Broadcom’s AI chip surge reframes the competitive threat to Nvidia, even as AMD also reports strong data-center results

Nvidia’s closest competitors in the AI data-center buildout are often framed as AMD on one side and custom silicon partners on the other. But the latest round of earnings coverage is spotlighting a different stress point, arguing that Broadcom may be a more immediate challenge than it typically appears in headline comparisons.

Broadcom reported Q2 fiscal 2026 results with the timing noted as June 3, and the company highlighted a major jump in AI semiconductors revenue, described in recent earnings coverage as up 143%. Broadcom’s broader relevance to AI infrastructure is that it supplies a range of components used throughout servers and the networks that connect them, not only stand-alone compute.

The contrast with traditional “GPU competition” narratives matters because the AI supply chain is not a single bottleneck. Chips, networking, and the surrounding system components all determine performance and total cost. If AI spend is pulling through multiple layers of the stack, vendors that can capture AI-linked demand across those layers can become competitive in ways investors may not fully map to Nvidia’s direct products.

In the same earnings window, AMD also posted data-center results that were characterized as forcing a “rethink” of Nvidia’s real competition. The coverage places both companies at the center of a question: whether Nvidia’s advantage is facing pressure primarily at the level of accelerators, or whether other semiconductor suppliers are increasingly capturing adjacent AI spending that can reduce share for any single supplier.

For Broadcom, the reported AI semi revenue acceleration suggests the company is benefiting from the continued expansion of AI infrastructure deployments. A 143% increase, if sustained, would imply that demand for the company’s AI-linked semiconductor offerings is rising faster than overall market growth. That kind of pace can change expectations for how much of AI infrastructure spend is being allocated to suppliers beyond GPU makers.

Still, the specific mechanics of what Broadcom’s AI semiconductor growth represents were not detailed in the market coverage described here. The post does not break down product lines, customer segments, or whether the revenue growth is concentrated in particular networking or server components. Without those disclosures in the cited coverage, it is not possible to say which exact sub-market is driving the acceleration.

From a sector perspective, the key takeaway is that “competition” for Nvidia is increasingly multidimensional. Companies that can scale AI-related semiconductor supply, and that can sell into the broader data-center system, may matter as much as chip-for-chip rivalry. If AI buildouts continue to pull through multiple layers of the hardware stack, Broadcom’s scale and growth trajectory could put more pressure on how buyers allocate budgets across suppliers.

What to watch next is whether Broadcom continues to report similar AI semiconductor momentum in subsequent quarters, and whether AMD’s data-center commentary points to sustained share gains in accelerators or other infrastructure areas. Also important will be any additional disclosure on the composition of Broadcom’s AI semi revenue, since that would help clarify whether its gains are broad-based or tied to a narrower set of products or customers. Until then, the competitive framing can be adjusted, but the exact threat profile remains partly unspecified.