Nvidia’s Outlook for the Next AI Spending Will Move the Entire Market

That wraps up our initial coverage of Nvidia’s Q2 results. Thank you for stopping by!

Nvidia delivered the kind of quarter that makes the near-term bear thesis much harder to defend. Revenue reached $96.2 billion, crushing the $92.4 billion consensus, while adjusted EPS hit $2.22 and gross margin held at 75%.

Nvidia’s $108 billion Q3 revenue outlook came in about $3.8 billion above consensus. That guidance confirms Blackwell 300 shipments and the early Vera Rubin ramp are translating directly into revenue without meaningful deployment bottlenecks or customer pushback.

Even better, the outlook assumes no Data Center compute revenue from China. If H200 shipments eventually resume, Nvidia could still have billions of dollars of upside that are not reflected in its current forecast.

While Nvidia stock is down 1% after earnings, the company’s AI infrastructure monopoly looks as untouchable as ever.

Nvidia ended Q2 with $93.9 billion in marketable and non-marketable equity securities, up from just $35.1 billion at the beginning of the fiscal year. The company purchased $42.4 billion of equity securities during the first six months of the year alone.

Those investments generated $7.8 billion in net gains during Q2, contributing materially to GAAP earnings.

The portfolio’s rapid expansion makes Nvidia increasingly dependent on the valuations of the same AI ecosystem it is helping finance, giving investors another reason to scrutinize the quality of its investment gains and customer demand.

Nvidia guided for $108 billion in Q3 revenue, nearly $4 billion above Wall Street’s $104.2 billion estimate. This guidance assumes no Data Center compute revenue from China.

That suggests Blackwell, networking, and other global AI demand remain strong enough to sustain sequential growth without help from a major potential market.

Vera Rubin is also entering full production, with racks already operating at Nvidia’s partners. The main pressure point is profitability: Nvidia expects adjusted gross margin to slip from 75% in Q2 to 74% in Q3, which may explain why shares initially fell despite the broad earnings beat.

Nvidia just reported earnings, with shares initially down 2% following the report. Here are the key numbers:

• Revenue: $96.22 billion vs. $92.17 billion expected
• Adjusted EPS: $2.22 vs. $2.10 expected
• Data Center Revenue: $89.0 billion, up 117% year over year
• Gross Margin: 75%

Guidance:

• Revenue: $108 billion vs. $104.19 billion expected

Quick Read:

• Nvidia comfortably beat revenue and earnings expectations, while its Q3 revenue outlook cleared Wall Street’s consensus by nearly $4 billion.
• Data Center revenue reached $89 billion and gross margin held at the critical 75% level, but the stock’s initial decline suggests investors entered the report expecting an even larger beat.

Nvidia received approval to sell H200 chips into China months ago, but shipments have reportedly remained minimal.

That leaves around $3 billion in potential revenue unavailable as investors focus on the company’s next-quarter outlook.

China revenue would offer upside to guidance without requiring additional demand from U.S. hyperscalers.

If shipments remain frozen, Nvidia may need even stronger Blackwell and networking growth elsewhere to deliver the outlook Wall Street expects.

Nvidia reportedly reduced the memory configuration in its next-generation Rubin platform as component availability tightened and memory prices increased.

The change highlights how supply constraints can affect product specifications even when customer demand remains exceptionally strong.

Investors will listen closely for a commitment to keeping gross margin near 75%. If higher memory costs, pricing decisions, or a less powerful Rubin configuration pressure profitability, the margin outlook could outweigh another headline earnings beat.

JPMorgan remains bullish with a $280 price target, but sees these near-term risks potentially driving volatility first.

With 5 minutes to Nvidia’s Q2 earnings release, Wall Street is looking for Q2 FY27 EPS of $2.0887 on revenue of roughly $92.07 billion, above NVIDIA (NASDAQ:NVDA)’s own $91.0 billion midpoint guide.

Shares closed the regular session at $209.66, down 1.59%, and are off 3.04% over the past week heading in. Polymarket puts the odds of a headline beat at 95.4%.

History urges caution: NVIDIA’s earnings-day change was negative after 4 of the last 5 beats, averaging -1.58%. The real trigger tonight is Q3 guidance clearing the $108 billion bar, gross margin holding at 75%, and confirmation that Vera Rubin production begins in Q3 as promised.

