Nvidia’s Valuations Show AI Rally Isn’t a Bubble, says DBS Group

Nvidia’s valuation and earnings growth suggest the artificial intelligence-driven stock rally is not yet a bubble, according to DBS Group Chief Investment Officer Hou Wey Fook. His argument rests on a relatively simple comparison: Nvidia is trading at about 17 times expected earnings over the next 12 months while analysts expect its earnings to grow by roughly 70% next year.

The comparison comes as Nvidia shares reach fresh highs and the company’s market value moves further above $5 trillion. That combination has revived debate over whether the AI investment boom has entered bubble territory. DBS argues that Nvidia’s valuation is difficult to classify as a classic speculative extreme because the company’s profits are growing at an unusually rapid pace and its valuation multiple remains far below the levels reached by technology leaders during the dot-com era.

Key takeaways

• DBS CIO Hou Wey Fook says Nvidia’s valuation does not currently resemble a classic AI bubble.
• Nvidia is trading at roughly 17 times its next-12-month expected earnings, according to Bloomberg data cited by DBS.
• DBS points to expected earnings growth of about 70% next year.
• Nvidia shares closed at a record $237.58 on October 5, giving the company a market value of roughly $5.76 trillion.
• Nvidia generated $96.2 billion of revenue in fiscal Q2 2027, up 106% year over year.
• Data Center revenue reached $89 billion, up 117%.
• Nvidia expects $108 billion of revenue in fiscal Q3 2027, plus or minus 2%.
• The argument does not eliminate risks around AI spending, customer concentration, competition, regulation or future returns on AI infrastructure investment.
• DBS recommends a “barbell” investment approach, combining AI-related growth assets with investment-grade fixed income and diversifiers.

Why DBS says Nvidia is not in a bubble

The central argument from DBS is not that Nvidia’s shares cannot fall.

Instead, Hou Wey Fook argues that the company’s valuation should be considered alongside the speed at which its earnings are increasing.

Nvidia is currently trading at about 17 times expected earnings over the next 12 months, according to Bloomberg data cited in the report. DBS compares that multiple with Cisco’s valuation during the dot-com period, when Cisco traded at roughly 100 times earnings before the technology bubble burst.

That comparison does not mean Nvidia and Cisco are directly comparable businesses.

The more important point is the relationship between valuation and underlying profits.

A company trading at a very high multiple requires enormous future growth to justify that valuation. Nvidia is certainly priced at a premium to the broader market, but its earnings have also been expanding at a rate that is far above normal large-cap corporate growth.

That is why DBS sees a fundamental difference between today’s AI rally and the late-1990s technology bubble.

Nvidia’s earnings provide the foundation for the argument

Nvidia’s recent financial results give investors a strong reason to take that argument seriously.

For the second quarter of fiscal 2027, Nvidia reported revenue of $96.2 billion, up 106% from a year earlier and 18% sequentially. Net income reached $59.7 billion, while GAAP gross margin was 75%.

The company’s Data Center business was responsible for most of the growth.

Data Center revenue reached $89 billion, up 117% from the same quarter a year earlier and 18% sequentially.

That means Nvidia is not simply benefiting from investors assigning a higher valuation to a company with uncertain future earnings. Its existing business is already producing enormous amounts of revenue and profit.

The distinction is important when assessing whether a market rally is fundamentally driven or primarily speculative.

The valuation argument is about growth-adjusted expectations

A price-to-earnings ratio cannot be evaluated in isolation.

A mature company growing earnings by 5% or 10% annually may look expensive at a high multiple. A company growing earnings at 70% can potentially justify a much higher multiple.

This is essentially the argument DBS is making.

Nvidia’s expected earnings growth means investors are paying a relatively modest multiple against a rapidly expanding earnings base.

If earnings continue to grow at the expected rate, today’s valuation could look considerably less demanding in hindsight.

The risk, of course, is that the growth assumption proves too optimistic.

That is where the debate over the AI bubble becomes more complicated.

