artificial

The artificial intelligence (AI) megatrend has dominated the stock market over the past three years, and with massive AI spending projections stretching out through 2030 and beyond, that doesn’t look likely to change anytime soon. Two of the most successful stock picks in that part of the tech sector in recent years have been Palantir (PLTR 0.11%) and Nvidia (NVDA 0.86%). They aren’t competitors, as they operate on different sides of the AI value chain. Nvidia is a largely hardware provider, while Palantir makes software.

Both have made their long-term investors a ton of money, but the question is, which one provides the better investment opportunity from here? Let’s break that question down and consider it by category.  

Business model

Nvidia makes the world’s leading graphics processing units (GPUs), and its wares are by far the most popular parallel-processor chips for powering and training AI models. Nvidia has enjoyed solid market dominance over the past few years, but rising competition from AMD (NASDAQ: AMD) and Broadcom (NASDAQ: AVGO) could challenge that. Additionally, the AI infrastructure buildout won’t last forever. There will eventually be a time when companies aren’t racing to add high volumes of computing capacity, but instead mostly replacing processors as they reach the end of their useful lifespans.

Palantir sells its customers artificial intelligence-powered data analytics platforms, and it has a large client base in both the commercial and government spaces. Palantir’s software enables those with decision-making authority to act quickly and with the most up-to-date information possible. Furthermore, it also offers automation tools that can task AI agents with jobs that humans have traditionally done, freeing up employees to do work that requires original thinking.

Even after the initial stages of the AI revolution are over, the use cases for AI will continue to grow. Palantir’s software is also a subscription service, so for customers to continue using it, they must pay their Palantir bills every year, while data center operators may be able to put off replacing their Nvidia GPUs for a while. This makes Palantir’s business model more sustainable, giving it the edge here.

Winner: Palantir

Growth rates

Both companies are growing at similar rates, although Nvidia’s sales have decelerated on a percentage basis, while Palantir’s continue to gradually accelerate.

NVDA Revenue (Quarterly YoY Growth) data by YCharts.

Whether Palantir will take the lead on growth or not, only time will tell, but with massive demand for AI services still out there, each will likely maintain a relatively rapid growth rate for the foreseeable future, leading me to view them as fairly evenly matched by this criterion.

Winner: Tie

Valuation

From a valuation standpoint, the comparison isn’t particularly close. Palantir’s stock has delivered incredible returns alongside Nvidia, but a large chunk of Palantir’s share price gains has come from its valuations rising to outsized levels, and that condition is not sustainable.

NVDA PE Ratio (Forward) data by YCharts.

Nvidia trades at a comparatively cheap 42 times forward expected earnings, while Palantir’s ratio tops 275. For comparison, if Nvidia had the same forward P/E as Palantir, it would be worth over $30 trillion, versus the $4.6 trillion it’s actually worth.

This shows that Palantir is overvalued. It will have to deliver years of sales and earnings growth to return the stock to a more reasonable level, even if it simply moves sideways from here. If we set Nvidia’s current valuation of 42 times forward earnings as that “more reasonable” level, and Palantir’s revenue rises at a 50% compound annual growth rate while it maintains a 35% profit margin, it would take over five years’ worth of growth to bring its P/E down to the target. And again, that assumes the stock doesn’t rise during those five years.

That means that Nvidia has basically a five-year head start on Palantir’s long-term stock performance. Given that AI spending is projected to grow rapidly over the next five years, Nvidia shares should easily outperform Palantir moving forward, as Palantir will spend most of the next five years growing into its already expensive valuation. Which is why, despite the two companies’ similar growth rates and Palantir’s more attractive business model, it’s not the AI stock I’d suggest buying now.

Overall winner: Nvidia

Artificial intelligence (AI) has become the megatrend of this decade and is fast transforming the enterprise landscape. According to Gartner, global AI spending will be nearly $1.5 trillion in calendar year 2025.

While the AI opportunity is massive, not every AI player can prove to be an exceptional business in the long run. Companies with proven technologies and well-established customer bases stand a better chance of sustaining high top-line and bottom-line growth rates in the coming years.

