To The Moon

Nvidia has once again reminded investors why it remains at the centre of the artificial-intelligence boom.

The latest quarter delivered another extraordinary set of numbers: revenue reached $96.22 billion, up 106% year over year, while data-centre revenue surged 117% to around $89 billion. Nvidia also guided to approximately $108 billion of revenue for the next quarter, while management expects around 70% revenue growth for the following fiscal year. (Reuters⁠)

At first glance, this looks like another straightforward Nvidia victory.

But there is a more interesting story underneath the headline numbers.

The AI boom is becoming so large that the biggest constraint may no longer be demand for Nvidia GPUs — it is the supply chain required to build them.

And that creates winners and losers far beyond Nvidia itself.

💰 The biggest issue: gross margins

Nvidia’s biggest question is no longer whether customers want its AI accelerators.

They clearly do.

The problem is how much it costs to build increasingly powerful AI systems.

Nvidia achieved an adjusted gross margin of roughly 75% in the latest quarter, but management expects margins to fall toward 71–72% in the January quarter, with margins around 72–73% in the following fiscal year. (MarketWatch⁠)

Why?

One major reason is memory.

Modern AI accelerators require enormous amounts of high-bandwidth memory, or HBM. As AI clusters become larger and more powerful, demand for HBM continues to increase.

The irony is fascinating:

The same AI demand that is driving Nvidia’s revenue higher is also increasing Nvidia’s component costs.

That means investors need to look beyond the GPU itself.

If Nvidia sells more AI systems, somebody has to manufacture the advanced chips, supply the HBM, provide networking equipment, build optical connections, assemble servers and provide the power infrastructure.

That is where the broader AI trade becomes interesting.

🧠 1. Nvidia — NVDA

Nvidia remains the core company in this entire ecosystem.

Revenue is exploding, demand continues to exceed supply, and the transition from Blackwell toward the Rubin platform provides another potential growth cycle.

Management’s statement that revenue could grow approximately 70% in the next fiscal year is particularly important because it suggests Nvidia does not believe the AI infrastructure boom is close to ending. (Reuters⁠)

However, investors should watch margins carefully.

If memory and other components become more expensive, Nvidia may have to accept lower margins or increase system prices.

That is why the stock’s next phase may depend less on simply beating revenue estimates and more on whether Nvidia can eventually turn its enormous AI demand into sustained free cash flow and margins.

💾 2. Memory stocks — Micron, SK Hynix and Samsung

If Nvidia’s margins are under pressure because memory is becoming more expensive, the other side of the transaction can potentially benefit.

This makes Micron (MU), SK Hynix and Samsung Electronics important stocks to watch.

HBM has become one of the critical bottlenecks in the AI hardware chain.

Recent market reactions after Nvidia’s results showed investors immediately looking toward memory companies. Micron and Sandisk both moved higher after Nvidia’s earnings, while Asian memory stocks such as SK Hynix and Samsung also benefited from the AI demand outlook. (MarketWatch⁠)

This creates an interesting dynamic.

Nvidia investors worry about rising memory costs.

Memory investors can potentially view the same development as evidence of pricing power and tight supply.

That is why the AI trade is not necessarily only about owning the GPU designer.

Sometimes the bottleneck can be where the better pricing power sits.

🏭 3. TSMC — the manufacturing backbone

Nvidia designs its own processors but relies heavily on external manufacturing and advanced packaging.

That makes TSMC (TSM) another important stock in the AI supply chain.

As AI accelerators become more advanced, manufacturing and packaging capacity become increasingly important.

The industry therefore has a simple equation:

More AI demand → more Nvidia GPUs → more advanced manufacturing → more packaging capacity → more semiconductor equipment demand.

This is why Nvidia’s success can spread across the semiconductor ecosystem.

🌐 4. Broadcom and Marvell — networking and custom AI infrastructure

AI data centres are not simply warehouses filled with GPUs.

Thousands of accelerators need to communicate with each other extremely quickly.

That creates enormous demand for networking, connectivity and custom silicon.

This makes Broadcom (AVGO) and Marvell Technology (MRVL) particularly interesting.

Nvidia itself has expanded its networking ecosystem, while Marvell has also become increasingly involved in custom AI infrastructure and silicon photonics.

Nvidia and Marvell announced a strategic partnership around NVLink Fusion, with Nvidia also investing $2 billion in Marvell. (NVIDIA Investor Relations⁠)

More recently, Marvell’s major Google partnership highlighted how hyperscalers are increasingly developing custom AI chips alongside Nvidia solutions. (Reuters⁠)

That creates two simultaneous trends:

Nvidia GPUs are growing rapidly, while custom AI accelerators are also growing.

The companies supplying the networking and connectivity layer could benefit from both.

🔌 5. Data-centre infrastructure — Vertiv and related stocks

There is another bottleneck investors sometimes overlook:

Power and cooling.

A modern AI data centre consumes enormous amounts of electricity and produces enormous amounts of heat.

Therefore, Nvidia’s growth indirectly increases demand for power distribution, cooling, electrical equipment and data-centre infrastructure.

