Listen to the audio version of this article (generated by AI).
Tom Yeung here with your Sunday Digest.
At the start of 2021, Nvidia Corp. (NVDA) was a $330 billion company. It had a significant data-center business, but still mostly built chips for gaming computers and consoles.
As you well know, Nvidia would transform into an AI behemoth. ChatGPT debuted in late 2022, a massive AI data-center buildout began the following year, and Nvidia is now the most valuable company on Earth. Investors during that stretch enjoyed a roughly 940% return.
One by one, other AI-related companies have also seen their “Nvidia moment” arrive:
- First, it was semiconductor companies like Broadcom Inc. (AVGO) and Advanced Micro Devices Inc. (AMD).
- Next came the data-center infrastructure firms, such as Super Micro Computer Inc. (SMCI) and Arista Networks Inc. (ANET).
- Then memory-chip makers went vertical. Micron Technology Inc. (MU) and SanDisk Corp. (SNDK) became household names.
And every time, many investors watched from the sidelines as a new group of stocks zoomed ahead. Even those lucky enough to get in would often wonder, “Why didn’t I buy more?”
Fortunately, there are still companies that are only at the beginning of the AI Revolution.
In a new special presentation, our three senior analysts – Louis Navellier, Eric Fry, and Luke Lango – join forces to talk about their AI Revolution Portfolio: a carefully selected group of 19 stocks that still have room to grow thanks to this new technology. They believe we’re reaching a point where AI is going super exponential and turning 10X returns from a slow-motion event into something that can happen in months… or even days.
.
To give a sense of the opportunity, I’d like to highlight several other picks that are still being overlooked by Wall Street. And to get the full details on their AI Revolution Portfolio – and a little bit more on their 19 official picks – .
The Eyes of the AI Revolution
Earlier last week, Chinese firm Unitree Robotics surged 629% after listing in on the Shanghai Stock Exchange.
You probably recognize the company from viral videos of its backflipping, dancing robots – moves I can only dream of doing.
Unitree’s $13,500 G1 humanoid robot costs a fraction of what competitors charge, and the company has already set up an “app store” for owners to download more skills. It’s easy to see a future where Unitree robots gain new abilities like “nanny mode” or “master electrician” at the click of a button.
While I do believe the firm will succeed in the long run, I do not recommend trying to buy this one stock. Shares were at least 5,500 times oversubscribed, and its abnormal first-day pop reflects China’s peculiar trading rules. In fact, Unitree soon fell around 20% after its IPO, leaving new investors with immediate losses.
Unitree investors are buying hype right now, and I never like buying hype by itself. That’s why I would recommend a “picks-and-shovels” play instead.
It’s a company that supplies vision systems for Unitree’s robots…
It ships 60% of the “eyes” of all robotaxis…
And best of all, shares are barely higher than they were at the start of 2025.
This company is Hesai Group (HSAI)
Hesai has spent the past several years consolidating the market for lidar sensors – the devices that use lasers to map the world around vehicles and robots.. It has driven at least a half-dozen competitors out of the business, and now supplies 8 out of the top 10 robotaxi companies globally. Hesai also counts major robotics manufacturers as customers, including Unitree.
The company did this with a cutthroat pricing strategy, which has kept share prices low. Operating margins were just 5.6% in 2025, and investors rightly stayed away. HSAI trades a hair below its 2023 listing price, even as AI-driven demand from robotaxis and humanoid robots have pushed ahead.
But Hesai is finally winning the lidar war. Analysts expect operating margins to hit 13% this year, 15% in 2027, and then 18% in 2028. That means Hesai is already on its way to transforming from a low-returning company like General Motors Co. (GM) to a high-returning one like Sensata Technologies Holding PLC (ST) – a specialist auto-parts manufacturer.
Now, this flip to higher profits is happening right as robotaxi growth is reaching an inflection point. Self-driving car companies are rapidly expanding in cities like Los Angeles, Beijing, and Dubai, and many are projecting 100%-plus annual fleet growth thanks to sudden improvements in AI-powered self-driving technologies.
Meanwhile, more and more people – including myself – are becoming OK with getting chauffeured around by these driverless cars.
That matters because robotaxis require far more lidar units than ordinary cars. Some carry as many as eight Hesai sensors to eliminate blind spots. Many higher-end passenger cars are also beginning to use multiple lidar rigs to enable “Level 3” self-driving – the type that allows drivers to take their eyes off the road. Humanoid robots will only add to this eventual demand.