Bull Case

• NVIDIA (NASDAQ:NVDA) has posted four consecutive EPS beats, with Q1 revenue of $81.61 billion, up 85.23% year over year.
• Polymarket assigns a 0.96 probability to a Q2 beat, and Data Center above $85 billion at 0.87.
• Locked-in demand from $119.0 billion in supply commitments and Blackwell 300 plus Vera Rubin ramps support the AI factory buildout thesis.

Bear Case

• Despite five straight beats, the stock’s average day-of change was -1.58%.
• China Data Center compute is excluded from guidance, and a substantial cash-tax increase hits Q2.
• At a P/E of 33, expectations are steep. Polymarket prices 0.986 odds NVDA finishes down today.
• Amazon (NASDAQ:AMZN) Trainium and custom silicon threaten the accelerator moat.

Tonight’s Q3 outlook will likely drive the market’s reaction to Nvidia’s Q2 earnings. Wall Street is anchoring on a Q3 revenue guide above $108 billion, well above the $91 billion plus or minus 2% Q2 bar management set in May.

NVIDIA (NASDAQ:NVDA) typically guides conservatively and beats by 3% to 6%. Investors want commentary on Data Center trajectory, gross margin sustainability at 75%, Vera Rubin timing, and any China H20 path. Supply commitments now sit at $145 billion, signaling future demand.

• Bullish scenario: Q3 revenue guide meaningfully above $100 billion, margins holding, and Vera Rubin pulled forward.
• Bearish scenario: a guide at or below $95 billion, margin compression below 74%, or hyperscaler digestion concerns.

Polymarket assigns a 0.92 probability that adjusted gross margin lands in the 74%–76% band.

Google, Amazon, and Broadcom are developing custom AI accelerators that could eventually capture as much as half of the market, according to JPMorgan’s framework.

These chips don’t necessarily need to outperform Nvidia everywhere, because they might just need to handle specific workloads cheaply.

That makes competitive positioning important in tonight’s earnings commentary. If management cannot show that Nvidia’s performance, software, and networking advantages will protect its market share, investors could lower the multiple they are willing to pay for future growth.

Cooling typically accounts for roughly 40% of a data center’s electricity consumption, making Vera Rubin’s fully liquid-cooled architecture a potentially major economic advantage.

The closed-loop system can operate with coolant temperatures as high as 45 degrees Celsius, reducing the need for energy-intensive chillers and consuming no additional water after installation.

Industry estimates suggest increasing coolant temperature by just one degree can reduce cooling energy costs by about 4%.

For a 50-megawatt facility, that could translate into more than $4 million in annual savings, helping customers deploy more AI computing within the same power constraints.

Vera Rubin’s modular, cable-free, hose-free, and fanless trays could reduce rack assembly time from around two hours to just five minutes.

The new design also allows equipment that previously required six rack units to fit into two. Faster assembly and greater rack density could help Nvidia ship systems more quickly, let customers install more computing power within existing facilities, and produce a much steeper Rubin ramp than what Blackwell achieved.

Vera Rubin could expand Nvidia’s addressable market far beyond the accelerators that built its AI dominance.

Its Vera CPU will be available as a standalone server processor, giving Nvidia access to an estimated $200 billion market it has never directly served.

Management expects nearly $20 billion in standalone CPU revenue this year, with every major hyperscaler and system manufacturer working to deploy Vera.

That means the Rubin cycle could deepen Nvidia’s control over the entire AI computing stack while creating a major new revenue stream independent of GPU sales.

This live blog is being updated by Thomas Richmond, a 24/7 Wall St. contributor. Simply stay on this page, and new updates will appear below automatically.

We expect Nvidia to release Q2 earnings shortly after 4:20 p.m. ET.

Options traders are pricing a roughly 5.5% move in Nvidia shares following tonight’s earnings report, with implied volatility sitting near 42%. Yet recent history suggests even that forecast could be generous.

Nvidia has not finished outside the options market’s implied earnings range since May 2024, when the company was worth approximately $2.3 trillion. That pattern has rewarded traders who sold volatility around recent reports.

