The fundamentals of investing are the same for newcomers as they are for established, successful investors. The challenge for everybody is to stick to what works and not let investment fads, the media limelight or bad advice from a broker or advisor steer you off course.

Artificial intelligence (AI) is an example of an investment idea that could boost your investment returns—or more likely end of costing you money.

Here’s at what AI is, and how it’s evolving. Later we let you know what stocks to buy to profit from AI—with the least amount of risk. We hope you enjoy and profit from this report.

Pat McKeough

Artificial intelligence (AI) is essentially the merging of today’s big computing with big data. This has resulted in breakthroughs in everything from creating machines that can recognize faces with more accuracy than a human, to building cars capable of driving themselves.

The concept fascinates a lot of people, especially investors. They visualize AI-equipped machines that will be able to, for example, carry on intelligent conversations without revealing they are less than human. No doubt AI will have great impact on our lives as time passes. But we think many people today overestimate the investor potential of AI technology.

Computers beat humans at games like chess because computers are inherently digital, and so is chess—there’s no better word. In a chess game, only a relative handful of moves exist at any one time. After each move, new possibilities come to light, but they all follow the same basic rules.

AI has an advantage over human intelligence in chess, because it has the data storage and computing power to look further into the future, based on all possible situations that follow each move on the chessboard.

But that advantage is pretty much limited to highly specialized situations, with a limited number of rules and moving parts—like a chess game.

In real life, human intelligence has the advantage. After all, human intelligence works with analogies as well as digits. Human accomplishment rises out of learning the rules in a variety of situations, over the course of years or decades, then applying those rules in different situations, often totally unrelated. Much of this learning takes place subconsciously, but it’s there when you need it. Human achievement also requires emotional intelligence and intuition. High IQ alone helps if you want to excel at chess, but by itself it falls short in other areas of achievement.

In late January 2025, Chinese firm DeepSeek launched a new AI-powered chatbot that claimed to run on fewer and less-powerful chips than current AI chatbots, like ChatGPT (part of Microsoft’s Bing search engine).

A number of prominent AI stocks, such as Nvidia (see below), fell on the news. Investors feared DeepSeek could hurt demand for Nvidia’s chips as well as add uncertainty to other AI developments underway.

However, DeepSeek’s claims remain unverified. And even if it has developed a cheaper way to run AI programs, that should spur the long-term need for more and better chips. For example, Nvidia’s outlook remains bright because most current AI services are restricted to collating and presenting information. Nvidia’s latest chips, on the other hand, can better “think” through complex problems (a process called inference). The company estimates that new inference algorithms will require much more computing power and chips than current chatbots, including DeepSeek.

When new business ideas come along, many investors get over-excited. They see the growth potential right away, but it takes them longer to spot the negatives. As a result, they may underestimate how long it will take backers of the idea to launch a business and start making money.

On the other hand, some investors worry that a new idea may destroy the market for existing companies and products.

This idea has been around at least since the 1920s. At the 1939 World’s Fair, GM exhibited a model of a self-driving car. (Search on Google for “timeline self-driving cars” and you’ll get 50 million results in less than one second.)

In 2009, I went to a TED Conference. I met some interesting people, including a young couple who had quit their jobs so they could work as consultants to Google on a self-driving car concept. They were pleased to inform me that self-driving cars would soon go on sale to “early adopters.”

This made me wonder how long it would take for the first self-driving car/penny-stock promotion to turn up on the market. But as the concept gained more notice, it had the opposite effect on some investors. They worried that it would throw hordes of truck and taxi drivers out of work, push unemployment up and spark a depression. Over a period of a few years, I had to talk a number of our portfolio-management clients out of selling their stocks because of that fear.

I bring this up now since it seems that AI, or Artificial Intelligence, has also sparked this kind of strong but divided response.

Financial Physics can go a long way toward clarifying the outlook for investors in something like AI.

You might think of Financial Physics as a distant cousin of theoretical physics. Or, you could call it a branch of the study of human nature, applied in a financial setting.

Here’s one of the first Financial Physics principles I ever stumbled upon: It’s much easier to package a new investment venture/idea and sell it to the public than to launch a new business and transform it into a money-maker. That’s why investment opportunities/propositions are always more plentiful than successful new businesses.