Of course, Hesai is not a risk-free investment. The company exports products to America, which has put up numerous trade barriers in recent quarters. The firm also has a customer concentration issue due to the sheer size of some robotaxi customers; the top 5 customers made up more than half of 2025 revenues. And, of course, a new price war could break out if a well-funded firm decides to start selling lidar components as well.
Nevertheless, the risk-reward looks considerably better than chasing a robotics IPO that opened 629% higher on its first trading day. Hesai does not need Unitree to become the world’s dominant robot maker. Nor does it need Baidu Inc. (BIDU), PonyAI Inc. (PONY), or WeRide Inc. (WRD) to win the robotaxi race.
It simply needs autonomous machines to keep proliferating. And that is something the AI Revolution almost guarantees will happen.
And the Brains
The bull story for this next stock is simple:
China is becoming a powerhouse in AI, and Beijing badly wants its homegrown companies to win.
To capitalize on this trend, I would recommend shares of the country’s most established AI firm: Alibaba Group Holding Ltd. (BABA).
The e-commerce giant has pursued Amazon.com Inc. (AMZN)-style vertical integration. It designs its own chips, runs its own data centers… and even builds its own AI models
Chips. The country lacks the specialized machinery used to make high-performance GPUs, and so Alibaba’s designers have developed clever workarounds. For instance, their “Panjiu” servers use 128 AI accelerators rather than the 72 that Nvidia uses. This brute force method helps offset weaker chips.
Models. Its Qwen family of AI models captured 30% of self-hosted AI usage at the end of 2025 (that’s when you download an AI model to run for yourself), and it has since created some of the best AI models in the world. Its latest model, Qwen3.8 Max, is almost as good as Claude Fable 5, Anthropic’s most advanced model. It even beats Fable in certain tests like operating normal desktop computers.
Datacenters. Superior chips and models have allowed Alibaba to provide cloud computing that is both effective and cheap. Its flagship Qwen3.8 Max model costs barely a third of Claude’s Fable 5 to run, and these leading models have helped Alibaba gain a 40% market share of China’s AI cloud computing market. That puts it ahead of the three next companies combined – ByteDance, Huawei, and Tencent.
Meanwhile, Alibaba’s e-commerce side is a cash cow. Earlier last week, the company reported that its global e-commerce sales increased 4%, and that adjusted earnings before interest, taxes, and amortization (EBITA) from the segment hit $5.9 billion. That figure easily offsets the $2 billion of losses its Cloud Intelligence segment currently generates from its heavy investments.
Best of all, Alibaba has not yet hit its “Nvidia moment.” The stock has fallen 19% since 2021 when Nvidia was beginning its epic run and trades at a discount to its Chinese peers.
I must note that Alibaba faces regulatory issues. The Chinese government “disappeared” the company’s co-founder in 2020 after he gave a speech that criticized the country’s financial regulators; it could do something similar again if the company steps out of line. China is also planning its own state-led data center buildout, which could cap profits in that sector.
Nevertheless, these risks are already priced into Alibaba’s shares. The company trades at just 19X forward earnings and is worth less than Nvidia was in 2021. In my view, this sets Alibaba up for its own “Nvidia moment.”
Riding The AI Revolution
One of the most common questions I get asked is, “Are we in another dot-com bubble?”
That’s smart to ask. Things turned out badly the last time investors went head over heels for a new technology.
But here’s what’s different from the 1990s.
You see, during the dot-com boom, everyone knew telecom companies were building too much capacity. By 1999, roughly half of America’s installed fiber network sat “dark,” and more than 90% of some newly constructed long-haul networks were unused.
We saw the same pattern during the housing boom of the 2000s. Buildings went up in the middle of deserts and swamps… and they had no residents.
The AI Revolution different. Today, companies are struggling to build enough cloud computing capacity because everything is bottlenecked. Roughly 40% of AI data-center construction projects are now delayed until next year, and certain components like flash memory and hard drives are sold out until 2028.
That’s why Louis, Eric, and Luke believe there’s still plenty of room for the rally to run. So, be sure to before that next company you’ve had your eye on reaches its “Nvidia moment” before you’ve had a chance to get in.
I will be out of town the next two weeks for travel, and so I’ll see you back here after Labor Day.
Thomas Yeung, CFA
蜜桃传媒 Analyst, InvestorPlace
