10 Stocks to Buy as Investors Look Beyond NVIDIA

For three years, the AI trade has been looking up to one face too often. Now it has an interim CEO. On November 12, Jensen Huang resigned as chief executive and board member of one of the largest companies in the world, hours after nerves gripped investors awaiting the outcome of the board’s crisis meeting. And on an interim basis, CFO Colette Kress is running the chipmaker while the board searches for a permanent successor.

The board opened an internal investigation on November 8 after reports that U.S. intelligence had probed the CEO’s ties to a Singapore data center executive suspected of funneling chips into China. A second reported affair, this time with a senior employee, followed. By the morning of November 12, the shares were down about 5% in premarket trading. Analysts were also cutting fast. Bernstein downgraded the stock to Market Perform and halved its target to $200, Baird cut it to Neutral with a $200 target from $500, and Morgan Stanley dropped it from its Top Pick list. Senator Elizabeth Warren’s office has also asked the administration to preserve all records on the company’s chip exports to China.

At about 11x forward earnings, the giant chipmaker trades far below the wider market’s roughly 19–20x, even after 106% revenue growth last quarter. A multiple that low reflects the market’s doubts about the sustainability of earnings. AI hardware is cyclical, margins above 60% invite competition, and a chipmaker trading below 15x forward earnings shows that investors are quickly pricing in a peak.

That is why money is spreading across the ecosystem: custom accelerators, the networking that ties clusters together, the clouds renting out compute, and the equipment that fills data halls. On the day Huang resigned, the leading GPU challenger rose 5.9%, and the biggest custom-chip designer gained more than 5%. For a cheaper entry point, see the cheapest AI data center stocks billionaires own, or the stocks Jim Cramer is bullish on amid the Jensen Huang controversy.

10 Stocks to Buy as Investors Look Beyond NVIDIA

Our Methodology

We started by filtering U.S.-listed companies that sell products or services that compete with, or substitute for, the AI chip giant’s. This covers alternative AI chips (GPUs and custom accelerators), AI networking, AI cloud services, and data center equipment. We only shortlisted companies with a market cap above $40 billion and at least 70% Buy ratings from analysts. We then selected the 10 largest by market cap and ranked them based on analyst upside as of November 13, 2026.

10. Microsoft Corporation (NASDAQ:MSFT)

Median Upside Potential: 21.21%

Microsoft is NVIDIA’s biggest customer, turning into one of its quiet competitors. Azure sells AI capacity by the hour, and the customer renting it rarely cares whose chip sits underneath. That gives Microsoft room to fill its data centers with its own Maia accelerators and AMD’s new Helios racks, as well as NVIDIA hardware.

The demand behind that strategy is visible in the fiscal Q1 report from October 28. Azure grew 48%, revenue of $92.3 billion beat estimates, and commercial remaining performance obligations jumped 73% to $677 billion. That is a backlog of contracted cloud demand Microsoft can route to whichever silicon gives it the best economics.

The risk is that the build-out runs ahead of the payoff. Capital spending more than doubled to about $39.4 billion in the quarter, and Microsoft still buys heavily from NVIDIA, so it is not yet a clean break. The stock has also lagged the market over the past year.

For investors who want AI exposure without picking a chip winner, Microsoft’s discipline on that spending is the story, and two well-known investors have taken notice.

9. Advanced Micro Devices, Inc. (NASDAQ:AMD)

Median Upside Potential: 21.71%

AMD is the only other company selling AI GPUs at scale, which makes it the first call for any customer that wants a second supplier. With NVIDIA’s leadership in flux, that pitch got easier this week: AMD shares jumped 5.9% on November 12, the day Huang resigned, before pulling back the next day.

The November 3 report showed customers were already moving. Revenue rose 45% to a record $13.4 billion, data center sales nearly doubled to $8.5 billion, and AMD guided to about $17 billion for Q4. More telling for the alternative story, OpenAI, Anthropic, Meta, Microsoft, and Oracle are already running workloads on AMD’s first Helios rack-scale systems, built on its MI450 GPUs. Those are full-rack systems, the format NVIDIA has used to lock in its biggest customers.

The hurdle is software and execution. NVIDIA’s CUDA ecosystem keeps developers loyal, and rack-scale systems are harder to ship on time than individual chips. After a run of more than 120% over the past year, the stock also has less room for missteps.

Whether this data center surge becomes lasting profit is the key question, and one we explored in depth here.