A move beyond 5.5% tonight would therefore mark a meaningful break from Nvidia’s increasingly subdued post-earnings reactions.

Investors may press Jensen Huang about Nvidia’s role in a proposed $500 billion financing effort for AI data centers and chip purchases.

Nvidia reportedly retains the option to backstop as much as 25% of eligible deals, implying potential support of up to $125 billion.

The proposal could expand the pool of capital available to Nvidia’s customers and sustain demand for its hardware. It also strengthens bears’ concerns about circular financing, particularly if Nvidia begins funding the same infrastructure spending that ultimately produces its revenue.

Any clarification about the program’s structure and Nvidia’s financial exposure could move the stock tonight.

Nvidia currently has four September appearances scheduled, culminating in a September 22 business update at Nvidia AI Day Singapore.

That conference calendar gives Nvidia several opportunities to expand on Vera Rubin, robotics, AI factories, and its broader growth strategy after earnings.

It could also keep event-driven volatility elevated well beyond tonight, particularly if Jensen Huang uses those appearances to announce new partnerships or provide additional product and demand updates.

Top 5 Analyst Questions:

1. Will Q2 revenue beat the roughly $92 billion consensus estimate and Nvidia’s $91 billion guide?
2. Can Data Center revenue surpass the expected $85.7 billion as hyperscaler AI spending accelerates?
3. Will Q3 revenue guidance come in above the $108 billion level Wall Street increasingly expects?
4. Can gross margin remain near 75% despite higher memory, packaging, and supply-chain costs?
5. Is the Blackwell 300 ramp progressing smoothly, and will Vera Rubin production begin on schedule?

Key Topics, Buzzwords, and Red Flags:

• Key topics management must address: Blackwell 300 shipments, Vera Rubin timing, Data Center networking growth, China restrictions, and the durability of hyperscaler AI spending.
• Buzzwords to listen for: “supply-constrained,” “sovereign AI,” “agentic AI,” “Vera Rubin,” and “accelerated computing.”
• Red flags: Low Q3 guidance, gross margin slipping materially below 75%, slower networking growth, Blackwell supply bottlenecks, or cautious commentary about customer spending.

Nvidia is valued near $5.2 trillion, with options traders pricing a roughly 5.4% post-earnings move, equivalent to approximately $280 billion in market value.

Options traders are pricing an approximately 5.4% move in Nvidia shares following tonight’s report.

With Nvidia valued near $5.2 trillion, that translates into roughly $280 billion of potential market value moving in either direction.

Technically, $216.40 is the bullish level to watch. A breakout could reopen a path toward Nvidia’s record-high range, while a close below $195.95 would invalidate the emerging reversal setup and potentially expose the stock to a considerably deeper correction.

Nvidia is expected to report approximately $92 billion in Q2 revenue and adjusted earnings of $2.08 to $2.09 per share, ahead of management’s $91 billion midpoint.

However, what investors really want to see is the company’s outlook for Q3.

Some analysts expect Q3 revenue guidance to come in above $108 billion if Nvidia delivers another beat-and-raise quarter.

Investors want confirmation that Blackwell demand remains strong, Rubin is on schedule, and gross margin can hold near 75%.

Data Center revenue remains the big number to watch when Nvidia reports tonight at 4:20 PM ET.

Wall Street expects roughly $85.7 billion, more than double the year-ago figure, as hyperscalers, sovereign governments, and enterprises continue pouring money into AI infrastructure.

A strong beat would suggest AI demand remains constrained more by chip supply than customer appetite.

Any slowdown, however, could revive fears that the industry is approaching peak AI spending and pressure the broader semiconductor sector alongside Nvidia.

Nvidia enters tonight’s Q2 earnings report with a $91.0 billion revenue guide and expectations for a major Blackwell 300 ramp.

Data Center networking revenue nearly tripled year over year last quarter, making the sustainability of that growth another critical test. The outlook still excludes China Data Center compute revenue, leaving a major source of potential demand outside the forecast.

A clean beat and strong Q3 guidance tied to Vera Rubin production would support Jensen Huang’s estimate that annual AI infrastructure spending could reach $3-4 trillion by the end of the decade.

Soft networking growth or cautious margin guidance could send the stock lower, as concerns could resurface about the durability of hyperscaler spending into 2027.