That 2009 TED event had no presentation about self-driving cars, but it did show a 3-minute video clip about a forerunner of artificial intelligence. It was a recording of what we’d now call an audio chat feature on a website. This one was an automated telephone answering program that clients could use to make an appointment with their hairdresser.

The program was able to do the job, but somewhat awkwardly by current standards and near-painfully slow. Still, it could pass for an actual human being at the end of the line, and it impressed the audience. Some clapped; a few gave it a standing ovation.

Today’s AI-powered phone-answering systems are smoother and undoubtedly cheaper than employing a receptionist. But you wouldn’t say they’ve raised the user’s standard of living, nor cut the time it takes to book an appointment.

You should keep our Financial Physics rule in mind if you consider investment in AI or anything related to it.

Several years ago, Elon Musk called AI the world’s biggest threat to civilization. (Films like Terminator and Matrix could give you that impression, and Mr. Musk does have a vivid imagination.) Since then, however, he has gone on to invest in AI.

AI has since made some gains, mostly in communications. (In contrast, early adopters are still waiting for a licensed, insurable, road-worthy self-driving car.) You hear a lot about AI-related start-ups. Most seem aimed at improving existing devices and/or cutting business costs. Many have highly specific goals.

As you’ve probably noticed, a boom is underway in the investment-newsletter publishing business, at least in its “GRQ” segment. (GRQ is an acronym for Get Rich Quick.)

GRQ publishers sell newsletters and related products to subscribers. Their expertise is in newsletter marketing, not investing. Many publish numerous newsletters that may offer conflicting advice. When one publication puts out a stream of bad recommendations that drive off too many customers, the publishers change the publication’s name and/or investment specialty. That way, they always have one or more fresh titles that still have customer appeal and can operate at a profit.

GRQ publishing has been around for many decades, if not centuries. But it really went into high gear in the early 2000s. That’s when email began to replace postal mail as the main carrier for newsletter advertising, and costs began to plummet.

In the days of postal mail advertising, it cost a publisher perhaps $1 per “name” to offer a newsletter subscription to prospective customers. Publishers had to create, print and mail elaborate mailing pieces. They had to rent prospect names from direct competitors, or from other publishers in the same or related fields.

Compared to the costs of paper/postal mailings a decade or two ago, today’s costs of email advertising are close to negligible. Now publishers spend heavily in other areas—direct marketing consultants, specialized writers of advertising copy for email marketing, and so on.

Some newsletter publishers seem to be using AI to help them create email ads in ever larger numbers, to send to investors who never asked for them—spam, in other words.

In response, investors are using AI-based anti-spam programs to keep from drowning in junk mail. I rely on a Microsoft anti-spam filter that uses AI to sort through my incoming email. It diverts spammy-looking emails into a junk mail folder; if I leave them there for 30 days, it deletes them.

Every time I’ve checked my spam folder in the past few months, it held around 7,000 emails. That’s how many emails the Microsoft system identified as spam and diverted from my inbox in the previous 30 days. I glance through them when I think of it but never come across anything worth opening.

The funny thing is that the same spammy subject lines get used repeatedly in spam from seemingly unrelated publishers. (After some exposure to spam, you can recognize it on sight, just as you can recognize a foreign accent, even if you don’t speak the language.) Maybe they are operating out of the same shop, or using the same AI software or consultants.

It’s interesting (at least to me) that publishers use AI to create spam to raise the payoff from their marketing, while investors use AI to keep spam from wasting too much of their time.

As AI improves, it will undoubtedly add value in all sorts of ways and activities. For one, AI has military applications. Plans are underway to use AI to help a single highly skilled fighter pilot lead (you might say “fly” or “manage” or “orchestrate”) a squad of three to six top-of-the-line planes, rather than just one.

My guess is that AI will create much more value as an aid to human effort than a replacement for it.

In light of the attention that AI is attracting, you will keep coming across startups and IPOs that are trumpeting their plans for AI success.

But we think you should downplay or avoid most startups and IPOs. That’s especially so when they launch themselves into a field—AI—that’s in the early stages of long-term growth but is already well ensconced in the broker/media limelight.