8. Amazon.com, Inc. (NASDAQ:AMZN)

Median Upside Potential: 23.75%

Amazon’s answer to NVIDIA is to make its own chips. Its Trainium accelerators let AWS offer AI training at a lower cost than renting NVIDIA systems, and Amazon now plans to sell Trainium outside AWS starting in late 2027. That would put it in direct competition with NVIDIA for the same customers.

The October 29 report showed the plan gaining traction. Revenue commitments for Trainium rose more than 20x year over year, and capacity on older Trainium instances is selling at a premium. AWS revenue grew 39% to $46 billion, its backlog reached $578 billion, and the unit produced $17.8 billion of operating income.

That ambition costs cash. Amazon.com, Inc. (NASDAQ:AMZN) expects about $220 billion of capital spending in 2026, and trailing free cash flow has swung to roughly negative $42 billion. For now, Amazon also remains a major NVIDIA buyer.

If custom silicon closes that gap, AWS keeps more of each AI dollar instead of passing it to a chip vendor. Whether it can do so fast enough is the crux, which we break down here.

7. Arista Networks, Inc. (NYSE:ANET)

Median Upside Potential: 24.18%

NVIDIA does more than just sell chips. It sells the networking that connects them, and that bundle is a key part of its lock-in. Arista breaks that bundle. Its Ethernet switches work with any accelerator, so a customer mixing NVIDIA, AMD, and custom chips can run them all on one open network.

Strong numbers back that pitch. On November 3, Arista reported Q3 revenue of $3.66 billion, up 58.5%, its fastest growth since going public. EPS of $1.20 beat the $1.08 estimate, and Q4 guidance of $3.9 billion to $4.0 billion came in above expectations. CEO Jayshree Ullal called it the company’s highest growth since its IPO, with record profitability.

The risk is that NVIDIA fights back with its own Ethernet products, and that a handful of hyperscalers drive most of Arista’s demand. Gross margin also dipped to 63.6% from 64.6%, reflecting the high costs the company faces when competing for AI deals.

Still, the more customers diversify their chips, the more they need networking that isn’t tied to one vendor. How much margin Arista can give up to win those deals is the debate, and here is how its profit cushion compares with a fellow AI infrastructure name.

6. Alphabet Inc. (NASDAQ:GOOGL)

Median Upside Potential: 30.17%

Alphabet may be the most complete NVIDIA alternative on this list. The company designs its own TPUs, rents them through Google Cloud, and runs its Gemini models on them, so it controls the full stack from chip to application. Increasingly, it is also selling that TPU capacity to outside customers.

The October 28 report showed how fast that is working. Google Cloud revenue more than doubled, up 103% to $30.7 billion, and total revenue rose 25% to $127.6 billion. CEO Sundar Pichai said Alphabet is seeing strong demand for its TPU systems “across a wide range of customers,” evidence that companies outside Google are choosing its chips over NVIDIA’s.

The trade-off is spending. Capex hit $51.4 billion in the quarter, full-year guidance rose to $195 billion to $205 billion, and Alphabet has signaled 2027 spending above $300 billion. Net income also fell year over year, to $26.2 billion.

Yet the stock still trades at roughly the market’s multiple, which suggests investors are pricing it as a search company rather than a chip-and-cloud rival to NVIDIA. That low multiple can be misleading, as our closer look at its valuation shows.

5. Broadcom Inc. (NASDAQ:AVGO)

Median Upside Potential: 51.29%

If the biggest AI buyers are designing their way off NVIDIA GPUs, Broadcom is usually the company turning those designs into silicon. Its custom accelerators let hyperscalers and AI labs build chips tuned to their own models, and its shares rose more than 5% on November 12 as investors looked for NVIDIA alternatives.

The growth is already large. AI semiconductor revenue jumped 221% to $16.7 billion in fiscal Q3, and Broadcom guided fiscal Q4 revenue up 93% to $34.8 billion. Management expects about $115 billion of AI revenue in fiscal 2027 and $230 billion in fiscal 2028, and CEO Hock Tan has said Anthropic is on track to become Broadcom’s largest custom-chip customer in 2027.

That concentration is also the risk. On November 13, Rothschild & Co Redburn reportedly cut the stock to Neutral, arguing that the 2028 forecast leans heavily on Anthropic and OpenAI, two customers that depend on access to credit markets. Gross margin is guided down to about 73% from 78%, as custom chips earn less than Broadcom’s traditional products.

Broadcom’s role as an NVIDIA alternative is clear. The open question is who keeps paying for those chips, a question we dug into recently.