Here’s why Palantir Technologies (PLTR -0.31%) and Snowflake (SNOW 0.18%) are two companies that could deliver strong returns, turning disciplined investors into millionaires over the long run.

Image source: Getty Images.

Palantir

Palantir has evolved from a pure data analytics company to a full-stack AI enterprise platform. The company’s software solutions are now used in mission-critical operations by both government and commercial clients.

In the second quarter, the company’s revenue soared 48% year over year to over $1 billion. The U.S. continues to be the biggest market, with revenue growing 68% to $733 million. The company also closed a record $2.27 billion in total contract value, up 140% over the year-ago period.

A significant part of this growth is driven by the rapid enterprise adoption of the company’s Artificial Intelligence Platform (AIP). AIP combines large language model reasoning with the company’s proprietary ontology framework (used to relate physical assets to digital twins) to solve complex, real-time business challenges.

Palantir is also focused on helping clients scale through automation. The company has added new tools and features to AIP, such as AI Forward Deployed Engineer (software engineer) and AI Workbench, to automate application development tasks and develop, debug, and automate workflows.

The company has also introduced the Ontology-as-a-Code feature to enable clients to leverage ontology in their preferred integrated development environments, tools, and workflows.

Palantir’s shares are currently trading at a very aggressive valuation of over 123 times sales. Although not an ideal scenario, this premium reflects Wall Street’s confidence in the company’s future growth trajectory.

Analysts expect Palantir’s revenue to rise at a compound annual growth rate (CAGR) of 39.9% from $2.86 billion in fiscal 2024 to $11 billion in fiscal 2028. Adjusted earnings per share (EPS) are also expected to grow at a CAGR of 40.7% from $0.41 in fiscal 2024 to $1.61 in fiscal 2028. Hence, the valuation can continue to remain elevated for several more years.

Considering these factors, Palantir can prove to be a smart pick in 2025.

Snowflake

Snowflake is transitioning from a cloud data warehouse to an AI data cloud (unified platform comprising AI technologies, data, and applications) for enterprises.

In Q2 of fiscal 2026 (ended July 31, 2025), product revenue grew 32% year over year to $1.09 billion, while non-GAAP operating margin reached 11%. The company had $6.9 billion in remaining performance obligations (RPO) at the end of Q2, up 33% on a year-over-year basis. With a large base of renewing customers, contracted billings, and large deals in the pipeline, the company has strong revenue visibility for the next few years. Snowflake’s healthy net-revenue retention rate of 125% also demonstrates its success in cross-selling and upselling to existing clients.

AI has become the key growth engine, influencing almost half of all new customer wins in Q2. AI is also powering nearly 25% of the deployed use cases. Currently, over 6,100 accounts use Snowflake’s AI capabilities on a weekly basis for various activities such as data migrations, analytics, and workflow transformations.

Snowflake has further strengthened its position in enterprise AI with Snowflake Intelligence, which enables enterprises to interact directly with their data and also build intelligent agents. The company has introduced Cortex AI SQL, which enables users to leverage AI models directly within SQL databases. This removes the need to move data between applications and unifies analytics and AI.

The company is also committed to improving performance and efficiency. The company launched Gen2 data warehouses , which offer double the performance in extracting insights and managing data without increasing costs. The company’s new OpenFlow capability allows enterprises to bring unstructured, structured, batch, or real-time streaming data into the Snowflake platform.

All these AI-powered capabilities have accelerated the company’s customer acquisition pace. Snowflake added 533 new customers in Q2, including 15 Global 2000 companies. The company now is trading at 19.4 times sales, which is not cheap for a loss-making company. However, the premium seems justified when we consider its accelerating AI adoption, expanding customer base, and robust backlog.

Hence, the payoff in investing in Snowflake can be impressive despite its elevated valuation levels.

Microsoft (MSFT -0.15%) and Apple (AAPL 0.54%) are forever rivals. They are competitors in the personal computer market and for years have been the leading tech companies in the world. Even today, their valuations are similar. As of Tuesday’s close, Microsoft had a slight edge in market cap ($3.82 trillion versus $3.67 trillion).