Meanwhile, lots of venture-capital deals and junior stocks promise to unlock the financial potential of AI. Some may evolve into profitable companies, but most will stumble, as they always have in venture capital and junior stocks.

All in all, we think that the biggest, surest gains from AI will come from investing in established businesses that are already profitable and growing, and that can profit all the more by applying AI to their operations.

Here are stock picks among our recommendations that are already profitably taking advantage of AI; they are also leaders in extending AI’s use. Note all figures updated March 1, 2025.

MICROSOFT, symbol MSFT on Nasdaq (Shares outstanding: 7.4 billion; Market cap: $2.92 trillion; www.microsoft.com), is the world’s largest computer software firm. Its main product is the Windows operating system, which powers about 85% of the world’s personal computers. Its other main product—its Office suite, with its word processor (Word), spreadsheets (Excel) and slide presentations (PowerPoint)—controls over half of its market.

Microsoft is a leading investor in Open AI. That’s the firm behind the ChatGPT chatbot/search engine, which uses artificial intelligence (AI) software to interact with users in a conversational way and produce human-like written responses.

The company is now incorporating this AI technology into its cloud and other software products. Annual revenue from its AI businesses is now $13 billion, which is up 175% in the past year.

Microsoft is now using OpenAI to improve the quality of its products and services. It’s also applying its huge library of AI patents to develop new applications for its customers. These include using AI to power chatbots in Skype, analyze data in Office 365, build apps in its Azure cloud platform, support interaction with its Cortana voice assistant and match searches with results in its Bing search engine.

For instance, the company recently added AI capabilities to its Bing online search engine and its Azure cloud-computing platform. Next, Microsoft will incorporate the technology in its Office suite of programs. This could let users quickly create documents and spreadsheets by simply asking the software for what they need.

Microsoft reported better-than-expected results for its latest quarter, despite slower-than-expected growth at its Azure cloud computing business.

In its fiscal 2025 second quarter, ended December 31, 2024, Microsoft’s revenue rose 12.3%, to $69.63 billion from $62.02 billion a year earlier. That also topped the consensus forecast of $68.87 billion.

All three of Microsoft’s divisions reported revenue gains: cloud computing (up 18.7%); business productivity software (up 13.9%); and consumer products (up 0.1%).

Overall earnings in the quarter rose 10.2%, to $3.23 a share (or a total of $24.11 billion) from $2.93 a share (or $21.87 billion). That also beat the consensus estimate of $3.11.

For fiscal 2025, the company’s earnings will probably rise about 11% to $13.11 a share. The stock trades at 29.9 times that estimate. While high, that’s an acceptable p/e in light of the growth potential of its AI software and the company’s high research spending (11% of revenue). The $3.32 dividend yields 0.9%.

Microsoft is a buy.

ALPHABET INC., Nasdaq symbols GOOG [class C: non-voting] and GOOGL [class A: one vote per share], (Shares outstanding: 12.4 billion; Market cap: $2.06 trillion; www.abc.xyz), is the parent of Google, the world’s leading Internet search engine—it handles over 80% of global search requests. Online advertising revenue supplies the bulk of its revenue.

In addition to search, Google also offers a variety of other services and products. They include Android (operating system for mobile devices), Chrome (operating system and browser), Google Cloud, Google Maps, Google Play (media downloads), Hardware (including Nest thermometers and smoke alarms) and YouTube (online videos).

The company’s “Other Bets” division consist of several unrelated businesses that are still in the early stages of their development. They include Waymo (self-driving cars) and Calico (medical research and drug development).

Google is now incorporating AI into its various products and service. For example, it continues to roll out its new artificial intelligence (AI) software called Gemini. Gemini can recognize and understand text, images, audio, and other data simultaneously. That lets it better understand and answer complex questions.

The software can also create images based on textual descriptions. Despite some problems, such as inaccurate depictions of historical figures, this technology will undoubtedly continue to improve. In fact, it could someday let anyone create long-form videos and movies without actors and cameras.

The system is different from Google’s regular search engine, as it uses AI technology to answer questions in a conversational style by analyzing vast amounts of data from the Internet. As Gemini processes more feedback, the quality and accuracy of its responses should improve over time.

