NVIDIA’s valuation looks compressed beside its recent history, while its AI business is still growing quickly. That mix matters to U.S. investors weighing whether current earnings can last as customers explore new options.
Gross margin is the share of revenue remaining after cost of revenue, before operating expenses. This report weighs the evidence and trade-offs around valuation, AI infrastructure, customer demand, and margin pressure without predicting a share price.
The useful question is whether earnings growth can outlast margin pressure and new customer alternatives.
Key Takeaways
- A forward P/E below 17 reflects estimates for future earnings, not a promise of future returns.
- Strong revenue growth can occur alongside lower gross margins.
- Data Center results are central to NVIDIA’s reported growth.
- Custom chips give some customers another option, but do not prove lost sales.
- Cloud spending plans matter most when they become working capacity and sustained workloads.
- Investors can compare results with guidance and a chosen forecast period.
Nvidia Forward PE Under 17 Margin Custom Chips 2026: What the Valuation Means
Yes, the October 2026 snapshot puts NVIDIA’s forward P/E below 17, at 13.67. NVIDIA shares closed at $216.00 on October 26, 2026, a dated view of the stock and its expected earnings.
A forward P/E compares a share price with expected earnings for a stated future period. The October multiple describes a relationship between price and forecast earnings, not a target price.
Treat that multiple as a dated reading, not a standing valuation or a buy signal.
The number can look low when expected earnings are large relative to the share price. It can also reflect doubt about whether those earnings will hold up.
The stock query connects valuation, margins, and custom chips because each can affect future earnings. That connection is a question about forecast durability, not proof that the multiple predicts a price move.
Investors also need to know which earnings period sits in the calculation. A ratio based on one forecast period cannot be compared cleanly with a ratio based on another.
The forward P/E is one valuation measure, not a full account of business quality or risk. Investors can read it alongside reported results, company guidance, and assumptions about future demand.
Its meaning also depends on the estimates used and the date of the share price. Treat that multiple as a dated reading, not a standing valuation or a buy signal.
How Has NVIDIA’s Forward P/E Changed Since 2025?
NVIDIA’s forward P/E fell from 39.84 in July 2025 to 22.89 in July 2026, with a modest increase in between. The overall direction was lower, though the series did not decline every period.
The forward P/E dropped across the listed dates, with a small April uptick. The table shows each dated reading in order.
| Date | Forward P/E | Place in the dated sequence |
|---|---|---|
| July 31, 2025 | 39.84 | Starting reading |
| October 31, 2025 | 31.54 | Lower than July |
| January 31, 2026 | 24.34 | Lower than October |
| April 30, 2026 | 24.58 | Small increase |
| July 31, 2026 | 22.89 | Lowest listed reading |
A falling multiple can come from different forces. Expected earnings may rise faster than the share price, the price may fall, or both may happen together.
That distinction matters because a lower ratio alone does not reveal which part moved. A stock can keep rising while its P/E declines if expected earnings rise faster.
Forecast updates and market prices can shift at different speeds. Investors comparing a stock with its past valuation should check whether the earnings period stayed consistent.
The modest April rise interrupts the falling series, but does not erase its broader direction. Compare valuation snapshots only when their measurement dates and earnings periods match.
What Earnings Estimates Support the Valuation?
The average diluted EPS estimate rises from $9.31 for fiscal 2027 to $15.91 for fiscal 2028. These are analyst estimates, not reported results or company guarantees.
The fiscal 2028 average is higher, but future earnings remain uncertain. The estimates reflect the views of 51 analysts for fiscal 2027 and 52 analysts for fiscal 2028.
| Fiscal year | Average diluted EPS estimate | Analyst count |
|---|---|---|
| Fiscal 2027 | $9.31 | 51 |
| Fiscal 2028 | $15.91 | 52 |
EPS means earnings per share. Diluted EPS includes the effect of shares that could become common stock, such as certain stock awards.
A forward multiple depends on the earnings period used in its denominator. A P/E tied to next year’s estimate may differ from one using a later year’s forecast.
Analysts can revise estimates as new results and guidance arrive. If expected EPS rises while the share price stays unchanged, the apparent valuation can fall without a matching price move.