There’s a new arena that could be their new battleground: artificial intelligence (AI). It’s still the early innings of AI deployment, and how their businesses evolve and adapt to AI remains a big question mark. But based on where they are today, which AI stock looks to be the better buy right now?

Image source: Getty Images.

Which company has the better overall growth prospects?

Both of these companies already have large, robust businesses that can benefit from AI. Apple is a big name in consumer electronics with its iPads and iPhones being highly coveted products and, in some cases, status symbols. Microsoft, meanwhile, has its core in the business world with companies all over using its office products and cloud software for their day-to-day operational needs. They also both sell personal computers, with Microsoft focusing more on practicality and real-world business use, while Apple’s focus has been on simplicity and ease of use for the average user.

They both have many opportunities where AI can enhance their existing products in services. But the edge for sheer growth potential has to go Microsoft, simply because of how much broader its business has become over the years, especially in gaming, with it wrapping up its massive $69 billion acquisition of Activision Blizzard a couple of years ago.

Which company will benefit the most from AI?

AI has tremendous potential applications for these businesses. Many Apple users have been eagerly awaiting the launch of new AI-powered features for the company’s iPhones and were disappointed when they learned many of the key ones related to Siri will be pushed back until next year.

When that happens, however, that could trigger a flurry of upgrades and growth for the business. I don’t think a slow-and-steady approach will necessarily hurt Apple. In fact, it could end up being a smart move for the tech company by taking its time and ensuring everything is rolled out smoothly, to ensure that user privacy is well protected in the process.

Microsoft has already been enhancing its products and services with AI capabilities. However, there’s been plenty of debate about just how successful its Copilot AI really is. Salesforce CEO Marc Benioff has referred to it as “Clippy 2.0,” in reference to the frustrating assistant that Microsoft had years ago that users didn’t find all that helpful.

Apple deserves an edge when it comes to AI potential, simply for the massive wave of upgrades that could be coming if it hits it out of the park with its new iPhone features.

Which stock has the more attractive valuation?

It’s always important to consider valuation when buying a stock, as buying at a high price may impact your ability to earn a good return from your investment in the future. For a while, Microsoft’s stock was trading at more of a premium to Apple’s stock, but in recent weeks, that gap has evaporated.

Data by YCharts.

This one is easy to decide: It’s a tie. Their price-to-earnings multiples are almost identical at this stage. But it is notable to see that prior to the announcement of reciprocal tariffs in April, it was Apple that was trading at more of a premium than Microsoft, and then the trend reversed, with Apple’s exposure to manufacturing its iPhones in China likely weighing down the stock for part of the year.

Which stock should you buy?

The stock I’d buy today is Apple. It has devoted fans who will be willing to upgrade to the newest iPhone, even under challenging economic conditions, if it means access to cutting-edge features. Apple may be slow in rolling out AI, but when it does, the execution can be a lot cleaner, polished, and better for users in the end than if it were rushed.

Microsoft, meanwhile, has been quick to rush out AI features for its software. However, in an increasingly crowded market for AI services, it may have a more difficult time keeping customers happy when there may be other, and potentially better, options to choose from.

Apple, may, in the end, benefit from being a bit slower in its AI deployment.

The Nasdaq Composite‘s brutal 3.6% sell-off on Oct. 10 was a painful reminder of how quickly growth stocks can sell off when doubt creeps in. Friday’s tumble marked the worst session since April during the height of trade tensions between the U.S. and China.

The sell-off was a reaction to the U.S. threatening an additional 100% tariff on Chinese imports as a retaliation for China’s stricter export controls on rare-earth minerals and magnets. These materials and products are used across economic sectors, including semiconductors and technological equipment with artificial intelligence (AI) applications.

On Oct. 12, reports indicated that China would not back down against escalated tariff threats from the U.S.