Google plans to eventually use Gemini to improve the performance of its search engine and other services, such as online email and documents. That will help it compete with Microsoft Corp. (Nasdaq symbol MSFT), which is using the popular ChatGPT chatbot in its search and other services.

In the three months ended December 31, 2024, Alphabet’s revenue rose 11.8%, to $96.47 billion from $86.31 billion a year earlier. Even so, that missed the consensus forecast of $96.56 billion. That was mostly due to weaker-than-expected revenue growth at Google’s cloud computing business.

The Google Services division (about 87% of Alphabet’s total revenue) includes the search, YouTube websites and subscriptions operations. It reported 10.2% higher revenue in the quarter. That gain was driven by a solid 10.6% rise in advertising revenue to $72.46 billion (86% of Google Services’ total revenue).

Revenue at the cloud computing business (12% of Alphabet’s total revenue) jumped 30.1% to $11.96 billion. However, capacity could not keep up with demand, which is why revenue fell short of the $12.19 billion consensus forecast.

Revenue from the Other Bets division (1% of revenue) fell 39.1%.

Alphabet’s overall earnings in the quarter improved 28.3%, to $26.54 billion from $20.69 billion. Earnings per share rose at a faster pace of 31.1%, to $2.15 from $1.64 on fewer shares outstanding. That beat the consensus estimate of $2.13.

The company continues to invest heavily in new products, particularly AI software. Its research costs rose 8.3% in the quarter, to $13.12 billion (or 13.6% of revenue) from $12.11 billion (14.0% of revenue) a year earlier.

Alphabet’s long-term outlook remains bright, particularly as businesses continue to shift away from traditional print and TV ads to more-targeted online ads. Its cloud computing services are also gaining new customers, and Alphabet could unlock shareholder value by spinning off YouTube or other smaller businesses.

Meanwhile, the company’s plans to incorporate AI just add to its prospects. In fact, the company has announced that it will spend $75 billion in 2025 on new datacentres and other equipment to power its artificial intelligence (AI) programs, which is more than the $59 billion consensus forecast.

Alphabet’s new datacentres and AI products should help lift its earnings in 2025 by 12% to $9.04 a share, and the class A stock trades at a reasonable 18.7 times that forecast. The $0.80 dividend yields 0.5% for both the class A and C shares.

You should note that the voting A shares typically trade at a premium to the non-voting C shares. While that’s not the case now, we see that as likely, and the price difference could widen in the coming decades. That’s why we recommend you buy the voting shares.

Alphabet (class A voting) is a buy.

NVIDIA CORP., Nasdaq symbol NVDA (Shares outstanding: 24.5 billion; Market cap: $2.93 trillion; www.nvidia.com), is a leading designer of 3D-capable video chips; they make video games run more smoothly and appear more lifelike. Nvidia has also adapted its chips for other applications, including artificial intelligence (AI), datacentres and self-driving cars.

Nvidia is also in a strong position to benefit from the expansion of AI-powered chatbots like Google’s Gemini (see above).

For instance, the company has launched Nvidia DGX Cloud, a new Internet-based service that lets businesses access computers and software capable of running their AI programs.

So far, Microsoft and Oracle have agreed to use the service to improve the performance of their own cloud computing platforms. Nvidia DGX Cloud will also help biotech-giant Amgen speed up the development of new drugs and help insurance technology firm CCC Intelligent Solutions improve the settlement of claims.

The new service will help Nvidia earn recurring revenue and cut its reliance on cyclical demand for its hardware products.

Meanwhile, Nvidia’s stock has soared over 621% since the start of 2023.

That’s largely due to strong demand for its chips that power AI applications. Those applications include the popular ChatGPT online chatbot/search engine.

Nvidia split its shares on a 10-for-1 basis effective on June 10, 2024. That makes the stock seem more affordable for individual investors and the company’s employees.

In its fiscal 2025 fourth quarter, ended January 26, 2025, Nvidia’s revenue soared 77.9%, to $39.33 billion from $22.10 billion a year earlier. Earnings before unusual items also jumped 71.9% in the quarter, to $0.89 a share (or a total of $22.07 billion) from $0.52 a share (or $12.84 billion). That also beat the consensus estimate of $0.56.