The analyst count also differs between the two fiscal years. That does not make one estimate more accurate by itself, but it shows each average draws on a separate group of forecasts.
Forecasts can help investors understand what earnings assumptions support a valuation. They do not show whether NVIDIA will meet those assumptions or how the market will value the results.
Check the forecast period before comparing a P/E headline with an EPS estimate.
Are Revenue Growth and Margins Moving in the Same Direction?
Fiscal 2026 revenue grew 65%, while both annual gross-margin measures declined from fiscal 2025. The different directions show that sales growth does not automatically mean stronger unit economics.
GAAP gross margin was 71.1%, down 3.9 percentage points, while non-GAAP gross margin was 71.3%, down 4.2 points. These measures use different accounting adjustments and should not be blended.
Total fiscal 2026 revenue reached $215.938 billion. Data Center revenue was $193.7 billion, up 68% year over year, making that business the main source of company sales.
That scale puts Data Center at the center of NVIDIA’s results. Yet revenue growth alone does not prove lasting pricing power, since costs and product mix can also affect gross margin.
- Revenue scale: $215.938 billion in sales shows that growth rests on a very large base.
- Margin direction: Both measures fell, signaling that more sales did not lift the share left after cost of revenue.
- Data Center contribution: Its $193.7 billion in revenue shows how strongly results depend on AI-related demand.
GAAP results follow standard accounting rules. Non-GAAP results adjust selected items, so each margin should be compared with the same measure across periods.
For investors, the key issue is whether sales expansion can keep producing sound economics. Read revenue growth alongside margin direction to see whether expansion is translating into stronger unit economics.
What Do Fiscal 2027 Results and Guidance Say About Demand?
Q2 fiscal 2027 revenue reached $96.221 billion, up 106% year over year, while the Q3 outlook calls for $108.0 billion, plus or minus 2%. The reported quarter and outlook point to strong demand, with different levels of certainty.
Q2 Data Center revenue was $89.0 billion, up 117% year over year. That result locates the main source of quarterly demand in Data Center.
| Period | Revenue | GAAP gross margin | Non-GAAP gross margin |
|---|---|---|---|
| Q1 outlook | $78.0 billion, plus or minus 2% | 74.9%, plus or minus 50 basis points | 75.0%, plus or minus 50 basis points |
| Q2 reported | $96.221 billion | 75.0% | 75.0% |
| Q3 outlook | $108.0 billion, plus or minus 2% | 74.0%, plus or minus 50 basis points | 74.0%, plus or minus 50 basis points |
The Q1 and Q3 revenue outlooks assumed no Data Center compute revenue from China. Those assumptions define what the two guidance periods included.
NVIDIA began including stock-based compensation expense in non-GAAP measures in Q1 fiscal 2027. The Q1 non-GAAP margin outlook included a 0.1% impact from that change.
Q2’s reported revenue and margins can be compared with the earlier outlook and the next guide. Yet reported results and forecasts are not the same kind of evidence.
The Q3 revenue outlook rises from Q2’s reported level, while the margin outlook is lower. Investors can track whether actual results meet both expectations as each quarter closes.
A strong sales figure does not describe the full quality of a quarter. Judge each new quarter against both its revenue guide and its margin guide, not revenue alone.
How Much of AI Chip Supply Does NVIDIA Hold?
NVIDIA held an estimated 63% of AI chip supply by computing power sold or shipped in Q2 2026. This supply-share estimate measures capacity, not physical chip units.
The measure uses H100-equivalents, a way to compare AI computing power sold or shipped. It does not count chips already installed or deployed.
| Designer | Estimated share in Q2 2026 | What the measure counts |
|---|---|---|
| NVIDIA | 63% | Computing power sold or shipped |
| 25% | Computing power sold or shipped | |
| Amazon | 5% | Computing power sold or shipped |
| AMD | 4% | Computing power sold or shipped |
| Huawei | 3% | Computing power sold or shipped |
H100-equivalents put different AI chips on a common computing-power basis. They do not show how many physical chips each designer shipped.
They also do not count all chips already installed in data centers. Sales and shipments measure supply entering the market, while deployments reflect what buyers have put to work.