Investors often talk about buying opportunities when the market is selling off. But it can be just as helpful to be aware of potential mistakes and prevent them before they do damage to your portfolio. Here are three that apply to AI growth stock investors who are preparing for next year.

Image source: Getty Images.

1. Having an overly concentrated AI portfolio

A common mistake is to overly focus on one facet of a value chain.

For example, an investor may own Nvidia (NVDA -0.17%), Broadcom, and Advanced Micro Devices as a way to diversify across different AI chip designers. The issue is that many of these companies have the same customers. For example, OpenAI is buying chips from all three companies to build out 10 gigawatts of data centers. If OpenAI were to cut its spending, it could affect the earnings of all three companies.

Similarly, equipment suppliers like Applied Materials, Lam Research, and ASML all share the same largest customers — which are semiconductor manufacturers like Taiwan Semiconductor, Samsung Electronics, and Intel. So if Taiwan Semi cuts its spending, it would reduce earnings across the semiconductor equipment supplier industry.

Further down the value chain are the cloud computing giants like Amazon Web Services, Microsoft Azure, Alphabet-owned Google Cloud, and Oracle. These companies benefit from increased AI spending, but they also serve general computing and storage needs. A slowdown in AI spending, or a widespread economic downturn, could reduce demand for additional cloud computing usage across major corporations.

By building out an AI portfolio across the value chain rather than focusing on one segment, you can help reduce volatility and limit the damage of an industry-specific pullback.

2. Ignoring position sizing

Portfolio sizing and allocation are just as important as the stocks and exchange-traded funds owned. You don’t want to be so diversified that your best ideas don’t make a big impact, but you also don’t want to be overly concentrated to the point where a handful of stocks can damage your financial health.

There’s no one-size-fits-all solution to diversification. But factors to consider include investment goals, investment time horizon, and risk tolerance.

A risk-averse investor would probably want to limit the size of a single stock in their financial portfolio, whereas an investor with a high risk tolerance and a multi-decade time horizon may not mind betting big on a handful of stocks, especially if they are still making new contributions to their investment accounts.

3. Buying stocks and not companies

Building a diversified portfolio isn’t enough. In fact, it’s not even the most important factor.

Arguably, the greatest mistake investors can make when approaching AI is to invest in stocks rather than companies. In other words, focusing too much on price action and potential gains rather than on what a company does and where it could be headed.

Peter Lynch’s investment advice to “know what you own, and why you own it,” still rings true today. Without conviction, a concoction of emotion and volatility can corrode the foundations of even the strongest portfolios. An investor may hold positions in stocks just because they are going up, even if those gains are temporary, because they don’t have to do with the underlying investment thesis.

The best investments are the ones you can put a decent amount of your portfolio into and be confident in owning, even if they suffer an extreme sell-off — like we saw in April during the height of trade tensions. If someone bought Nvidia just to make a quick buck, they may have been tempted to sell it when it fell by over 37% from its high in early April. Or when it dropped over 55% from its high in 2022. But someone investing in Nvidia for its multi-decade potential in AI data centers would have had an easier time holding the stock throughout these volatile periods.

Unlocking lasting success in the stock market

Diversifying across the AI value chain in companies you understand and with an awareness of portfolio sizing can help you build a portfolio that’s built to last, rather than one that can get hot only if the conditions are right.

Long-term investors know that success is more about making consistently good decisions over an extended period, rather than a few great ideas wedged between mediocrity and mistakes.

AI stocks have generated monster returns for patient investors, and many have the potential to create lasting generational wealth going forward. But those gains could take time, with many bumps along the way.

No one knows when the next major stock market sell-off will occur. Instead of guessing the timing and severity of a sell-off, it’s better to put your effort into following great companies and limiting mistakes.

In sum, diversification, conviction, and good companies are components that can help you build an investment suspension system capable of absorbing sell-off shocks.

On October 14, 2025, CCLA Investment Management disclosed it had sold its entire position in NICE (NICE -1.26%) in an estimated $120.03 million transaction.

What Happened

According to a filing with the Securities and Exchange Commission dated October 14, 2025, CCLA Investment Management exited its holding in NICE by selling all 710,865 shares, with an estimated trade value of $120.03 million.

