The artificial intelligence trade is not falling apart.
Demand is growing.
After a spectacular first-half advance in chip and memory stocks, investors are starting to distinguish companies that gain from AI investing from those that can develop without bearing the brunt of the spending.
That’s why it’s worth hearing from Victoria Greene, chief investment officer of G Squared Private Wealth.
She likes “capex light” companies like Nvidia (NVDA) and Apple (AAPL), Greene told CNBC. That’s the buried investor implication underlying the recent AI-stock whiplash.
The market is no longer treating all companies in the AI buildout the same.
It’s starting to wonder which companies can turn AI demand into revenue, margins, and cash flow without having to fund the full infrastructure race themselves.
That makes Nvidia and Apple both significant, but for different reasons.
NVIDIA is at the heart of the AI infrastructure demand. Apple is closer to the consumer and services layer. But both may give investors something more significant in the next phase of the AI trade: access to the theme without the same capital intensity facing memory makers, cloud builders, and data-center operators.
Greene is now favoring “capex light” companies such as Nvidia and Apple, or what she calls “old school” tech, CNBC reported.
Nvidia and Apple give investors a different AI setup
In the initial phase of the AI rally, exposure was rewarded. Many investors considered companies selling chips, memory, storage, servers, networking equipment, or anything related to data centers to be in the same winning trade.
That worked the first part of this year.
The PHLX Semiconductor Index (SOX) nearly quadrupled in the first half of the year before falling in July. It referenced data from Bespoke Investment Group showing that 22 S&P 500 equities, which had more than doubled in the first six months of 2026, were recently down an average of 16% in July. Many of those stocks were linked to technology and memory chips.
That sort of reversal doesn’t indicate the demand for AI is over. It shows investors are shifting from a broad “buy anything AI” phase to a more focused phase.
This is where Greene’s capex-light argument comes in.
Related: Bank of America’s unmistakable signal to Nvidia stock investors
Capital investments are important, since AI is not a cheap growth story.
Training and operating complex models involve huge spending on chips, servers, networking gear, energy, and data center capacity. The corporations developing that infrastructure might reap enormous revenue, but they also have significant capital demands and severe investor scrutiny on return on investment.
LPL Financial echoed that sentiment in its view on AI, noting that the market is shifting “from pricing in promise to pricing in execution.” AI is a key driver for equities markets, but investors are more and more concerned about whether the spending is actually resulting in monetization, the business said.
That helps explain why Apple (AAPL) and Nvidia (NVDA) might be in the middle of a more selective AI trade.
The argument is not that either company avoids investment. Nvidia spends heavily on research and product development. Apple spends heavily across its supply chain, services ecosystem, and hardware roadmap.
The distinction is that neither business is aiming to carry the full weight of building out global AI infrastructure on its balance sheet like hyperscalers, data center operators, or memory manufacturers.
Nvidia is selling the tools and shovels, and Apple has control over the consumer ecosystem. That may be the edge investors are beginning to realize.
Nvidia proves AI demand without owning the heaviest buildout
Nvidia continues to be the best example of a firm that is already converting AI demand into financial outcomes.
For the first quarter of fiscal 2027, the corporation recorded a record $81.6 billion in revenue, an increase of 85% from a year ago. Nvidia’s official statistics show data center revenue was $75.2 billion, up 92% from a year ago.
Those stats demonstrate why Nvidia isn’t just another AI play. It’s monetizing the infrastructure boom on a scale few companies can match.
That’s important for Greene’s cap-ex-light argument because Nvidia gets to enjoy the benefit of AI capital spending without having to build every data center, finance every power deal, or carry every cloud-computing facility on its own books.
It sells the processors, systems, and networking products that make those initiatives happen.
That’s still a tough position. Investors are betting on Nvidia to keep producing tremendous growth, great profits, and strong product cycles. Any slowdown in orders from hyperscalers, export limitations, or competition from in-house chips might put pressure on the stock.
But Nvidia is a significant player in the AI chain with its commercial model, and it can join the spending race without being the one that actually pays for it all.
More Nvidia:
- Nvidia’s workplace culture sends Big Tech a warning
- Nvidia’s $25B bond deal sends investors a clear signal
- Bank of America resets Nvidia stock forecast after CFO meeting
That difference will become more relevant as AI commerce grows up.
A business selling scarce, high-margin components into an infrastructure boom can seem quite different than a business that has to spend much up front and wait for years to justify the return.
That’s why Nvidia might still be appealing after a monster run.
Greene’s thesis isn’t only that Nvidia is an AI leader, according to Investopedia. That is, it may currently appear more tempting than the more capital-intensive parts of the chip cycle.
The value point is also important. Nvidia’s forward price-to-earnings ratio for 12 months was roughly 19, compared to about 20 for the S&P 500. Nvidia’s ratio was 25 at the end of December versus 22 for the broader index.
That doesn’t mean Nvidia is inexpensive in every investor’s mind. But it does indicate why some pros would evaluate the stock differently after the recent AI reset. If the earnings have accelerated enough to pull the multiple down, the stock can look less extended, even after a big gain.
The danger is that investors will now want to see proof every quarter, while the opportunity is that Nvidia remains in the sector of the AI economy where demand is easiest to see.
Apple may offer a quieter AI advantage
Apple is another kind of AI narrative. It doesn’t have the data-center income engine of Nvidia, nor does it have Micron’s direct exposure to high-bandwidth memory demand. It doesn’t feel like the purest method to play AI infrastructure.
That may be the idea, however.
The encouraging news for Apple is that standard artificial intelligence may be used to enhance an existing environment, rather than drag the business into an all-or-nothing infrastructure competition.