The estimate shows NVIDIA held the largest share in that period. It does not guarantee the company will hold the same share in later quarters.
Customer demand, supply, and competing products can change over time. For more context on the companies across the AI stock supply chain, investors can compare their different business roles.
Use the H100-equivalent measure to compare computing capacity, not to infer a count of chips.
Could Custom Chips and Rising Input Costs Pressure Pricing Power?
Customer-developed silicon and memory costs create different risks: one can change the alternatives available to buyers, while the other can raise system costs. Both can affect pricing power, but through separate paths.
Custom chips may give large customers more choices, while memory costs can raise the cost of building AI systems. Neither point alone proves that NVIDIA’s earnings will fall.
How Google, Amazon, and Meta Are Developing Alternatives
Google TPU, Amazon Trainium, and Meta in-house AI chips are examples of customer-developed silicon. These chips could give their developers alternatives for particular workloads.
A custom chip is a processor designed by a customer for particular workloads. The design can fit a company’s needs, but it must also work within its software and data-center systems.
These projects do not show that customers have displaced NVIDIA at scale. Their competitive effect depends on how much usable capacity reaches the market and which tasks the chips can handle.
Three factors help clarify the possible effect:
- Shipment scale: More units in service could make an alternative matter more to a buyer.
- Workload fit: A chip built for a narrow task may not replace general-purpose computing.
- System integration: Software and data-center links can shape how easily customers use the chip.
Why Memory Costs Matter for Gross Margins
Commodity DRAM contract prices rose 93%-98% quarter over quarter in Q1 2026. A further 58%-63% increase was projected for Q2 2026.
Memory is one part of the cost of an AI system. If its price rises and sellers cannot pass that cost to buyers, gross margins may face pressure.
Later market estimates also pointed to continued memory price increases. TrendForce forecast conventional DRAM contract prices would rise 10%-15% quarter over quarter in Q4 2026, in its Q4 memory price outlook.
Cost pressure can fall on different parts of the supply chain, depending on contracts and product design. Treat a custom-chip announcement as a competitive signal, not proof of lost NVIDIA revenue.
Can Cloud Spending Sustain AI Infrastructure Demand?
Four major cloud providers’ projected combined 2026 capital expenditure is substantially higher than their 2025 figure. Amazon, Microsoft, Alphabet, and Meta projected $760 billion for 2026, compared with $413 billion in 2025.
That plan signals expected investment, not completed purchases or proven returns. NVIDIA’s chip demand depends partly on buyers’ willingness and ability to build AI infrastructure.
Capital expenditure, often called capex, can pay for data centers, servers, and related equipment. Higher plans can support demand for chips, but project timing affects when purchases happen.
Large budgets also do not show whether new capacity will stay busy. Providers need ongoing workloads and funding to keep investment moving beyond initial construction.
Investors can follow four signs to judge whether the spending case is becoming more durable:
- Revised provider capex plans: Higher or lower budgets can change expected infrastructure demand.
- Data-center capacity delivery: Delays can push equipment purchases or use into later periods.
- Sustained AI workload use: Ongoing demand can support capacity after it opens.
- Ability to fund investment: Buyers need enough cash or financing to continue building.
These checks separate stated budgets from actual buying and use. A spending plan does not establish that providers will earn a return on every new facility.
The four providers’ projected total is a major demand signal, but it remains an expectation. A budget announcement supports the demand case only when spending turns into deployed capacity and continuing workloads.
Is It Time to Buy or Sell NVIDIA Stock?
There is no single best date to buy or sell NVIDIA that follows from a forward P/E reading alone. The right decision depends on an investor’s outlook, risk limits, and time horizon.
A low P/E by itself does not show that the earnings outlook will hold. NVIDIA trades publicly under ticker NVDA and reports both GAAP and non-GAAP results.
When Is It Reasonable to Buy NVIDIA Stock?
The best time to buy NVIDIA stock is when an investor’s thesis, time horizon, and tolerance for volatility fit the business outlook. A lower valuation can be one part of that case, not the whole case.
A buyer can ask what would support expected earnings over the chosen holding period. Company guidance and reported results give concrete measures to compare with that expectation.