What Else to Know

CCLA Investment Management sold out of NICE, reducing its post-trade stake to zero; the position now represents 0% of 13F AUM.

Top holdings following the filing:

• NASDAQ:MSFT – $369.63 million (5.9% of AUM) as of September 30, 2025
• NASDAQ:GOOGL – $345.87 million (5.5% of AUM) as of September 30, 2025
• NASDAQ:AMZN – $269.0 million (4.3% of AUM) as of September 30, 2025
• NASDAQ:AVGO – $207.92 million (3.3% of AUM) as of September 30, 2025
• NYSE:V – $180.65 million (2.9% of AUM) as of September 30, 2025

As of October 13, 2025, shares of NICE were priced at $132.00, marking a 23.8% decrease over the year ended October 13, 2025. Over the same period, shares have underperformed the S&P 500 by 35.5 percentage points.

Company Overview

Company Snapshot

NICE Ltd. delivers AI-powered cloud software solutions designed to optimize customer experience and enhance compliance for enterprises and public sector organizations worldwide. The company leverages a broad portfolio of proprietary platforms and analytics tools to address complex business needs in digital transformation, financial crime prevention, and operational efficiency.

The company offers AI-driven cloud platforms for customer experience, financial crime prevention, analytics, and digital evidence management, including flagship products such as CXone, Enlighten, and X-Sight.

NICE Ltd. serves a global client base of enterprises, contact centers, financial institutions, and public safety agencies seeking advanced automation, compliance, and customer engagement solutions. It operates a subscription-based business model, generating revenue from cloud services, software licensing, and value-added solutions for enterprise and public sector clients.

Foolish Take

In a recent regulatory filing, CCLA Investment Management revealed that it has completely sold out of its ~$120 million position in NICE, an Israeli software company. This move comes following a tough period for NICE stock.

Over the last five years, the company’s stock has consistently underperformed the broader market. Shares have logged a total return of (44%) over this period, equating to a compound annual growth rate (CAGR) of (11%). This compares quite unfavorably to the S&P 500, which has generated a total return of 105% over the last five years, equating to a CAGR of 15%.

All that said, NICE’s stock performance doesn’t reflect its underlying fundamentals. Total revenue, net income, and free cash flow have all increased significantly over the last five years, indicating strength in the company’s business model, which relies on artificial intelligence (AI) to power applications serving contact centers, financial institutions, and public safety organizations. Moreover, the company recently announced plans to buy back up to $500 million worth of its outstanding shares, which could help put a floor under its share price.

While CCLA’s recent sale does indicate the deterioration of some institutional support, retail investors may want to take a look at NICE — an under-the-radar AI growth stock.

Glossary

13F reportable assets: Assets disclosed by institutional investment managers in quarterly SEC Form 13F filings.

AUM (Assets Under Management): The total market value of investments managed by a fund or investment firm on behalf of clients.

Quarterly average price: The average price of a security over a specific quarter, often used to estimate transaction values.

Post-trade stake: The number of shares or value held in a position after a trade is completed.

Flagship products: A company’s leading or most prominent products, often representing its brand or core offerings.

Cloud platforms: Online computing environments that provide scalable software and services over the internet.

Digital evidence management: Systems for storing, organizing, and analyzing electronic data used in investigations or compliance.

Financial crime prevention: Technologies and practices designed to detect and stop illegal financial activities, such as fraud or money laundering.

Compliance: Adhering to laws, regulations, and industry standards relevant to a business or sector.

TTM: The 12-month period ending with the most recent quarterly report.

Operational efficiency: The ability of a company to deliver products or services using minimal resources and costs.

Serve Robotics (SERV -0.55%) develops autonomous last-mile logistics solutions. It has a major deal with Uber Technologies (NYSE: UBER) that will see thousands of its latest robots deployed into the Uber Eats food delivery network. But this is more than just a commercial partnership, because Uber is also one of Serve’s largest shareholders.

