To The Moon

#Reward: Tech Stocks — Buy the Dip or Run for the Exit?

My answer: I would buy the dip selectively—but I would not blindly chase every AI stock.

This does not necessarily mean that the AI story is ending. Now the market has moved from asking, “Is AI real?” to asking, “Which companies can convert AI spending into sustainable revenue, profit and cash flow?”

AI investment is finally producing measurable returns

The latest results from the major cloud companies provide strong evidence that AI capital expenditure is beginning to generate real commercial returns.

Microsoft reported quarterly Microsoft Cloud revenue of $59.3 billion, up 27% year on year, while Azure and other cloud-services revenue increased 43%. More importantly, its commercial remaining performance obligations reached $678 billion, up 84%. This represents contracted business that should be recognized progressively as revenue—not merely optimistic management projections.

Amazon delivered an equally important signal. AWS revenue increased 37% to $42.2 billion, its fastest expansion in more than four years. AWS operating income rose from $10.2 billion to $16.6 billion, showing that cloud growth is translating into higher profit instead of only higher expenditure.

Alphabet $GOOGL$ also demonstrated stronger AI monetization. Google Cloud revenue rose 82% to $24.8 billion, while operating income more than tripled to $8.8 billion. Its cloud backlog reached $514 billion, supported by demand for AI infrastructure and enterprise AI solutions.

In my view, these results answer one of the market’s biggest concerns: Big Tech is no longer spending on AI without seeing a return. Cloud growth, higher operating profit and expanding contractual backlogs show that customers are paying for AI capacity.

The strength is visible throughout the supply chain

If AI demand were weakening, we should see warning signs among the companies supplying the infrastructure. Instead, the latest results remain strong.

Nvidia $NVDA reported first-quarter fiscal 2027 revenue of $81.6 billion, up 85%, with data-centre revenue increasing 92% to $75.2 billion.

AMD $AMD$ reported a 107% increase in data-centre revenue to $6.7 billion, supported by demand for EPYC processors and Instinct accelerators.

Broadcom $AVGO$ generated $10.8 billion of AI-semiconductor revenue in its fiscal second quarter, representing growth of 143%. Its custom AI accelerators and networking products show that hyperscalers are building more diversified AI infrastructure instead of relying on only one type of processor.

The contracts and partnerships are also becoming larger and longer-term. AMD has separate strategic agreements supporting planned deployments of up to six gigawatts of GPUs for OpenAI and another six gigawatts for Meta. Nvidia and OpenAI announced plans for at least 10 gigawatts of Nvidia systems, representing millions of GPUs, while Nvidia intends to invest progressively as each gigawatt is deployed.

These are not small experimental orders. They are multi-year infrastructure commitments measured in gigawatts.

Why memory remains my highest-conviction segment

Within the AI supply chain, memory is my preferred long-term theme.

Processors perform the calculations, but memory and storage hold the model parameters, training data, inference context and newly generated information. Even if improvements in chip efficiency eventually slow the growth in accelerator units, the amount of memory and storage required per AI system can continue increasing.

Micron $MU$ estimates that the high-bandwidth-memory market could expand from approximately $35 billion in 2025 to around $100 billion by 2028—an annual growth rate of about 40%.

Micron has also signed 16 strategic customer agreements that generally run from 2026 through 2030. Collectively, these agreements cover approximately 20% of its DRAM volume and one-third of its NAND volume over the period. This gives Micron greater demand visibility than it had during previous memory cycles.

Sandisk $SNDK$ has disclosed eight agreements with six customers worth a combined $93.9 billion, with an average duration of four years. By fiscal 2028, the company expects around two-thirds of its output to be sold under long-term agreements.

Kioxia, meanwhile, expects the NAND supply-demand balance to remain tight throughout 2027 because of strong AI data-centre demand and constrained industry supply.

The evidence also extends beyond semiconductor memory. Seagate $STX$ - which supplies high-capacity hard-disk storage—reported fiscal third-quarter revenue of $3.11 billion and free cash flow of $953 million. Management believes AI-driven data creation is supporting a new period of structural storage growth.

HBM, conventional DRAM, NAND flash and hard drives serve different purposes, but together they tell the same story: more AI usage creates more data, and more data must be processed, moved and stored.

Think about how our personal storage needs have changed. Years ago, a 2GB pen drive felt sufficient. Today, even a 1TB drive can fill up quickly with photos, videos, applications and backups. Now multiply that progression across autonomous vehicles, humanoid robots, AI agents, smart factories and millions of enterprise users.

An autonomous vehicle may eventually generate terabytes of sensor data. A robot must continuously process visual, audio and movement information. AI agents will create and retrieve documents, images, videos and transaction histories. The long-term storage requirement therefore extends well beyond training large AI models.

Strong fundamentals do not eliminate volatility

I am bullish, but I do not expect technology stocks to move upward in a straight line.

Valuations are elevated, capital expenditure is enormous and some companies are experiencing free-cash-flow pressure. Export restrictions, electricity constraints, higher bond yields and a future increase in semiconductor capacity could also affect margins and share prices.

Recent market reactions illustrate this risk. Some storage companies fell sharply even after reporting strong revenue because their guidance did not exceed already-extreme expectations. At this stage of the cycle, being “good” is sometimes insufficient—the market expects results to be exceptional.

That is why I prefer accumulating strong companies during corrections instead of chasing them after a vertical rally.

Geopolitical conditions are another important variable. If tensions ease, lower oil, freight and supply-chain risk could reduce inflation pressure. In turn, lower inflation and bond yields would generally support the valuations of long-duration growth companies. However, geopolitical improvement would be an additional tailwind—not the foundation of my AI thesis.

My conclusion: Buy the dip, but be selective

I do not believe the recent volatility proves that the AI bubble has burst. The latest cloud earnings, chip demand, multi-year deployment plans and long-term memory-supply agreements indicate that the underlying demand remains strong.

My preferred approach is to invest gradually, focus on companies with real revenue and cash flow, and avoid assuming that every company carrying an “AI” label will succeed.

Among the different layers of the ecosystem, memory remains my strongest conviction because every new AI model, autonomous vehicle, robot and digital service will require more bandwidth, capacity and storage.

The road will be volatile, but the direction of global data creation remains upward.

For me, this is not “run for the exit.” It is “buy the dip carefully, hold for the long term and let the fundamentals decide.”

This reflects my personal investment view and is not financial advice.

Insta: marketlensdaily89

 $AMZN$ $GOOGL$ $MU$ $SKHY$