Nvidia 2026 Forecast: Does NVDA Have Room for US$295?

The Nvidia 2026 forecast is back at the centre of Wall Street conversation after the quarter released on 20 May: the company delivered US$81.6 billion in revenue, +85% year-over-year, and projected US$91 billion for the next quarter. With the Rubin roadmap entering production in the second half, the average analyst target has risen to US$295 — but noise about a possible AI bubble has also grown.

TL;DR

• NVDA trades around US$213.95 on 27 May 2026, with a market cap of US$5.28 trillion.
• Q1 FY27: US$81.6 billion revenue (+85% YoY), adjusted EPS of US$1.87 — above consensus.
• Q2 FY27 guidance: US$91 billion (+95% YoY), well above the expected US$87.2 billion.
• Average target of 61 analysts: US$295.69; range from US$180 (-16.2%) to US$500 (+133.7%).
• Vera Rubin is already running in production on Microsoft Azure; volume shipments begin in the second half of 2026.
• Bear case: AI capex reaches US$539 billion in 2026 — real risk of digestion.

NVDA today: the stock price in May 2026

On 27 May 2026, Nvidia's stock (NASDAQ: NVDA) trades around US$213.95, oscillating between US$212.00 and US$218.36 throughout the session. Over the past 12 months, the stock has moved in a wide band — a low of US$132.92 and a high of US$236.54 — reflecting two opposing facts: the continuation of the data centre capex cycle and growing nervousness around the valuation of the entire AI chain.

The market cap touched US$5.28 trillion, a number that keeps Nvidia as the world's most valuable company, ahead of Microsoft and Apple. The level is historic, but it needs to be read against the revenue base that supports it. And it is precisely this base that has changed in the last two weeks.

Q1 FY27: the quarter that rewrote estimates

On 20 May 2026, Nvidia reported the first quarter of fiscal year 2027 (February to April). The numbers, in order of relevance:

• Total revenue: US$81.6 billion, versus US$78.8 billion expected by the FactSet consensus — up 85% YoY.
• Adjusted EPS: US$1.87, beating the consensus of US$1.76.
• Gross margin: rose to the 72% range, compared to 43% in the equivalent cycle last year.
• Q2 FY27 guidance: US$91 billion, +95% YoY, and nearly US$4 billion above the consensus of US$87.2 billion.

The most important detail is not the result itself: it is the guidance. Nvidia rarely has linear quarters; when it targets US$91 billion for the next quarter, it is signalling that demand for Blackwell continues to accelerate — and that the first orders for Rubin are already appearing in the pipeline. The four major hyperscalers (AWS, Google Cloud, Microsoft Azure, and Oracle Cloud) have already publicly made capacity allocations.

What moved the price right after the result

The stock reacted curiously: it rose more than 4% in after-market trading and gave back most of it in the following session. It wasn't due to the quality of the number — it was because part of the market was already working with US$90 billion as internal guidance. When the official consensus was US$87.2 billion, but the "buy-side whisper" was US$92 billion, US$91 billion effectively became just an "in-line".

Price target: what the 61 analysts covering NVDA say

Nvidia is one of the most covered stocks in the US market. According to S&P Global, 61 analysts maintain a "Strong Buy" consensus on the stock, with an average target of US$295.69 for the next 12 months — implying 37.6% above the price on 27 May.

The dispersion is wide:

Metric	Value	Implied vs US$213.95
Average target	US$295.69	+37.6%
Median target (estimate)	US$280.00	+30.9%
Lowest target	US$180.00	-16.2%
Highest target	US$500.00	+133.7%
Range (high/low)	2.78x	—

Two details help read this table. First, the range is wide because assumptions vary greatly: the bear case discounts a sharp slowdown in capex in 2027; the bull case extrapolates the Rubin cycle through 2028 and considers a new product generation in 2029. Second, even the lowest target (US$180) does not assume a collapse — it assumes a compressed multiple, not plummeting revenue.

There are also more speculative analyses. One of them, published by The Motley Fool in May 2026, projects the stock at US$357 by year-end if Nvidia trades at 40 times forward earnings — which would give a potential gain of 66% from the current level. It is a possible scenario, but it requires both earnings and the multiple to cooperate at the same time.

Vera Rubin: the roadmap that supports the bullish scenario

The Nvidia 2026 forecast is not just about Blackwell. It is about what comes next.

At GTC 2026 in March, Jensen Huang announced that the first rack of Vera Rubin — the direct successor to Blackwell — was already operating in production on Microsoft Azure. Volume production begins in the second half of 2026, with AWS, Google Cloud, Microsoft Azure, and Oracle Cloud announced as first customers to offer cloud instances.

The platform consists of six chips and promises to reduce inference cost per token by up to 10 times compared to Blackwell. In terms of training, Nvidia claims that one NVL72 rack with Rubin trains the same models with 1/4 of the GPUs compared to the previous generation — and delivers 10 times more inference throughput per watt.

Why does this change the thesis? Because the unit cost falls, but Nvidia maintains premium pricing. More value delivered to the customer, same margin for the company. Huang projected at GTC US$1 trillion in combined Grace Blackwell + Vera Rubin revenue by 2027 — a number that two years ago would sound absurd, but with US$91 billion per quarter becomes a reasonably conservative extrapolation.