Uber acquired a company called Postmates in 2020, and in 2021, it spun Postmates’ robotics division out into a new company that became Serve Robotics. Serve is still relatively small with a market capitalization of just $890 million, but at the time of this writing, its stock has soared by 67% over the past year alone.

Serve has identified an enormous addressable market for its delivery robots, so should investors join Uber and buy the stock?

Image source: Getty Images.

A potential $450 billion opportunity

Existing last-mile logistics networks are quite inefficient, because they rely on cars with human drivers to deliver relatively small commercial loads from restaurants and retail stores. Serve is betting those workloads will increasingly shift to autonomous robots and drones, creating a potential $450 billion opportunity by 2030.

Serve’s latest Gen 3 robots have achieved Level 4 autonomy, meaning they can safely operate on sidewalks in designated areas without any human intervention. This makes them ideal for transporting small food orders, which is why 2,500 restaurants in five U.S. cities have used them to make 100,000 deliveries since 2022.

The Gen 3 robots use Nvidia‘s Jetson Orin platform, which includes all of the computing hardware and artificial intelligence (AI) software they need to operate autonomously. Having such a powerful technology partner will help Serve scale as quickly as possible, which is key to bringing costs down to management’s target of just $1 per delivery. At that point, using robots will be substantially cheaper than using human drivers.

Serve has a contract with Uber Eats to deploy 2,000 robots across Los Angeles, Miami, Dallas, Atlanta, and Chicago before the end of 2025. The company rolled out its 1,000th robot on Oct. 6, meaning its capacity will double in just the next few months.

But it won’t stop there, because last week Serve announced a new multiyear deal with DoorDash, which operates the largest food delivery network in the U.S. The two companies are yet to provide firm numbers, so it’s unclear how many more robots Serve will have to deploy.

Scaling a robotics business is not cheap

Despite its status as a publicly traded company, Serve is still very much a start-up. Its revenue tends to be quite lumpy, which is typical when a product is in the early stages of commercialization. The company brought in just $642,000 in revenue during the second quarter of 2025 (ended June 30), which is a tiny amount relative to its $890 million market cap.

But Serve’s business could scale extremely quickly. Management thinks the company will generate up to $80 million in annual revenue once all 2,000 Gen 3 robots are up and running, which bodes well for 2026. Wall Street predicts Serve will generate $3.6 million in total revenue this year (according to Yahoo! Finance), so $80 million would be a monumental jump.

But so far, the road to commercialization has been paved with substantial losses. Serve lost $33.7 million on a generally accepted accounting principles (GAAP) basis during the first half of 2025, so it’s on track to exceed its 2024 loss of $39.2 million by a very wide margin. The company spent $16 million on research and development alone during the first half of this year, so based on its minuscule revenues, its losses are no surprise.

Serve had $183 million in cash on hand as of June 30, and it raised a further $100 million from investors in October, so it has enough cushion to sustain its losses for the next few years (assuming they don’t materially increase). However, if the company doesn’t chart a pathway to profitability by then, it might have to raise even more money, which will dilute existing shareholders.

As a result, there is a lot riding on the successful commercialization of Serve’s 2,000 Gen 3 robots.

Serve stock trades at a sky-high valuation, but is it a buy?

Serve stock is extremely expensive right now. Its price-to-sales (P/S) ratio is a mind-boggling 486, making it substantially more expensive than any other major AI stock. Palantir Technologies, which also trades at a sky-high valuation, looks cheap by comparison because its P/S ratio is 128. For some further perspective, Nvidia stock has a P/S ratio of just 27.

SERV PS Ratio data by YCharts

With that said, if we assume Serve will generate around $80 million in revenue next year, its forward P/S ratio is just 11. In other words, it almost looks like a bargain.

But investors can’t always rely on management’s guidance, especially in this case because it assumes a perfectly smooth transition to commercialization for the Gen 3 robot. As with any new product, there will probably be bumps in the road, and we simply don’t know if it will scale successfully.

As a result, investors might be better off waiting a few more quarters to see if the rollout of the robots actually translates into as much tangible revenue as management expects. If it doesn’t, Serve stock could suffer a sharp correction because of its current valuation.

