Amazon Adds Qualcomm to Reduce Nvidia Dependence
Advisors Capital Management partner JoAnne Feeney argues that Amazon’s expanded Qualcomm relationship is a supply-chain and bargaining move: another chip designer can reduce reliance on Nvidia, improve Amazon’s pricing leverage and give Qualcomm greater confidence to invest in custom-chip capacity. Speaking with Bloomberg’s Riley Griffin, Feeney frames the arrangement as a form of risk-sharing often labelled circular financing, linking buyer demand more directly to supplier investment as AI infrastructure expands.
Amazon is buying leverage, not merely another chip
JoAnne Feeney sees Amazon’s Qualcomm arrangement as a way to diversify the supply and design base for chips used in connectivity and, ultimately, AI inference. Amazon already works with Marvell, she noted; adding another partner reduces the risk attached to any one relationship.
The benefit is not just additional capacity. Multiple suppliers and co-designers give Amazon more alternatives when it acquires chips in the future, strengthening its pricing position. Feeney said the arrangement also offers protection from Nvidia’s price levels as competition among chip designers increases. Broadcom is another relevant competitor, though Amazon does not currently work with it.
For Feeney, this is a move toward a less concentrated supply chain: Amazon can spread sourcing risk, secure more potential supply, and negotiate with more than one credible designer.
Circular financing turns demand into a supplier commitment
Riley Griffin raised the Qualcomm arrangement in the context of concerns about “circular financing” during the AI boom, particularly where buyers, customers, and suppliers make cross-investments.
Feeney said that concern will arise whenever companies invest across those relationships. Her framing is that these arrangements can serve as a route to greater vertical integration—not necessarily through ownership alone, but through greater control over the inputs and technology a company needs to expand capacity.
Circular financing is really sort of an alternative way for companies to become more vertically integrated, to have more control over the inputs, the technology that they need to use to build out their capacity in the future.
The financial rationale, in her account, is that companies face different borrowing limits and balance-sheet capacities even when all must make large, long-lived commitments to AI infrastructure. Cross-investment and commercial arrangements can combine exposures that would otherwise be separate: owning Amazon may, as Feeney put it, mean owning “a little bit of Qualcomm.” Investors can still choose where to concentrate risk, but the relationship links the fortunes of buyer and supplier more directly.
The causal chain matters most at the supplier level. Long-term contracts and Amazon’s deeper involvement give Qualcomm more confidence that demand will support investment in capability. That can make it more willing to take the risk of building large design teams and committing resources to custom-chip work. Feeney described the result as risk-sharing: Amazon has more skin in the game, Qualcomm can invest more confidently, and innovation may happen faster than it would if the supplier bore more of the risk alone.
Nvidia can lose share in a market that still expands
JoAnne Feeney expects Nvidia’s dominant AI-chip market share to fall as more competitors enter. But she argued that a declining share should not be mistaken for a broken investment case.
Nvidia remains a holding across several Advisors Capital Management strategies, Feeney said. The more important question, in her view, is the size and growth of the overall market. AI infrastructure demand is growing quickly enough that multiple chip companies and the large technology companies deploying their products can participate in the expansion.
She named Broadcom, Nvidia, and AMD among the companies she favors, alongside major users including Amazon, Microsoft, and Google. Revenue growth at Anthropic and OpenAI, based on the information available about them, supports her view that infrastructure build-out and application deployment remain a substantial long-term growth opportunity.
Amazon’s search for more chip partners therefore does not, in Feeney’s framing, amount to a winner-take-all reversal for Nvidia. It is part of a market in which customers seek alternatives and suppliers need credible future demand before committing capital and design resources.
The physical build-out may be limited by labor
JoAnne Feeney identified labor as an operational constraint on the infrastructure strategy. Diversified chip supply and committed supplier investment do not by themselves determine how quickly new data-center capacity can be built.
She described labor being drawn into data-center construction amid a strong jobs report, with softness in some information-technology hiring but strength elsewhere. In her view, the current AI build-out is creating jobs in areas where labor is already scarce, rather than producing broad job destruction. She also pointed to H-1B visa fees that, she said, have constrained applications despite a need for more technical labor.
The more immediate pressure may be among the workers building facilities. Feeney said immigrant labor is an important part of both housing and commercial construction, and that shortages could constrain those sectors. As an example from a Texas client, she said a shortage of roofers had forced planned housing construction to be cut in half for the year.
Data-center developers have so far been able to pay more to obtain the construction labor they require, Feeney said. But pulling workers into those projects can raise costs and reduce available labor for other parts of the economy—making workforce availability a risk investors should keep in view.