The question is not simply whether Nvidia is profitable today. It is whether the enormous amount of capital being invested across the AI ecosystem will continue producing sufficient revenue and earnings growth to justify current valuations.

Nvidia’s $5.7 trillion valuation makes the debate unavoidable

Nvidia’s shares closed at $237.58 on October 5, giving the company a market capitalization of approximately $5.76 trillion.

That makes Nvidia one of the most valuable companies ever created.

At that scale, investors cannot expect the company to grow simply by selling a few more chips.

Nvidia now needs to add tens of billions of dollars of incremental revenue to sustain its growth trajectory.

So far, it has been doing exactly that.

Revenue increased from $46.7 billion in the year-earlier quarter to $96.2 billion in fiscal Q2 2027.

The challenge for investors is that the comparison will become progressively harder.

Once a company reaches a revenue base approaching $100 billion per quarter, maintaining 50%, 70% or 100% growth becomes mathematically more difficult.

That does not make the growth impossible, particularly during a structural technology transition, but it raises the level of execution required.

Nvidia expects another major jump in revenue

Nvidia’s own guidance reinforces the bullish argument.

The company expects fiscal Q3 2027 revenue of approximately $108 billion, plus or minus 2%.

That would represent another major increase from the $96.2 billion reported in the previous quarter.

Importantly, Nvidia’s guidance does not assume any Data Center computing revenue from China.

That means the current outlook is being built largely around demand outside China’s Data Center market.

The company’s results show that hyperscalers, AI-native companies, enterprises and sovereign customers are continuing to invest heavily in accelerated computing infrastructure.

AI infrastructure spending is the real engine

The Nvidia story is ultimately a story about capital expenditure by other companies.

Cloud providers and technology companies are spending enormous amounts of money on data centers, networking, power infrastructure and AI accelerators.

Nvidia captures part of that spending through GPUs, CPUs, networking equipment and increasingly complete AI infrastructure systems.

Its latest financial filings show the scale of this transition.

Nvidia said Data Center revenue increased 117% year over year in fiscal Q2, with growth driven by the ramp of its Blackwell Ultra infrastructure.

The company also said hyperscale revenue more than doubled from a year earlier.

That provides a mechanism behind the earnings growth.

Nvidia is not simply raising prices on the same products. The AI infrastructure market itself is expanding rapidly.

Why the dot-com comparison matters

The dot-com bubble is the benchmark frequently used when investors discuss whether AI has gone too far.

During the late 1990s, technology companies attracted enormous valuations based largely on expectations of future internet adoption.

Many businesses had little or no sustainable profit.

When expectations reversed, valuations collapsed.

Nvidia is different in one crucial respect: it already generates enormous profits.

Its fiscal Q2 2027 net income was almost $60 billion.

That does not make the company immune to a valuation correction. But it makes the comparison with unprofitable dot-com companies less straightforward.

DBS therefore sees the current AI market as being supported by actual corporate earnings rather than only by expectations.

The distinction matters for investors trying to determine whether today’s AI enthusiasm is a speculative mania or the early stages of a long-lasting technological investment cycle.

But a profitable bubble is still possible

The strongest counterargument to DBS is that profitability alone does not prove a market cannot be in a bubble.

A company can have excellent earnings and still be overvalued if investors assume that unusually high growth will continue for too long.

Nvidia’s current valuation incorporates significant expectations about future AI infrastructure spending.

If hyperscalers eventually reduce capital expenditure, if AI applications fail to generate sufficient returns or if custom chips become more competitive, Nvidia’s earnings growth could slow.

The stock’s valuation could then come under pressure even if Nvidia remains highly profitable.

This is why the AI bubble debate should not be reduced to a simple comparison between Nvidia’s P/E ratio and historical technology valuations.

The durability of AI demand matters just as much as the current earnings multiple.

Nvidia’s gross margins are another important signal

Nvidia’s profitability is also unusually strong.

The company reported a 75% GAAP gross margin for fiscal Q2 2027.

Such margins provide Nvidia with substantial financial flexibility.