4. Vertiv Holdings Co (NYSE:VRT)

Median Upside Potential: 63.56%

Vertiv doesn’t compete with NVIDIA’s chips; it makes the power and cooling systems that every AI rack needs, whoever designs the accelerator inside. As customers spread their spending across GPUs, TPUs and custom chips, Vertiv gets paid either way.

Its October 28 report showed how much that neutrality is worth right now. Q3 revenue of $4.31 billion and adjusted EPS of $2.21 both beat estimates, and organic orders more than doubled year over year. Vertiv raised its 2026 guidance to $14.45 billion to $14.85 billion in revenue and $7.10 to $7.20 in adjusted EPS.

The threat is political, not competitive. The stock fell 13% on November 4 and another 16% on November 5 after the midterm results raised fears of limits on data center construction. Texas’s governor-elect campaigned on pausing new projects, and Mississippi’s senator-elect has pledged to push for a pause until operators pay for their own power and water. Vertiv’s equipment can be shipped to other states, but order timing could slip.

If projects get delayed, the question is how much of Vertiv’s cash flow depends on customer prepayments, which we examined here.

3. Dell Technologies Inc. (NYSE:DELL)

Median Upside Potential: 74.15%

Dell builds the AI servers that put accelerators to work, and it offers systems built on both NVIDIA and AMD chips. That lets investors own the AI build-out without betting on a chip winner, and it gains if customers spread orders across vendors.

The scale of that demand is clear from its last report. Dell posted record AI orders of $60.9 billion and an AI backlog of $95 billion, and raised full-year revenue guidance by $25 billion to $192 billion, including about $74 billion of AI server revenue. On November 24, it reports against guidance of $49 billion in revenue and $6.50 in adjusted EPS, the next test of whether that backlog turns into cash.

The weak point is margins. Dell’s trailing gross margin is under 20%, so it keeps far less of each AI dollar than the chip designers, and it has little pricing power if component costs rise. The stock also fell sharply in the November 4 to 5 data center selloff.

Dell won’t capture the profits of an NVIDIA alternative, but it ships every one of them. For more names across the server and power stack, see our list of AI infrastructure picks for the long haul.

2. Oracle Corporation (NYSE:ORCL)

Median Upside Potential: 91.46%

Oracle has turned itself into one of the largest suppliers of rented AI compute. The company is building this business with more than one chip vendor. It is among the first operators of AMD’s Helios racks, which gives its customers an alternative to NVIDIA systems inside Oracle Cloud Infrastructure.

The demand is enormous. OCI revenue grew 121% in fiscal Q1, while remaining performance obligations reached $664 billion. Management is guiding to fiscal 2027 revenue of at least $90 billion and non-GAAP EPS of $8.10. Few cloud providers have that much contracted AI capacity to fill.

The problem is paying for it. Q1 capex of $28.5 billion exceeded operating cash flow, free cash flow was negative $5.4 billion, and interest expense rose 55%. The company has given no timeline for returning to positive free cash flow. Much of the backlog is also tied to OpenAI. Those worries explain why the shares are down about 41% over the past year.

That gap between the backlog and the share price is why Oracle ranks near the top of this list. Its December 14 report is the next test of whether it can pay for the buildout it has promised.

1. Marvell Technology, Inc. (NASDAQ:MRVL)

Median Upside Potential: 92.33%

Marvell is the second major designer of custom AI chips, and it sells the optical links that move data between them. That makes it a double play on hyperscalers building their own silicon instead of buying NVIDIA’s. The stock rose 9.4% on November 11 as the NVIDIA news broke.

Management laid out the scale of that opportunity at its October 6 investor day: a path to $70 billion to $90 billion of revenue by fiscal 2031, about $30 billion of it from custom silicon, across more than 35 chip sockets at the top four U.S. hyperscalers. An expanded partnership with Google includes a warrant tied to revenue milestones. In its last quarter, data center sales rose 46% and made up 79% of revenue.

The bears argue that those targets are years away, and custom chips usually carry lower margins than NVIDIA’s merchant GPUs. Marvell also depends on its customers’ chip programs staying on schedule. Furthermore, after nearly doubling in a year, the stock leaves little room for delays.

Its early-December report will show how fast those wins become revenue. Google’s next AI chip is said to be running late, though that may matter less for Marvell than it seems.

READ NEXT: 10 Stocks Defying Tech Market Turmoil and Sell Alert: 10 Stocks Analysts Downgraded on Wednesday

Follow Insider Monkey on Google News.