Nvidia continues to spend heavily on the development of new products. In the latest quarter, its research costs rose 50.7% to $3.71 billion, or a high 9.4% of its revenue.

Thanks to that spending, the company launched its Blackwell AI platform and chips, which use 25% less energy than its preceding chips. Major computing firms such as Alphabet (Google), Amazon.com, Microsoft, Dell, Oracle and Meta Platforms now plan to use the Blackwell platform to run their AI applications.

The company also recently launched its new H100 graphics chips for datacentres, which are about 9 times faster than the preceding model. They are also up to 30 times faster at processing AI applications.

The company’s earnings are forecast at $4.45 a share for the fiscal year ending January 31, 2026. The stock trades at 27.5 times that 2026 estimate. While that’s a high multiple, the company has a clear technological advantage over most of its competitors, and its big commitment to research (about 10% of revenue) will help it maintain that lead. The $0.04 dividend yields 0.03%.

Nvidia is a buy, but only for highly aggressive investors.

IBM, symbol IBM on New York (Shares outstanding: 927.3 million; Market cap: $234.8 billion; www.ibm.com), is one of the world’s largest computer companies, with operations in over 175 countries.

In the past few years, IBM has shifted its focus to its more-profitable cloud computing, consulting and mainframe businesses. It now gets over 75% of its revenue from its software and consulting.

Meantime, IBM has one of the largest and most well-funded AI research programs in the world. As part of that, it has developed its Watsonx AI platform.

Watsonx has three key applications:

Watsonx.ai is a new studio that provides a comprehensive set of tools to create new foundation models, generative AI and machine learning. It offers increased productivity compared to traditional approaches, enabling users to work with AI models efficiently.

The Watsonx.data lake house combines a data lake’s flexibility and a data warehouse’s performance. This component is a centralized repository to store various data types, including structured, unstructured, semi-structured and multimodal data.

The Watsonx.goverance toolkit enables AI workflows with responsibility and transparency. It helps users track the data used to train models, understand model lineage, identify biases and monitor model drift. Enterprises can ensure reliable and accountable AI deployments by consolidating governance processes into a single platform.

The company reported better-than-expected results for its latest quarter. That’s mainly due to strong demand for its artificial intelligence-related software and consulting services.

In the three months ended December 31, 2024, revenue rose 1.0%, to $17.55 billion from $17.38 billion a year earlier. That topped the consensus forecast of $17.54 billion. If you exclude exchange rates, revenue in the quarter rose 2%.

Earnings before unusual items also improved 2.9%, to $3.69 billion from $3.59 billion. Due to more shares outstanding, per-share earnings rose at a slower pace of 1.3%, to $3.92 from $3.87. That also beat the consensus estimate of $3.78.

IBM now expects its overall revenue (excluding currency rate movements) will rise at least 5% in 2025.

This year’s earnings should also rise about 4% to $10.75 a share, and the stock trades at 23.3 times that estimate. That’s a reasonable multiple in light of the company’s high research costs (12% of revenue) and AI expertise. The $6.68 dividend yields 2.7%.

IBM is a buy.

PALANTIR TECHNOLOGIES INC., symbol PLTR on New York (Shares outstanding: 2.2 billion; Market cap: $198.8 billion; www.palantir.com), is a developer of data-mining software platforms for both government and private industry. It was founded in 2003 to build software for the U.S. intelligence community. The CIA was one of its original financial supporters. The name Palantir is a reference to magical stones in J.R.R. Tolkien’s Lord of the Rings.

The company was co-founded by prominent tech investor Peter Thiel, who also co-founded PayPal Holdings. Thiel was also one of Facebook’s first major investors.

Palantir began trading on New York on September 30, 2020.

The company has four primary software platforms: Palantir Gotham, Palantir Foundry, Palantir Apollo—and its newest, Palantir Artificial Intelligence Platform (AIP).

Palantir launched AIP in April 2023. The AI platform aims to let clients combine public datasets—such as the totality of text and information available on the Internet, with privately held repositories of information maintained by those clients, including across the commercial and government sectors. At the same time, it keeps client data private and secure.

