graphics processing units (GPUs) have grabbed most of the headlines in the AI boom for their role in training large models, CPUs remain essential for the broader computing workload — especially as AI moves from training to everyday use.
BofA’s revised forecast effectively splits the CPU opportunity into two buckets: a slower-growing “traditional” business and much larger AI-linked segments by 2030. If that mix shift is correct, investors may start paying less attention to short-term swings in AI training budgets and more to who is positioned for the steadier, usage-driven phase: running AI systems at scale across many applications.
AMD vs. Intel: the supply story
The note leans heavily on supply consistency. BofA raised its price objective for AMD and lowered Intel’s, arguing that the winners will be the companies that can reliably deliver leading-edge server parts as demand shifts toward inference-heavy deployments. This makes manufacturing and product cadence a bigger part of the valuation story, because missed supply windows can mean lost share in data-center refresh cycles.
AMD has been gaining market share in server CPUs over the past few years, while Intel has struggled to keep pace with its manufacturing roadmap. The bank’s relative stance is a signpost of what the market is trying to price: durable share gains and pricing power as inference becomes a larger slice of data-center spending.
What it means for investors
For everyday investors, this forecast is a reminder that the AI boom is not just about the handful of companies that make the most famous chips. The entire data-center supply chain — including server CPU makers — stands to benefit if AI workloads expand as expected. But it also highlights the risks: if the AI buildout slows or if companies shift spending away from CPUs, the market could look very different.
BofA’s $240 billion figure is a forecast, not a guarantee. It depends on assumptions about how quickly agentic AI is adopted and how much computing power those applications require. Investors should watch for signs of that adoption in earnings reports from major cloud providers and chipmakers, as well as any updates to capital spending plans.
The note also underscores the importance of supply reliability. In a market where data-center operators plan upgrades years in advance, a company that can consistently deliver high-performance chips can lock in customers and pricing. That is why BofA’s preference for AMD over Intel is not just about current products, but about who can execute on future roadmaps.
For those tracking the broader tech sector, this story ties into the ongoing debate about whether AI spending is sustainable. Nvidia's leadership transition has added some uncertainty to the AI trade, but the underlying demand for computing power remains strong. Similarly, market reactions to Nvidia's CEO exit show how sensitive investors are to any sign of disruption in the AI supply chain.
In the end, BofA’s revised forecast is a bet that the AI story is far from over — it is just entering a new phase. For investors, the key takeaway is to look beyond the training hype and consider who will benefit when AI systems are actually deployed and used every day.
