Nvidia (NASDAQ:NVDA) Record US$81bn Revenue, But Has the AI Trade Peaked?

Delivers 85% Growth, But the Market Still Shrugs

Is Nvidia now priced for perfection?

Each quarter, we cover Nvidia because it has become the foundation of the AI trade. It is the company investors look to for the clearest read on where AI demand, data centre spending and semiconductor growth are heading.

Nvidia is not just another chip company anymore. It is deeply embedded across the entire AI ecosystem. In many ways, it has become the market’s AI bellwether.

That is what makes the recent share price reaction so interesting.

Nvidia can beat non-GAAP EPS by US$0.10, deliver revenue of US$81.6bn and still grow at 85% YoY, yet the market barely reacts. We saw this last quarter as well. Strong numbers, strong guidance, and still, investors shrugged.

After hours, the stock was down 1.8%.

That tells us something important. The issue is not whether Nvidia is performing well. It clearly is. The issue is whether the market has already priced in near perfection, and relies on NVDA string outlook to keep the AI trade going.

When a company delivers numbers like this and the stock still falls, expectations have become extremely high. For Nvidia, good is no longer enough.

The Numbers Are Hard to Comprehend

A lot of analysts and investors have started arguing that GPUs could become obsolete as AI shifts from training to inference, where CPUs may play a larger role.

We do not think that is right.

GPUs should remain in demand as AI training, model development and inference workloads continue to scale. Nvidia is also not just a GPU company anymore. Its platform now includes CPUs within the broader accelerated computing stack, including Vera Rubin, which Nvidia has positioned as a processor platform built for the age of agentic AI & AI Infrastructure.

The real issue is not whether GPUs disappear. They will not.

The real issue is competition and growth deceleration.

AMD and Intel are both building more competitive CPU and accelerator products, while hyperscalers are increasingly investing in their own custom silicon. That adds pressure over time, even if it is not showing up in Nvidia’s numbers yet.

Gross margins are the clearest proof. Nvidia’s gross margin is still around 75%, compared with roughly 60% last year. If competition was already eroding the business, margins would be showing it. They are not.

The bigger concern is the growth rate.

The headline guidance still looks enormous. Revenue of US$91bn is a huge number by any standard. But when we look at the sequential growth profile, the slowdown is becoming clear.

• Q3 FY26 to Q4 FY26 revenue grew 24% QoQ.
• Q4 FY26 to Q1 FY27 grew 20% QoQ.
• Q1 FY27 to Q2 FY27 guidance implies 11% QoQ growth.

That is still strong growth, but the direction matters. Nvidia is not falling apart. It is simply moving from hypergrowth to slower growth, and for a stock priced close to perfection, that shift can matter a lot.

The Vera Rubin Margin Risk Is Underpriced

Management has guided for gross margins to remain stable, which the market appears to have taken as a reassuring signal.

But we think investors need to be careful here.

Vera Rubin production is expected to begin in H2 2026, which means Q3 FY27 should be the first quarter where Rubin starts to ramp meaningfully. That matters because new platform ramps can pressure margins, even when demand remains strong.

We saw this with Blackwell.

When Blackwell began scaling, Nvidia’s gross margin fell from 78.4% to 72%, a 6 percentage point compression over three quarters. That was not a demand problem. It was a product transition issue.

Rubin could create a similar dynamic.

The added risk is that Rubin includes a larger CPU stack, and that part of the market is facing more competition from AMD, Intel and custom hyperscaler silicon. So while Nvidia’s margins remain exceptional today, investors should not assume they stay untouched through the next product cycle.

The Networking Number Is Underappreciated

While we have identified some risks that investors may have glossed over, the release also highlighted something the market should not ignore. Networking revenue reached US$15bn, up 199% YoY.

This is the NVLink and InfiniBand infrastructure that connects GPU clusters together. It allows GPUs to talk to each other, transmit data quickly and operate as one large compute system rather than a collection of separate chips.

This is why Nvidia is not just a GPU commodity business. It is becoming an integrated compute and networking platform. As data centres buy more AI chips, connecting those chips efficiently becomes critical. The bigger the cluster, the more important the network becomes.

That is why networking is such an important segment to follow.

The fact that networking grew almost 3x faster than compute revenue tells us customers are not just buying more GPUs. They are building larger, more complex AI systems. That strengthens Nvidia’s ecosystem advantage because the value is no longer just in the chip. It is in the full stack that connects, optimises and scales the entire AI data centre.

Is NVDA a buy, hold or sell

So, what should investors in the semiconductor industry take away from this?

Nvidia has delivered almost five years of extraordinary growth. It has become the centre of the AI trade and one of the clearest signals for where the broader semiconductor cycle is heading. But the setup is now changing.

As AI moves deeper into the inference phase, competition is increasing. AMD, Intel and hyperscaler custom silicon are all trying to capture more of the value chain. Nvidia remains the clear leader, but the growth rate now looks to be decelerating. That is the key point.

We do not think Nvidia looks like a buy at current levels. The valuation has already stretched and the market reaction is telling. When a company beats expectations for two straight quarters and the share price still falls, it usually means investors have already priced in a lot of the good news.

Nvidia is still an exceptional business. But exceptional businesses can still become risky investments when expectations move too far ahead of reality. Investors should probable look at other names in the semiconductor space.