Meanwhile, Palantir is also using AI in the products it offers to clients. For instance, last year Palantir closed a $178.4 million contract with the U.S. Army for a ground station project, referred to as Titan, powered by AI. In the two-year agreement, the company will deliver ten prototypes which use machine learning to enhance the speed and accuracy of combatant targeting from various sensors. Military contracts are likely to expand in the future given Palantir’s superior technology in this area.

In the quarter ended December 31, 2024, Palantir’s revenue rose 36.0%, to $827.5 million from $608.4 million a year earlier. Customer count grew 43% year-over-year. The company closed 129 deals of at least $1 million, 58 deals of at least $5 million, and 32 deals of at least $10 million.

Excluding one-time items, the company earned $341.9 million, or $0.14 a share, in the latest quarter. That was up 80.3% from $189.6 million, or $0.08.

Under its listing on New York, Palantir established four different tiers of stock: Class A shares with 1 vote each; class B shares with 10 votes; preferred shares for certain investors that are not typically counted in the same way as other classes of stock; and class F shares for founders. The voting rights of those F shares can vary, meaning the company’s founders can have control whenever they want.

Palantir operates in a highly competitive market. In addition, we are wary of new stock issues. That’s all the more so when insiders control the firm through multiple voting or special shares.

Still, despite those risks, the company’s outlook is positive:

Palantir’s balance sheet is strong. It holds cash of $5.2 billion, and it has no debt.

The company spends a very high 18% of revenue on research and development, and that makes it appear a lot less profitable than it really is. Although, as mentioned, it operates in competitive markets—markets very dependent on U.S. defence spending levels—high spending helps it stay on top of industry changes.

Going forward, AIP will continue to be a major growth area for Palantir—building on its already rapid success. In fact, in the latest quarter, the company’s U.S. commercial revenue increased by 64%—spurred by its AIP platform.

Palantir is a buy for AI investing, but only for aggressive investors.

With over four decades of experience as an advisor, commentator, editor and publisher, Pat McKeough has a long record of determining which stocks are bound to reward investors most.

Over the past two decades he has been the editor and publisher of a growing series of investment newsletters through TSI Network. Pat also offers two investment advice services, Inner Circle and the advanced Inner Circle Pro. Since 1999, he and his team have put his investment approach to work for private clients in his Successful Investor Wealth Management business.

His philosophy is anchored in safety and a balanced portfolio to generate accelerating gains for subscribers and clients. TSI Network now publishes seven newsletters for every kind of investor:

1. The Successful Investor—Pat’s flagship advisory continues to be a beacon for Canadian investors seeking growing gains and reduced risk with the best Canadian stocks.
2. Power Growth Investor—If you like the idea of “a conservative approach to aggressive investing”, this advisory has Canadian and U.S. stocks with escalating growth potential.
3. Wall Street Stock Forecaster—Your portfolio is much stronger with at least 20% in U.S. stocks—and this special advisory covers the 70 best U.S. stocks for Canadians.
4. Canadian Wealth Advisor—A ‘safety-first’ advisory offering you the best conservative strategies based on well-established Canadian dividend stocks, ETFs and REITs.
5. Dividend Advisor—In this advisory, our exclusive Dividend Sustainability Ratings® will change the way you look at dividend stocks—and the way you invest in them.
6. Spinoffs & Takeovers—If you’d like “the closest thing to a sure thing in investing,” this advisory on spinoffs and other special opportunities is utterly unique.
7. Best ETFs for Canadian Investors—This ground-breaking publication shows you how to get the best results with ETFs as these investments explode in popularity.

In 2002, Pat founded his Inner Circle, offering investors more personal attention, plus access to his four original publications. Members can ask Pat personal investment questions. They also get his commentaries and answers to questions posed by other Inner Circle Members. In 2017 he launched Inner Circle Pro, an advanced group that receives all seven of his newsletters.

Through Successful Investor Wealth Management, Pat and his team manage over $900 million for individual Canadian investors. Free of comprising ties to brokerages, with no hidden costs or commissions, the team charts an independent course for clients. For the past 18 years the portfolios they manage for clients have enjoyed an uncommonly high annual average return.