Initial public offerings, or IPOs, can generate a lot of excitement, especially when there hasn't been a public offering in a long time. And last week's debut of Arm Holdings (ARM 1.53%) had that in spades. After launching at $51 per share on Thursday, amounting to around a $53 billion market cap, Arm rocketed on its first day of trading. And while Friday saw a sell-off -- true across the entire tech sector -- the stock still finished the week 19% higher.

However, buying an IPO can be pretty risky. After all, when a company goes public, existing shareholders are cashing out, meaning both the company and its investment bankers have incentives to market the company to public shareholders -- that means you! -- at a high price.

Looking through Arm's registration document shows that there weren't one, but numerous red flags that stood out. Here are five factors that should keep new investors far, far away from the stock.

A nosebleed valuation

First, the company's IPO valuation of over 100 times earnings and nearly 20 times sales is not exactly cheap. In fact, that type of valuation is usually reserved for only the highest-growth, widest-moat companies.

For reference, Arm's P/E ratio is basically the same as that of Nvidia (NVDA 3.56%), which grew revenue a whopping 101% last quarter. And its price-to-sales ratio is in line with the likes of Cloudflare, an asset-light software company that grew 32% last quarter.

But Arm isn't growing like these companies are. In fact, in its recent fiscal year ended in March, revenue slightly declined.

Of course, that's not surprising given the down-year in the cyclical semiconductor industry. And it is true that chipmakers might have pulled forward their licensing activity in fiscal 2022, before Nvidia was supposed to acquire the company; the FTC scuttled that deal. Still, CFO Jason Child remarked in an interview that investors should look at the past three years' average growth rate, which was around 15%, as indicative of Arm's "normalized" growth today.  That's still not exactly eye-opening growth that would warrant such a high price.

It's true that Arm is an asset-light, profitable company that earns licensing revenue, and therefore doesn't need to invest a lot in capital expenditures. But if revenue can ebb and flow with the volatile and cyclical semiconductor sector, it's not at all clear that Arm's financials aren't as risky as those of other fabless chip designers in the sector. And those companies don't go for nearly Arm's valuation, unless you're posting the crazy growth rates Nvidia is achieving.

How strong is the moat?

Arm also may seem to have a long-lived moat. After all, for CPU processors, there are really two alternatives today: x86 architecture or Arm. The x86 architecture is used by processor companies including Intel (INTC 0.35%) and Advanced Micro Devices (AMD 1.00%), and it's generally seen as a more powerful architecture. On the other hand, Arm is known for its power efficiency. With more mobile phones and other edge devices needing that low-power computing, Arm has grown a lot over the past decade.

In fact, Arm-based semiconductors seem as if they're taking market share. In 2022, Arm had 48.9% market share for relevant processors, up from 42.3% in 2021.

However, part of that gain could be due to the movement of end markets. As most know, the PC market, which is dominated by x86 processors, had its worst downturn in history in 2022. Therefore, part of that "market share gain" could just be the result of end-market movements, and not necessarily a sign that Arm is gaining that much traction. Arm right now dominates the smartphone business with basically 100% market share. That doesn't leave any more room for share gains, and the smartphone market is plateauing.

At its valuation, investors may also be anticipating that Arm will be able to raise its licensing rates. Last year, Arm disclosed that it made 1.7% average royalties for all Arm-based chip sales. That doesn't seem like much, and it's possible Arm could raise prices down the road, further justifying its valuation.

However, some large chipmakers are now looking to develop an open-source alternative to Arm, called RISC-V, especially for non-smartphone devices. In its registration form, Arm noted, "in August 2023, a group of our customers and other competitors announced a joint venture aimed at accelerating the adoption of RISC-V." And earlier this week, analyst Charles Shi of Needham noted that RISC-V has been closing the performance gap with Arm for non-smartphone applications.

RISC-V poses a definite threat. Even if it doesn't gain much share, it could keep Arm honest about raising its license fees too much.

