DBS rests its case against an AI bubble on Nvidia's 17x forward earnings multiple
DBS CIO Hou Wey Fook says AI stocks are not in a bubble, citing Nvidia at about 17x forward earnings and a 70% revenue rise forecast for fiscal 2028. The multiple stays cheap only while the profit forecasts beneath it hold.
The Investor · Invest desk
What happened
- Nvidia reported $96.2 billion in revenue for its fiscal second quarter ended July 2026, up 106% from a year earlier.
- Data-center sales of the chips and systems used to train and run AI models made up nearly $89 billion of that quarter's revenue.
- Nvidia's management has described supply as constrained, with demand running ahead of what the company can produce.
- Hou recommends a barbell portfolio that pairs tech and AI growth stocks with defensive assets meant to hold steady in choppy markets.
Why it matters
- Who's exposed exposure With about 92% of quarterly revenue coming from data centers, Nvidia's forecast profits, and with them its 17x multiple, move with the capital budgets of a small group of AI infrastructure buyers.
- What it limits constraint The 70% forecast depends on manufacturing capacity as much as on demand, because Nvidia cannot book revenue on systems it has not yet been able to produce.
- Who has to decide decision A client who acts on DBS's no-bubble call still keeps part of the portfolio in defensives, so the bank is stopping short of recommending concentrated exposure to chip stocks.
Next story · Invest · yesterday
Nvidia's Huang says he would cut off an unsafe AI customer after a conversationConfirmed3 publishers
A forward multiple is a share price divided by a profit forecast. The roughly 17x Hou cited for Nvidia on October 5 [1][2] is only as firm as the forecast beneath it. The cryptobriefing report gives the same caveat: the figure assumes Nvidia delivers the expected profits, and the 70% revenue call is a DBS estimate [7]. Hold the share price still and cut next year's earnings by 30%, and the multiple becomes about 24x [8]. Halve them and it is 34x [9].
The dot-com comparison gives DBS the most room. According to the research behind the bank's view, peak dot-com valuations ran at about 100x earnings [3]. Nvidia at 17x is about a sixth of that. For today's price to carry a 100x multiple, forward earnings would have to come in about 83% below forecast [10]. Hou set the speculation of that era against the earnings visibility AI companies offer now [4].
The growth forecast already assumes a slowdown. The July quarter's 106% rise came off a year-earlier quarter of about $46.7 billion [16]. The 70% DBS projects for fiscal 2028 [12] is a lower rate on a far larger base. Fiscal 2028 begins in February 2027, so the July quarter belongs to fiscal 2027 [11]. The report does not give the fiscal 2027 total DBS is growing from. As a rough guide only, four times the July quarter is about $385 billion [17], and 70% on top of that is about $654 billion [18].
If estimates hold, the multiple stays near 17x and DBS is right on its own terms. If the supply limits management has described [6] push revenue past the 12-month window, the forward multiple climbs on an unchanged price even though demand is intact. If the large tech companies buying data-center systems [14] slow their orders, estimates and the share price fall together. In that case the 17x figure means little until the size of the cuts is known.
I think the case holds on the evidence given. Even halved profits leave Nvidia at about a third of the dot-com peak [9][3]. The counter-thesis is about scope. Every figure DBS cited is Nvidia's [2][12], but the verdict covers AI-related tech stocks as a group [1]. A 17x multiple shows investors paying a modest price for forecast profits. It does not show the forecasts are right. The thesis is wrong if Nvidia's forward estimates get cut while its share price holds, pushing the multiple toward the mid-20s [8].
What to watch
- Nvidia's fiscal third-quarter report: whether data-center revenue and management's supply commentary support a pace that leads into 70% growth in fiscal 2028.
- Capital spending plans from the largest tech companies buying AI infrastructure, the customers behind nearly all of Nvidia's revenue.
- Whether DBS revises its 70% fiscal 2028 forecast or shifts the weighting of its barbell toward defensives.
Information, not investment advice. This story is general information about companies and markets, drawn from cited reporting. It is not a recommendation to buy, sell or hold any security; check the cited sources before acting.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence55
- Adoption70
- Hype gap+20
- Incentives
- Not scored
- Confidence50
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
On October 5, 2026, DBS Chief Investment Officer Hou Wey Fook argued that Nvidia's earnings trajectory shows AI-related tech stocks have not drifted into bubble territory.
- [2]
Hou put Nvidia's 12-month forward price-to-earnings ratio at approximately 17x.
- [3]
Historical dot-com peak valuations ran at approximately 100x earnings, according to the research underpinning the DBS view.
- [4]
Hou contrasted the speculative nature of the pre-dot-com bubble with the earnings visibility AI companies currently offer.
- [5]
Hou recommends a barbell strategy, with one side holding tech and AI growth stocks and the other holding defensive assets meant to hold steady when markets turn choppy.
- [6]
Nvidia's management has described a supply-constrained environment, with demand outrunning what the company can produce.
- [7]
The roughly 17x forward P/E figure assumes Nvidia delivers on expected profits, and the 70% fiscal 2028 revenue projection is a DBS estimate, not a guarantee.
- [8]
At an unchanged share price, a 30% cut to forward earnings would lift Nvidia's forward multiple from about 17x to about 24x.
- [9]
At an unchanged share price, halving forward earnings would lift the multiple to about 34x, roughly a third of the 100x dot-com peak.
- [10]
Nvidia's 17x is about a sixth of the 100x dot-com peak; at an unchanged price, forward earnings would have to come in about 83% below forecast for the multiple to reach 100x.
- [11]
Because fiscal 2028 begins in February 2027, the quarter ended July 2026 falls within fiscal 2027.
- [12]
DBS projects Nvidia's revenue will climb 70% in fiscal 2028, a period that begins in February 2027.
- [13]
In its fiscal second quarter, which ended in July 2026, Nvidia reported $96.2 billion in revenue, up 106% from a year earlier.
- [14]
The data-center segment contributed nearly $89 billion of the $96.2 billion quarterly total; it sells the chips and systems that train and run AI models for the world's largest tech companies.
- [15]
Data centers supplied about 92% of Nvidia's fiscal second-quarter revenue.
- [16]
The year-earlier quarter's revenue was about $46.7 billion.
- [17]
Four times the July 2026 quarter's revenue is about $385 billion.
- [18]
Illustratively, if fiscal 2027 revenue equalled four times the July quarter, a 70% rise would put fiscal 2028 revenue near $654 billion.
Sources
2 independent publishers whose own reporting we read for this story.
- cryptobriefing.comNvidia’s earnings growth signals AI stocks are not in a bubble, says DBS
1 article · October 5, 2026
- qz.comDBS says Nvidia's valuation shows the AI rally isn't a bubble
1 article · October 5, 2026
Topics and entities
Topics
- Forward price-to-earnings ratio
- Barbell portfolio strategy
- AI Bubble and Debt-Bomb Debate
- AI data-center spending