The detail that part of Wall Street ignores

Those who are mid-cycle on the GB200 (the best-selling Blackwell configuration) will likely finish the rollout in 2026 and adopt Rubin only in 2027. The supply of HBM4 memory — required for Rubin — does not have enough volume to serve the enterprise market before mid-2027. The "H2 2026" window for Rubin is, in practice, exclusive to hyperscalers. This favours the bull case, because it guarantees allocation for higher-margin customers and pushes enterprise orders to FY28, extending the cycle.

The bear thesis: US$539 billion capex and the risk of digestion

None of this eliminates the noise. The most aggressive estimates point to US$539 billion in AI capex in 2026 — summing Microsoft, Meta, Google, Amazon, Oracle, CoreWeave, and the constellation of neoclouds. It is a number that makes analysts ask: what if real inference demand does not keep up with investment?

The counterarguments usually come from four fronts:

1. Revenue concentration. About 40% of Nvidia's data centre revenue comes from four customers. Any capex revision in one of them has an amplified effect on the top line.
2. Circular financing. Nvidia invested in CoreWeave, which pays Nvidia. It invested in OpenAI, which pays Microsoft, which pays Nvidia. The loop is not fraud, but it inflates the apparent growth of the sector.
3. Software sell-off. The AI software sector (Snowflake, MongoDB, Datadog) suffered a sharp correction in February 2026 — a sign that end customers have not yet monetised what they promised to their own investors.
4. Multiple. NVDA trades at around 30x forward earnings. In almost any other sector this would be expensive; in AI chips, it is the recent historical average. The question is how much of that average remains valid in a higher interest rate cycle.

The countermeasure comes from Morgan Stanley, which described the "fear of bubble" as premature. The argument: the median cash flow and capital reserves of the 500 largest US companies are about three times larger than in classic bubble periods (dot-com, sub-prime). Companies are funding capex with their own generation, not with debt — a financially healthy pattern, even if the cycle is excessive in some names.

How Brazilian investors access NVDA

For those in Brazil, there are three main routes — each with different frictions:

• BDR NVDC34. Traded on B3, in reais, with reasonable liquidity. The advantage is simpler income tax and no direct currency cost. The disadvantage is the spread relative to the US stock, which can reach 1-3% in times of high volatility.
• International broker (Avenue, Inter, Nomad, XP US). Direct purchase of NVDA on NASDAQ, with IOF on currency conversion and income tax via monthly DARF on gains. Good option for larger tickets and for those who trade frequently.
• ETF. IVVB11 (S&P 500 with currency hedge) or sector ETFs like SMH and SOXX (US semiconductors) provide diversified exposure — a way to capture part of the thesis without concentrating everything in a single stock.

For those entering now, two points matter more than the entry price: position size (Nvidia is already more than 7% of the S&P 500, so anyone with a US ETF is probably already exposed without realising it) and horizon (the thesis is 2-3 years, not one quarter). Trading NVDA with a weekly horizon is virtually playing volatility — something few individual investors do well.

What to monitor until December 2026

The Nvidia 2026 forecast will depend more on five events than on any isolated analyst target:

1. Q2 FY27 results (August 2026): the test is whether Nvidia delivers the US$91 billion guidance and revises the second half upward.
2. Start of Rubin volume shipping (Sep–Nov 2026): official announcements with measurable units, not just "first racks".
3. Hyperscaler capex guidance in Q3: Microsoft, Google, and Meta release projections for 2027 — any downward revision is a correction trigger.
4. HBM4 supply: Samsung, SK hynix, and Micron need to deliver volume. A delay here pushes the Rubin revenue curve to 2027 and breaks the US$1 trillion by 2027 thesis.
5. Fed stance: the rate cut or stabilisation cycle changes the multiple applicable to growth stocks. Each 25 bps of hawkish surprise typically costs 3-5% in NVDA's multiple.

For Brazilian investors, it is worth monitoring the exchange rate in parallel. BDR and direct stock react differently to Real vs Dollar movements — in 2025 this caused NVDC34 to underperform NVDA at certain points, even with the underlying asset rising.

Those who want to better understand the macroeconomic backdrop of generative AI can revisit our coverage on AI agents and what Gemini Spark changes for businesses and the analysis of how Nvidia's silicon also powers the Nintendo Switch 2 — two perspectives that show how broad Nvidia chip consumption is today, from hyperscale data centre to portable console.

Conclusion: the Nvidia 2026 forecast in one sentence

If the consensus is right, NVDA will close 2026 between US$270 and US$300 — up 26% to 40% from the current level. If the Rubin cycle brings forward revenue and Q2 guidance comes in above US$91 billion, there is room for US$350 or more. If hyperscaler capex begins to slow or Blackwell digestion drags on, the US$180–200 range returns to the map.

The good news for those who invest rather than speculate is simple: the core thesis — Nvidia as monopolistic AI infrastructure for one more product cycle — remains intact. What has changed in the last quarter is the margin of safety, which has decreased. And it is precisely at this point that the investor needs to decide whether to enter now, scale in monthly instalments, or wait for the next volatility trigger to build a position.

This text does not constitute investment advice. Financial decisions depend on each investor's risk profile, horizon, and objectives — consult a CVM-certified advisor before trading.