High profitability allows the company to invest heavily in research and development while still generating significant cash.

It also provides room for shareholder returns.

Nvidia recently increased its share-repurchase authorization by $150 billion, bringing the remaining authorization to approximately $235 billion through fiscal 2028.

That is another factor investors can consider when assessing the company’s valuation.

The market is not valuing a company that merely expects to become profitable. It is valuing a company already generating enormous cash flows and using part of those resources to return capital.

The AI ecosystem still carries significant risks

DBS’s view does not eliminate the risks surrounding the AI investment cycle.

One major concern is whether the world’s largest technology companies can earn attractive returns on their AI spending.

Amazon, Microsoft, Alphabet, Meta and other companies are committing tens of billions of dollars to data centers and AI infrastructure.

Those investments make Nvidia’s current revenue growth possible.

But eventually, those customers need their AI businesses to generate enough revenue or productivity gains to justify the capital expenditure.

If the economics disappoint, spending could slow.

That would eventually affect Nvidia.

Another risk is competition.

Nvidia has a major advantage through its CUDA software ecosystem, hardware roadmap and networking portfolio, but AMD, custom cloud-provider accelerators and other specialized chips continue to develop.

Over time, customers could attempt to diversify their computing infrastructure.

Nvidia’s software advantage matters

One reason Nvidia has maintained such a strong position is that its competitive advantage extends beyond the GPU itself.

CUDA provides developers with a mature software ecosystem for building and optimizing AI workloads.

Nvidia has also expanded its software stack around AI agents, inference, networking and enterprise applications.

This creates switching costs.

A customer that has built large AI workloads around Nvidia hardware and software cannot necessarily replace that infrastructure simply by purchasing a different chip.

That ecosystem effect is one reason investors continue to assign Nvidia a premium valuation.

The company is increasingly positioning itself as an infrastructure platform rather than simply a semiconductor manufacturer.

The next phase is AI inference

Another potential source of growth is AI inference.

Training involves using massive computing resources to develop an AI model. Inference occurs when that model is actually used to generate responses, recommendations, predictions or other outputs.

As AI becomes embedded in consumer applications and enterprise software, inference workloads could grow dramatically.

Nvidia is developing products specifically for this stage of the market, including its Vera Rubin platform and new inference-focused accelerators.

If AI agents become widely deployed, the amount of computing required to serve those systems could increase substantially.

That could extend the demand cycle beyond the current training-focused infrastructure boom.

Nvidia is building an entire AI factory ecosystem

Nvidia’s strategy also increasingly extends beyond individual processors.

The company has introduced rack-scale systems, networking platforms, CPUs and infrastructure software designed to function together as complete AI factories.

Its fiscal Q2 disclosures said Vera Rubin was entering full production, with systems running at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius.

Nvidia also announced partnerships intended to mobilize more than $500 billion of third-party capital for AI infrastructure over time, subject to definitive agreements.

This is important because it potentially expands Nvidia’s economic exposure to AI infrastructure beyond the sale of individual GPUs.

The more components and services Nvidia supplies, the larger the potential revenue opportunity per AI deployment.

DBS recommends a more balanced portfolio

Despite his bullish assessment of AI stocks, Hou is not advocating an all-in technology strategy.

He recommends what is known as a “barbell” approach.

Under this strategy, investors combine higher-growth assets such as technology stocks with more defensive assets such as investment-grade fixed income.

Gold and hedge funds can act as additional diversifiers.

The underlying logic is straightforward.

An investor can participate in the AI growth story without making the entire portfolio dependent on the continued rise of technology valuations.

That recommendation also acknowledges that even a fundamentally justified AI rally can experience significant volatility.

What could prove DBS wrong?

Several developments could challenge the argument that the AI rally is not a bubble.

The first would be a sustained slowdown in Nvidia’s earnings growth.

If earnings growth fell sharply while the stock remained on a high multiple, the valuation would become considerably more demanding.

The second would be a reduction in hyperscaler capital spending.

