This article digs into how Advanced Micro Devices (AMD) and Broadcom are carving out their spots in the exploding AI infrastructure market. It covers the product launches, strategy pivots, and what analysts are eyeing for 2026 and beyond.
You’ll find a look at what’s driving hyperscale investments, the main platforms and chips on the table, plus how investor moods are shaping valuations and ratings.
Industry momentum behind AI infrastructure
The global race to scale up artificial intelligence workloads is pushing data-center hardware spending higher and higher. Hyperscalers plan to pour a lot of capital into AI accelerators, high-bandwidth interconnects, and rack-scale platforms as we move toward 2026.
In this climate, AMD and Broadcom are leaning into their product lines to grab a bigger slice of the action from faster AI model training, inference, and deployment. AMD is rolling out its “AI Everywhere, for Everyone” strategy, along with a bunch of new platforms and processors meant to widen AI adoption for both enterprises and hyperscalers.
Broadcom, on the other hand, keeps reporting steady gains in AI-related revenues. Custom accelerators and networking products are at the heart of this, showing that software-defined AI workloads are actually leading to real hardware demand.
AMD: Strategy and catalysts shaping the AI push
AMD’s short-term focus is all about a broad AI-centric platform approach. They’re mixing data-center accelerators with rack-scale systems, aiming for flexibility.
The lineup looks pretty strong: there’s the Helios rack-scale platform, the Instinct MI400 family, and the Ryzen AI 400/AI PRO 400 processors. AMD wants these products to deliver scalable AI compute across both cloud and on-premises environments.
AMD expects its data-center AI revenues to grow at a compound annual growth rate (CAGR) of over 80% in the next three to five years. Overall data-center revenues should rise more than 60% annually in that same stretch. Will those numbers hold? Hard to say, but here’s what they’re banking on:
- MI450 chips linked to a 6 GW deal with Meta, anchoring a big chunk of AI workloads.
- Early momentum for Helios through partnerships with folks like OpenAI and HPE.
- The MI500 series, built on the 2-nanometer CDNA 6 architecture, aiming to boost efficiency and performance for huge AI pipelines.
Still, AMD faces some near-term headwinds. The first-quarter outlook for 2026 looks a bit soft, with expected revenues around $9.8 billion, hinting at a sequential dip. Competition isn’t letting up either—NVIDIA and Intel remain tough rivals in the data-center AI acceleration game.
Broadcom: AI momentum and its impact on revenues
Broadcom’s latest quarterly results really highlight its AI-driven growth. In fiscal Q1 2026, semiconductor revenue jumped 52% year over year to $12.52 billion.
Custom accelerators (XPUs) shot up 140%, and AI-related revenues surged 106%. Broadcom thinks AI revenues will hit about $10.7 billion in the next period, with total revenue close to $22 billion for fiscal Q2 2026.
One detail that stands out: AI networking is set to make up around 40% of Broadcom’s AI revenues. Their networking and accelerators are clearly becoming essential for AI workloads, from model training to low-latency inference in data centers.
- Analysts have bumped up earnings expectations: Broadcom’s fiscal 2026 EPS consensus is now $11.12.
- AMD’s 2026 consensus crept up to $6.61 per share as investors look for more AI-fueled growth.
Valuation, sentiment, and what to watch
Looking at year-to-date stock moves, both companies are feeling some pressure—AMD is down about 6.9% and Broadcom is off roughly 8.7%. Long-term AI demand still looks strong, though. On valuation, AMD’s forward price-to-sales (P/S) multiple is around 6.7x, a good bit lower than Broadcom’s 12.6x.
That gap might mean expectations for AMD are a bit more muted, or that Broadcom’s premium multiple bakes in more near-term quality. Right now, analyst sentiment puts Broadcom at a #2 Buy and AMD at a #3 Hold in a lot of rating systems. Some investors seem to like Broadcom for the short run, but there’s a sense that AMD’s longer-term AI play could eventually deliver bigger data-center growth.
Bottom line for researchers and investors
AMD and Broadcom both stand out as big winners in the current AI infrastructure wave. They’re telling very different stories, though.
AMD leans into a broad platform approach, pushing for higher AI compute density and building big partnerships. Broadcom, meanwhile, relies on a high-margin mix of XPUs and AI networking, aiming for stable, predictable AI revenue.
Researchers watching the AI hardware world should keep an eye on the next 12 to 24 months. It’ll be fascinating to see how MI450 rollouts, Helios uptake, and CDNA 6 efficiency actually impact real-world workloads and profit margins in such a cutthroat market.
With the AI revolution in full swing, it’s smart for stakeholders to track the rhythm of data-center AI launches, hyperscale capex trends, and how supplier relationships shift as new node technologies mature. Will the expected boom in AI infrastructure spending really turn into lasting leadership for AMD and Broadcom? The next few quarters should make things a lot clearer—though, let’s be honest, nothing’s ever set in stone in tech.
Here is the source article for this story: AMD vs. Broadcom: Which Semiconductor Stock is a Buy Right Now?