Female technician holds a square microchip in gloved hand.

ARM's high IPO valuation leaves much to be desired. Image source: Getty Images.

It bills itself as an AI company, but it's really not

Some may think that because Nvidia tried to buy Arm, and that Nvidia is a customer of Arm, that Arm is an artificial intelligence play. But Arm doesn't earn licensing revenue from Nvidia's GPUs, which run on their own proprietary architecture. That's really where the lion's share of AI growth is.

Arm estimates that it had about 10.1% market share in the cloud computing vertical in 2022, up from 7.2% in 2020, which is where a lot of AI computing will happen. But remember, this is primarily for CPUs.

It's true that Nvidia uses Arm architecture for its new Grace CPU in the data center, but that is a relatively small part of the Nvidia supercomputing systems. It's also unclear whether customers will adopt lots of Grace chips or mix and match with current Intel and AMD CPUs that are the current standard. AMD has been innovating impressively on the leading edge, and Intel is catching up fast, with a data center roadmap that will develop x86 chips either maximized for performance or power efficiency. The upcoming power-efficient x86 processors could make Intel more competitive with Arm in applications where low power is needed.

Moreover, according to a recent Bernstein report, Arm-based CPUs are dominated by just one customer: Amazon, whose Graviton processors reportedly account for more than 50% of ARM-based data center CPUs. While Graviton has grown impressively thanks to Amazon's lead in the cloud, that's a lot of customer concentration. Meanwhile, a significant portion of Arm's other data-center CPUs are produced by Chinese chipmakers. Given the evolving sanctions regime on China, it's unclear whether those Chinese chipmakers will be allowed to progress their roadmaps.

In any case, Arm should be able to grow its data center footprint, but it is highly unlikely to see the type of growth other more pure "AI" companies are seeing.

Like other semis, it has China issues... but even more so

Speaking of China, another red flag is Arm's China exposure, and arrangement with, subsidiary Arm China.

Arm got 25% of its revenue from China last year, which comes mostly from licensing revenue from its subsidiary, Arm China. But Arm China is mostly independent, with Arm having just a 4.8% nonvoting stake in the subsidiary. So it's as if Arm gets licensing revenue from a separate Chinese company with obscure financials.

The registration document notes that Arm has, in the past, has "had issues obtaining timely and accurate information from Arm China." While the company suggests these past issues have been resolved, those who have followed Chinese businesses know that bad actors can occasionally manipulate or obscure the numbers. In fact, in 2022, Arm China's former CEO sued it over corporate governance practices.

While it's possible China will continue to grow Arm-based CPUs rapidly, with a questionable arm's-length (forgive the pun) arrangement with the main Arm business, that 25% of revenue is more at risk, especially given growing U.S.-China tensions.

Softbank is selling you goods, but still in control

Finally, investors should understand that when a company goes public, it's selling for some sort of reason. In this case, the seller is Softbank (SFTB.Y 0.86%), the Japanese conglomerate that runs the giant Vision Fund that invests in new technology companies. However, after piling into lots of tech companies at the top of the tech market, many of Softbank's bets, such as the infamous WeWork, has gone bad. So Softbank is now looking to raise cash.

Why? Well, in a recent interview, Softbank CEO Masayoshi Son said Softbank will be looking to get into more AI-related investments. But if Arm is supposed to be an AI play, why would Softbank be selling shares?

In addition, following the IPO, Softbank will still retain about a 90% stake. That means public shareholders will have little say in how Arm is governed. And if Softbank seeks to raise more cash in the near future, it could very well sell more Arm stock. That could put an overhang on the stock and limit further gains -- especially at this valuation.

Hold off for a cheaper valuation

Arm certainly has some positive attributes, as the world looks toward more energy-efficient computing solutions in data centers and at the edge. However, at this high of a valuation, Arm stock isn't worth your investment dollars given these myriad risks. That's especially true when there are much, much cheaper semiconductor stocks with similar prospects.