If Microsoft, Alphabet, Amazon or Meta began cutting AI infrastructure budgets, Nvidia’s revenue outlook would be affected.

The third would be rapid technological disruption.

If customers increasingly shifted from Nvidia GPUs toward custom accelerators or competing products, Nvidia’s market share could decline.

The fourth would be disappointing AI monetization.

If businesses spend heavily on AI but fail to generate sufficient economic returns, future infrastructure investment could slow.

Those are the variables that matter more than whether Nvidia’s P/E ratio looks high or low in isolation.

What Nvidia’s valuation says about the wider AI market

Nvidia’s valuation provides a useful test for the broader AI rally.

If the market leader is generating rapidly rising revenue, extremely high margins and substantial profits while trading at a relatively moderate forward earnings multiple, then at least part of the AI rally is being supported by fundamentals.

But that does not mean every AI stock is similarly justified.

Smaller companies with limited revenue, high valuations and weak profitability may still exhibit bubble-like characteristics.

Investors therefore need to distinguish between the economics of AI infrastructure leaders and speculative companies benefiting simply from the popularity of the AI theme.

Nvidia’s financial results offer evidence for the former.

The bigger question is what happens after the buildout

The most important unanswered question is what happens once the current AI infrastructure buildout matures.

Companies are spending aggressively today because AI computing capacity is scarce and demand is expanding.

At some point, supply will catch up with demand.

When that happens, growth rates could normalize.

Nvidia’s valuation will then depend more heavily on recurring AI workloads, inference demand, software revenue and replacement cycles.

The company is already positioning itself for that transition.

Its move into CPUs, networking, software, AI factories and inference infrastructure suggests management expects AI computing to become an ongoing technology platform rather than a one-time hardware cycle.

The Bigger Picture

DBS’s argument does not prove that Nvidia cannot be overvalued. It instead challenges the idea that the entire AI rally is inherently speculative. Nvidia is producing extraordinary profits, revenue is more than doubling, Data Center sales are growing even faster, and the company is generating cash at a scale that was difficult to imagine only a few years ago.

The real test will be whether that earnings growth can continue as Nvidia’s revenue base becomes enormous. A 70% growth rate on a company already generating nearly $100 billion in quarterly revenue requires an extraordinary expansion in global AI infrastructure spending. That is the assumption investors ultimately need to evaluate.

FAQs

Is Nvidia in an AI bubble?

DBS CIO Hou Wey Fook argues that Nvidia’s valuation does not currently resemble a classic bubble because the company is generating enormous profits and is expected to deliver very strong earnings growth. That is an investment view, not proof that the stock cannot be overvalued.

What is Nvidia’s forward P/E ratio?

Nvidia was trading at about 17 times expected earnings over the next 12 months, according to Bloomberg data cited by DBS.

Why is Nvidia’s valuation considered relatively reasonable?

The argument is based on the relationship between its valuation and earnings growth. Nvidia’s forward P/E is relatively modest compared with its expected earnings growth of about 70%, according to the DBS analysis.

How fast is Nvidia’s Data Center business growing?

Nvidia’s fiscal Q2 2027 Data Center revenue reached $89 billion, up 117% year over year.

What could cause Nvidia shares to fall?

A slowdown in AI infrastructure spending, weaker-than-expected earnings growth, stronger competition, lower customer returns on AI investments or a compression in Nvidia’s valuation multiple could pressure the stock.

Looking Ahead

Nvidia enters the next phase of the AI cycle with an unusually strong combination of growth and profitability. The company’s immediate financial outlook remains supported by enormous demand for Blackwell and newer AI infrastructure, while its expansion into networking, CPUs, inference and complete AI systems gives it additional avenues for growth.

The harder question for investors is whether those conditions can persist as Nvidia becomes a $5 trillion-plus company. DBS believes the current valuation remains compatible with the company’s earnings trajectory, but the AI rally will ultimately be judged by whether real-world AI revenue and productivity gains continue to justify the extraordinary amount of capital being committed to the technology.