This puts companies such as Vertiv (VRT) and other electrical/data-centre infrastructure suppliers on the radar.

The broader AI infrastructure opportunity is increasingly moving beyond semiconductors.

The next bottleneck could be electricity.

Then cooling.

Then transformers.

Then networking.

Then land and data-centre construction.

The AI buildout is becoming an entire industrial cycle rather than simply a chip cycle. (DBS Singapore⁠)

🚗 6. Tesla — TSLA

Tesla is another interesting Nvidia-related name.

Tesla has historically used Nvidia GPUs for AI training. Its Cortex supercomputer included approximately 50,000 Nvidia H100 GPUs, demonstrating just how important Nvidia hardware has been to Tesla’s autonomous-driving and AI ambitions. (Data Center Dynamics⁠)

That creates an indirect Nvidia relationship:

More Tesla AI training → more demand for computing infrastructure.

But there is an important twist.

Tesla is also developing its own AI chips and expanding its internal semiconductor capabilities.

The recently announced Texas Terafab project with SpaceX shows the long-term ambition to bring more semiconductor manufacturing capability in-house. The project is initially expected to involve around $16.8 billion of investment. (Reuters⁠)

So Tesla is both:

A customer of Nvidia today

and potentially

a competitor to Nvidia’s hardware dominance over the long term.

That makes TSLA an interesting secondary AI infrastructure stock rather than a pure Nvidia beneficiary.

🤖 7. Palantir — PLTR

Palantir is another stock that investors should not overlook.

Unlike Nvidia, Palantir does not primarily sell the physical AI accelerator.

It sits much higher in the stack.

Nvidia provides the computing infrastructure, while Palantir provides software that helps organisations actually deploy AI into real-world operations.

The companies have been strengthening their relationship, with Palantir integrating Nvidia accelerated computing, CUDA-X libraries and Nvidia’s AI models into its AI platform. (NVIDIA Newsroom⁠)

Palantir’s latest growth also demonstrates the demand for enterprise AI: revenue reached approximately $1.94 billion, up 93% year over year, while U.S. commercial revenue increased sharply. (MarketWatch⁠)

This is important because it shows another layer of the AI investment chain:

Nvidia sells the compute.

Data-centre companies provide the infrastructure.

Palantir helps companies turn that compute into business applications.

If AI adoption continues moving from experimentation into actual production, software companies such as Palantir could become increasingly important.

⚡ The bigger picture

The Nvidia story is therefore much bigger than one company.

Think about the AI ecosystem as a chain:

AI demand → Nvidia/AMD accelerators → HBM memory → TSMC manufacturing → Broadcom/Marvell networking → optical components → servers → power & cooling → data centres → AI software such as Palantir.

Every layer can benefit from the same underlying trend.

But every layer also has different risks.

Nvidia faces margin pressure.

Memory companies face potential supply cycles.

TSMC faces geopolitical risks.

Networking companies face competition.

Data-centre companies face electricity and construction constraints.

AI software companies face valuation risk.

And Tesla is attempting to build more of its own semiconductor capability.

📊 Stocks I would put on the AI watchlist

Core AI:

🔵 NVDA — AI accelerators and full-stack infrastructure

🔵 AMD — alternative AI accelerators

Memory:

🟢 MU — HBM and memory exposure

🟢 SK Hynix — major HBM exposure

🟢 Samsung — memory and semiconductor exposure

Manufacturing:

🟡 TSM — advanced semiconductor manufacturing

Networking/custom chips:

🟣 AVGO — networking and custom AI silicon

🟣 MRVL — custom AI silicon, networking and optical connectivity

Data-centre infrastructure:

🟠 VRT — power and cooling infrastructure

AI software:

🔴 PLTR — enterprise and government AI deployment

AI/autonomous systems:

🚗 TSLA — AI training, autonomy, robotics and custom-chip development

🎯 My takeaway

The most important lesson from Nvidia’s latest earnings is that AI demand has not disappeared — the bottleneck is changing.

Nvidia is still producing extraordinary growth, but rising memory costs show that the AI boom is putting pressure on the entire supply chain.

That could actually create a new rotation within the AI sector.

Instead of asking only:

“Should I buy Nvidia?”

Investors should also ask:

“What does Nvidia need in order to sell more AI systems?”

The answer includes memory, semiconductor manufacturing, networking, optical connectivity, servers, electricity, cooling and software.

That is why I think the next phase of the AI trade could be broader than Nvidia alone.

🚀 Nvidia may remain the leader — but the real opportunity could be following the bottlenecks.

And whenever the bottleneck moves, the winning stocks can move with it.

Not financial advice. These are stocks to research based on their exposure to the AI infrastructure theme, and each carries different valuation and business risks.

Find out more here: Tiger Brokers | Market Rebound: Rally or Pullback? Capture potential opportunities. Stay Flexible with Options

Market volatility is picking up. How can investors use options to navigate it? Explore different options strategies based on market conditions, current positions, and investment objectives, and complete the activity to unlock exclusive options rewards.

@AI_FocusedTrader @Daily_Discussion @TigerStars @TigerEvents