AMD Slides 8.75%: Can Real Orders Rescue the AI Supplier Rotation?
The sharp reversal in AMD and Broadcom caught my attention because it exposed a weakness in the recent AI supplier-diversification trade.
AMD dropped 8.75% and Broadcom fell 4.76% on Friday, November 13, after both stocks had benefited from the market’s initial reaction to Jensen Huang’s departure from Nvidia.
At first, investors appeared to be positioning for a possible shift towards alternative AI chip suppliers. But the subsequent sell-off showed that enthusiasm can disappear quickly when the market has not yet seen evidence of actual business moving between companies.
My view is that the diversification opportunity is real, but the market may have priced in the benefits before the orders were fully demonstrated.
A leadership change does not automatically transfer customers
Nvidia’s leadership transition is significant. Jensen Huang resigned as CEO and board member, with CFO Colette Kress appointed interim CEO while the board searches for a permanent successor.
However, a change in leadership does not mean Nvidia’s customers will immediately abandon its products.
Nvidia still has a powerful combination of GPUs, networking, software, developer support and an established ecosystem. Its latest fiscal Q3 results reinforce that point: revenue reached $115.8 billion, up 103% year over year, while Data Center revenue climbed 110% to $107.6 billion.
The company also guided for approximately $130 billion in fiscal Q4 revenue.
That is not a business suddenly losing customer demand.
For AMD and Broadcom, the opportunity is therefore not simply to wait for Nvidia to stumble. They need to convince customers that their own solutions can deliver the performance, availability, cost efficiency and reliability required for large-scale AI deployments.
AMD has real momentum, but execution is the next test
AMD is the more direct alternative in the GPU and server CPU markets.
Its latest Q3 2026 results provide encouraging evidence. Revenue reached $13.4 billion, up 45% year over year, while Data Center revenue increased 97% to a record $8.5 billion.
AMD attributed the growth to strong demand for EPYC server processors and the launch of its Instinct MI450 Series GPUs.
More importantly, AMD has announced actual customer programmes rather than simply talking about potential demand.
Its partnership with Anthropic includes plans to deploy up to 2 gigawatts of MI450 GPUs, with the first gigawatt scheduled to begin deployment in the first half of 2027. AMD has also identified major customers running workloads on its first Helios rack-scale systems.
These are meaningful commercial developments.
But I would be careful about calling them confirmed Nvidia order transfers. They demonstrate that customers are adopting AMD infrastructure; they do not establish that those customers cancelled Nvidia purchases and redirected the same orders to AMD.
The next proof point is whether AMD can ship at scale, meet deployment schedules and convert these programmes into recognised Data Center revenue.
Broadcom is a different kind of AI opportunity
I would not analyse Broadcom in exactly the same way as AMD.
AMD competes directly in GPUs and server CPUs. Broadcom’s AI opportunity is more closely tied to custom accelerators, networking and the infrastructure required to connect large AI systems.
Its latest reported Q3 fiscal 2026 results were particularly strong.
Broadcom generated $16.7 billion in AI semiconductor revenue, up 221% year over year and 54% sequentially. Management expected that figure to increase to $21.7 billion in Q4, representing 236% year-over-year growth.
That is substantial evidence of demand for custom AI chips and networking.
The distinction matters because hyperscalers may not want to rely entirely on general-purpose GPUs. Some may develop custom accelerators for specific workloads where performance, power consumption and total cost of ownership make economic sense.
Broadcom is well positioned to benefit from that trend.
However, the same questions apply: Are customers committing to production deployments? Are the chips being delivered on schedule? Is AI semiconductor revenue growing because of repeatable demand rather than a small number of large projects?
Those answers matter more than a one-day share-price reaction.
What would prove that orders are genuinely shifting?
If I were monitoring AMD and Broadcom over the next few quarters, I would focus on four indicators.
1. AMD Data Center revenue and shipment growth
I want to see sustained growth from Instinct GPUs and EPYC processors, supported by actual customer deployments. Announced gigawatt-scale plans are encouraging, but delivery and revenue recognition are the real tests.
2. Broadcom AI semiconductor revenue
I would watch whether its custom accelerator and AI networking business continues to grow sequentially. The $21.7 billion Q4 AI semiconductor revenue outlook is an important benchmark against which future results can be measured.
3. Customer concentration and repeat orders
A few large customers can produce impressive growth, but they also create concentration risk. A healthy supplier-diversification cycle should eventually show a broader and more repeatable customer base.
4. Margins and cash generation
Revenue growth alone is not enough. I want to see whether AMD and Broadcom can deliver that growth profitably, without excessive spending, financing support or deteriorating economics.
My investment view: I would follow the orders, not the rotation
I remain interested in both AMD and Broadcom because they offer different ways to participate in the expanding AI infrastructure market.
AMD has the potential to gain from customers seeking an alternative GPU and server CPU platform. Broadcom benefits from the growing adoption of custom accelerators and high-speed networking.
But I would not assume that Nvidia’s leadership transition automatically creates a winner-takes-share opportunity for either company.
There is also a bigger risk that investors should not overlook: if AI infrastructure customers slow their spending, all three companies could be affected. Supplier diversification does not protect semiconductor companies from a broader reduction in AI capital expenditure.
For now, Nvidia’s latest results show that its underlying business remains exceptionally strong. AMD and Broadcom need to demonstrate their own ability to capture incremental demand, rather than relying on uncertainty surrounding a competitor.
My conclusion is simple: I would rather see two or three consecutive quarters of stronger shipments, rising AI-related revenue and healthy cash generation than buy a stock simply because investors believe Nvidia’s market share might eventually decline.
The AI supplier rotation becomes investable when it appears in the financial statements, not just in the share-price movements.
$Advanced Micro Devices(AMD)$ $Broadcom(AVGO)$ $NVIDIA(NVDA)$ $Taiwan Semiconductor Manufacturing(TSM)$ $VanEck Semiconductor ETF(SMH)$
Comments