Some of the biggest winners in the stock market over the last three years have been companies riding the rising wave of generative artificial intelligence.

Palantir (PLTR -5.39%), with its artificial intelligence platform, has seen its stock rise by over 2,000% in three years. Nvidia (NVDA -4.84%), the poster child for AI chipmakers, is up by more than 1,300% in the same period. And neo-cloud providers like Nebius Group (NBIS -2.37%) and CoreWeave (CRWV -3.32%) have soared by triple-digit percentages since their IPOs.

If you had invested $10,000 in any one of these big winners ahead of their surges, you’d be well on the way to having a million-dollar holding in the long term, even if they produce merely average returns from here on out. But identifying which companies will be a new technology’s big winners ahead of time is difficult. If it were easy, everyone would be rich.

If you’d like to profit from the ongoing growth of AI, you could put a little bit of money into a lot of different AI stocks, or you could buy an ETF that specializes in finding generative AI opportunities. That’s what the Roundhill Generative AI & Technology ETF (CHAT -5.03%) does. Investors who are still trying to strike it rich with generative AI stocks may find it a compelling alternative to attempting to pick individual AI stocks themselves.

Image source: Getty Images.

Looking under the hood

The Roundhill team is focused on building a portfolio of companies that are actively involved in the advancement of generative AI. Its holdings include companies developing their own large language models and generative AI tools, companies providing key infrastructure for training and inference, and software companies commercializing generative AI applications.

Since it’s an ETF, investors can see exactly what the fund holds. Here are the largest holdings in the portfolio as of this writing.

• Nvidia
• Alphabet
• Oracle
• Microsoft
• Meta Platforms
• Broadcom
• Tencent Holdings
• Alibaba Group Holdings
• ARM Holdings
• Amazon

There aren’t a lot of surprises in the list. Perhaps the biggest standout is Arm, which is relatively small compared to the other tech giants with large weightings in the portfolio. Still, its market cap comes in at a healthy $165 billion.

In total, the ETF holds 40 stocks and several currency hedges for foreign-issued shares as of this writing. That diversification gives it a good chance of holding a few companies that will be big winners from here, which may be all it takes to produce market-beating returns. Indeed, the portfolio includes some of the best-performing stocks of 2025, including Palantir.

Since its inception in 2023, the Roundhill Generative AI & Technology ETF has returned an impressive 148% compared to a 66% total return from the S&P 500. And that’s factoring in the drag of the ETF’s 0.75% expense ratio.

Could $10,000 invested make you a millionaire?

In order to turn $10,000 into $1 million, the ETF would have to increase in value 100-fold. That may be difficult, considering the current sizes of its top holdings.

Nearly one-third of the portfolio is invested in companies with market caps exceeding $1 trillion, and the larger a company becomes, the more raw growth it takes to move the needle on its size on a percentage basis. For Nvidia to grow by even 25% now would be the equivalent of creating a whole new trillion-dollar business. And while such growth is certainly possible for some of those megacap companies, there’s still a finite amount of money in the global economy.

Meanwhile, there are only a handful of relatively small businesses in the ETF’s portfolio that could reasonably be expected to multiply in size significantly.

Additionally, many stocks in the portfolio have high valuations. Palantir shares trade for a forward P/E ratio of 280. Nebius trades for 54 times expected sales. Even CoreWeave’s sales multiple of 12.5 looks expensive, given its reliance on debt to continue growing. That said, some of the best performers of the last few years also looked expensive a few years ago (including Palantir and Nvidia). Still, the expected return of stocks with such high valuations isn’t going to be as high as those offering more compelling values.

As such, it seems unlikely the Roundhill Generative AI & Technology ETF will produce returns strong enough to turn $10,000 into $1 million over a reasonable time frame. That doesn’t mean that it’s not worth owning. For investors looking to gain exposure to the generative AI trend without going all in on one or two stocks, buying the Roundhill Generative AI & Technology ETF is a simple way to do that.