Apple reported revenue of $111.2 billion for the fiscal second quarter, up 17% year over year, and diluted profits per share of $2.01, up 22%. The corporation also claimed services revenue reached a record high and approved another $100 billion share repurchase program.
Those metrics indicate why Apple is a different kind of tech investment.
Its AI payout might come from iPhone upgrades, services engagement, developer tools, productivity features, and tighter consumer lock-in. That’s less dramatic than a chip scarcity, but it may be strong if it makes the company’s installed base more valuable.
That’s the beauty of the capex-light model. Apple does not have to win the AI infrastructure war, unlike Nvidia. It needs AI to make it difficult to leave its devices and services behind.
That might give Apple a more muted ride through the next leg of the AI cycle. The tech giant may use AI to enhance Siri, photo editing, device search, messaging, productivity, and app experiences.
If those features help secure the iPhone ecosystem or increase service consumption, the financial impact can occur without Apple having to build the same capital-heavy company as cloud operators.
But investors still have reasons to worry. Apple needs to show that AI features are significant enough to change user behavior. The corporation also confronts regulatory scrutiny, competition in China, and worries about whether iPhone upgrade cycles will gain traction.
Still, Apple has an edge that many other AI infrastructure names lack. It’s already secured the consumer relationship, giving it a different avenue to monetize AI. That’s why Greene’s “old school” tech point matters to investors.
Key takeaways from the capex-light AI trade
- The AI trade is not ending, but it is becoming more selective.
- G Squared’s Victoria Greene is favoring capex-light names such as Nvidia and Apple.
- Nvidia benefits from AI infrastructure spending without carrying the full data center buildout burden.
- Apple can use AI to deepen its device and services ecosystem rather than compete as a pure infrastructure company.
- Memory and semiconductor stocks can still benefit from AI demand, but capital intensity and cyclicality matter more after a sharp rally.
- Investors should watch whether AI leaders can turn demand into durable revenue, margins, and free cash flow.
These findings suggest a more mature market. The first AI-stock surge rewarded investors who picked the correct subject. Those who can identify the right company model may now benefit from the next stage.
That’s a tougher test, since memory stock can have tremendous revenue growth and still trade like a cycle. AI spend can be good for a semiconductor supplier, while doubts remain on margins, supply, and customer concentration. A cloud business can spend hundreds of billions on AI and yet be judged on whether consumers pay enough to support the buildout.
In this framework, the semiconductor business is where Nvidia and Apple shine. Nvidia captures the infrastructure spend, and Apple might dictate how users are treated.
Both can engage in AI without being judged simply on how much capital they put to work.
Investors should watch who pays for the AI boom
The future of the AI trade may hinge on one question: Who pays?
That’s where the capex-light argument really comes in.
For example, the AI memory trade can be compelling yet capital-intensive. For example, consider Micron Technology (MU). Capital expenditures, net, were $7.1 billion in the business’s fiscal third quarter, the company said. Adjusted free cash flow was $18.3 billion. Micron said its record performance and improved outlook also underlined the strategic relevance of memory in the AI era.
That’s a good business tale. It’s not the same tale as Nvidia or Apple, however. To meet the AI-driven demand, Micron needs to pour a lot of money into technology, products, and supply. That can create considerable upside when memory pricing is favorable, but it also reminds investors that some AI winners are still beholden to capital cycles.
Nvidia’s danger is another. Its clients are the ones putting most of the physical AI infrastructure in place. Nvidia still has to innovate and fulfill demand, but its economics are tied to selling into that buildout.
Apple faces another danger. It must show that AI makes its products more useful and its ecosystem even more lucrative.
That turns the capex-light concept from a catchphrase into an investing filter. It also asks investors if a company is funding the AI boom, supplying the AI boom, or profiting from the AI boom via an established platform. They are separate positions and deserve different appreciation.
The real question for retail investors isn’t whether AI remains a powerful subject. The better question is, “Is the stock they own playing the correct role in the theme?” That’s why Greene’s Nvidia-and-Apple call should be the story’s centerpiece.
It indicates that the AI trade is not decoupling from technology. It’s heading towards companies that can demonstrate their economics without investors footing the bill for infinite capital expenditure.
The AI rally is moving toward business-model discipline
The AI rally hasn’t run out of fuel; it has run into a taller barrier.
Investors still want to pay for artificial intelligence growth, but they seem less prepared to pay for any firm approaching the theme. That’s the difference between an early-stage rally and a more mature market.
Nvidia and Apple matter because they give investors two alternative ways to think about what’s next. Nvidia is the cash cow in the AI infrastructure revolution. Apple is the ecosystem firm that can turn AI into increased engagement, greater services demand, and more sustainable customer loyalty.
Nvidia has to keep proving data center growth can stay spectacular. Apple will need to demonstrate that standard artificial intelligence can deliver more than just the capabilities investors were previously expecting.
But both are a better fit for Greene’s capex-light structure than many sections of the AI trade that have already soared.
That’s the key takeaway for investors from the latest chip whiplash: The market isn’t saying AI isn’t relevant anymore. It says the business model is more important.
Investors can be disappointed even when a firm is exposed to AI if the benefits are not evident, the burden of spending is too high, or the stock already reflects perfection.
This scenario means the next leg of the AI trade is less about owning the broad theme and more about owning the companies with the clearest path from AI demand to earnings.
For Greene, that points toward Nvidia and Apple. For investors, it points toward a more disciplined question.
Not “Which companies are AI stocks?” but “Which companies can profit from AI without having to pay the highest price to chase it?”