Investors who want broader context can review AI stocks across the industry and compare their revenue sources and risks. Different companies can benefit from AI spending in different ways.
When Might Selling NVIDIA Stock Make Sense?
Selling or trimming may fit when the original earnings thesis weakens, the position exceeds risk limits, or an investment goal changes. A short-term price move alone does not settle whether the business case has changed.
Is it time to sell NVIDIA? Investors can compare their thesis with reported results and guidance, then decide whether those facts still support their expectations.
Exposure also matters. A large position can create more portfolio risk than an investor intended, even when the underlying company remains strong.
This is a decision framework, not personal financial advice. It cannot account for an individual investor’s savings needs, tax situation, or full portfolio.
Before placing an order, write down what evidence would change your view and how much loss you could tolerate.
What Can a Forward P/E Tell You, and What Can It Miss?
A forward P/E is only as comparable as the share-price date and forecast period behind it. The ratio divides share price by estimated EPS for a chosen period.
A trailing P/E uses EPS already reported, while a forward P/E uses expected earnings. Estimate providers, measurement dates, and periods can produce different readings, so the inputs matter.
A low ratio does not establish that future earnings are secure. Forecasts can change even when a company’s current results remain strong.
Before comparing a valuation headline, ask:
- Which share-price date does the calculation use?
- Which earnings period sits in the denominator?
- Which analyst estimate set supplies the forecast?
- Does the EPS measure use GAAP or non-GAAP results?
These details can change the ratio without changing the company’s current business. A forecast based on a later year may look cheaper because it assumes more earnings.
That difference is useful only when investors see the assumptions clearly. Compare like-for-like measurements before drawing a conclusion from a valuation headline.
Conclusion
NVIDIA’s lower multiple makes the stock worth examining, but does not settle whether its earnings path is durable. Investors need to connect valuation with the earnings assumptions and operating results behind it.
Choose one forecast period and compare NVIDIA’s next reported results and guidance with that expectation. This gives investors a clear way to judge whether the business is tracking the case they considered.
Put your forecast period and risk limit in writing, then revisit both when NVIDIA reports again.
FAQ
Why can two websites show different NVIDIA forward P/E figures?
Two websites can show different NVIDIA forward P/E figures because they may use different measurement dates, earnings periods, or analyst estimates. One calculation might use a coming fiscal year, while another uses a later year. Share prices also change over time. Comparing the displayed ratio is more useful when both websites use the same date and forecast period.
What is the difference between forward and trailing P/E?
Forward P/E uses estimated EPS for a future period, while trailing P/E uses EPS already reported. The forward ratio can change when analysts revise their forecasts, even if the share price stays the same. The trailing ratio relies on past results, so it does not show what analysts expect next. Each measure answers a different valuation question.
Does NVIDIA’s 63% AI chip supply share mean it shipped 63% of all chips?
No, NVIDIA’s 63% share measures H100-equivalent computing power sold or shipped, not physical chip count or deployments. H100-equivalents compare computing capacity across different AI chip designs. They do not show how many chips each company sold or how many buyers installed. The figure describes estimated supply share for the stated period, not a future market guarantee.
Why can NVIDIA’s P/E fall even if its share price does not?
NVIDIA’s P/E can fall while its share price holds steady if forecast EPS rises. The ratio compares price with expected earnings, so a larger earnings estimate lowers the ratio’s value. Analysts may revise forecasts as they assess results and demand. A lower P/E can reflect stronger expected earnings, though it does not prove that forecasts will be met.
Should investors use GAAP or non-GAAP gross margin?
Investors should check which gross-margin measure they are comparing and use the same one across periods. GAAP follows standard accounting rules, while non-GAAP results adjust selected items. NVIDIA began including stock-based compensation expense in non-GAAP measures in fiscal 2027 Q1. That change affects how investors should read comparisons across the reporting shift.
What is a practical way to track NVIDIA’s next report?
Compare NVIDIA’s reported revenue and gross margin with the company’s guidance for the same period. Check GAAP and non-GAAP margins separately, since the measures differ. Revenue can beat expectations while margin performance misses, or the reverse can happen. Looking at both results gives investors a clearer read on demand and the economics of sales.
