A Demand Cycle That Refuses to Cool
The current wave of artificial intelligence demand is unlike anything markets have seen in years, and the language used by industry leaders to describe it has matched the moment. Explosive, exciting, accelerating — these are not marketing flourishes but accurate descriptions of what is happening on the ground. When you look at the backdrop of hyperscaler spending, the partnerships forming around chip capacity, power supply, and networking equipment, it becomes apparent that both AI adoption and AI usage are climbing in parallel. The leaders of the chip industry have every reason to keep delivering upbeat commentary, and the supply chain confirms it.
That said, enthusiasm carries a hidden risk. The so-called whisper numbers — the unofficial expectations that float above published consensus — have a tendency to run overly aggressive for a company like Nvidia. The froth around the name means that even strong earnings can occasionally fall a hair short of these unofficial targets, triggering near-term pullbacks. The stock can be squeezed lower in the short run by expectations that were never realistic to begin with.
A Multi-Year Chart That Speaks for Itself
The longer-term picture, however, tells a different story. Nvidia traded around $29 just three years ago and now sits near $221. Against a trajectory of that magnitude, most dips have proven to be buying opportunities. The complication today is technical. Like many other leadership names, Nvidia is trading well above the key support levels investors typically watch — the 50-day, 100-day, and 200-day moving averages. Those usual reference points are simply too far below the current price to be useful. The 20-day moving average is now a more practical near-term gauge for where the stock may find footing.
But provided AI adoption and usage continue to climb across consumers, enterprises, and governments, demand for AI compute, data center capacity, and networking will keep expanding. That dynamic plays directly into the hands of Nvidia, but also Marvell, Broadcom, and a handful of others positioned in the same supply chain.
AI as a Lever — and as an Excuse
The labor market story attached to all of this is more complicated than the headlines suggest. AI is going to upend, if not outright disrupt, how work gets done, because the people and companies adopting it are unlocking real productivity gains. The more productivity you can squeeze out of each worker, the fewer workers you ultimately need. Anyone working day-to-day in a knowledge-intensive role can already feel this — whether by leaning on ChatGPT, Claude, or voice-to-text tools like Whisper to bypass keyboard input. These are no longer novelties; they are quietly compounding into real workflow changes.
At the same time, there is a fair suspicion that some companies are leaning into AI as a convenient excuse to trim headcount and become leaner — even when they do not have a clear internal picture of what that AI strategy actually looks like. Call it AI washing if you like. The reality is probably a mix: genuine automation gains running alongside opportunistic cost cutting under an AI banner.
The numbers attached to these decisions are not small. One major social platform recently announced layoffs of roughly 10%, or about 8,000 people. A major international bank disclosed plans to cut more than 15% of its support staff by 2030, from a base of around 52,000 employees. Between now and the end of the decade, that bank — and many like it — will be reshaping its hiring patterns, possibly pivoting what hiring remains toward AI-related roles. They are not alone, and more announcements of this type are coming.
More Disruptive Than the Internet
Where does this end up over a longer horizon? The fair comparison may be the early internet — and the comparison flatters AI rather than diminishing it. When the internet first arrived, it was effectively a glorified digital catalog. You could not buy things, you could not pay for things, you could not stream, you could not do nearly any of the activities we take for granted today. The capability set expanded over years, and with it, the disruption.
AI has the potential to be even more disruptive. Adoption is one metric worth tracking, but usage is the variable that matters most. If usage replicates even a fraction of what the internet ultimately delivered in terms of economic transformation, the ceiling is enormous. That implies a multi-year buildout of AI infrastructure, data center capacity, and networking capacity — not a one- or two-quarter story.
A Rising Tide and a Sleeper Theme
Because the buildout is long and broad, a rising tide will lift many boats over the next one to two years. Numerous companies in chips, infrastructure, and platforms will benefit simultaneously. Picking individual winners becomes harder several years out, once usage growth begins to mature and the field narrows.
The most overlooked piece of the trade may be networking equipment. Everything we do on the internet, in streaming, and in AI requires connecting to a network. That makes networking infrastructure a quiet but essential beneficiary of every dollar spent on compute. It is the sleeper category that does not get the same attention as the marquee chip names, even though no AI workload runs without it. Long portfolio positioning aimed at this theme naturally clusters around Nvidia, Google, Broadcom, Meta, and Marvell — a mix that captures compute, platform, and connectivity.
The bullish positioning is visible in the options market as well. Call volume running roughly four-to-one over puts on a name like Nvidia is unusually lopsided, signaling how aggressively traders are leaning in. Some commentators have floated valuations approaching six trillion dollars, a number that would have seemed absurd not long ago.
Geopolitics, Rates, and Retail Cross-Currents
The AI story does not exist in isolation. Geopolitical risk remains a live variable, with Iran signaling it will escalate and expand its posture if the United States and Israel push harder. A self-imposed deadline of just a few days — Friday into the weekend, possibly Monday — means the next several trading sessions deserve close attention. The rate environment is another open question. Yields are elevated, but whether the highs are already in is genuinely unclear.
In the meantime, what is driving day-to-day market behavior, while everyone waits on the next signal from the AI giants, is retail earnings. Strong comparable sales prints — a 6% comp at one major off-price retailer — have rightly captured attention. But the standout, on a monthly adjusted basis, remains the leading warehouse club, whose comps continue to run even stronger. The consumer is not as exhausted as some narratives suggest, and that resilience matters for everything sitting above it in the market.
The Bottom Line
The picture that emerges is of an economy mid-transition. Infrastructure is being built at a scale that will support years of compute-intensive work. Productivity tools are quietly changing how individuals operate, while corporations adjust their headcount strategies to reflect — or to claim to reflect — that change. Some of the labor disruption is real automation; some of it is rhetorical cover. Markets are pricing in enormous upside in the names tied most directly to the buildout, and the technical setup demands more careful timing than in earlier phases of the rally. The longer story, though, is straightforward: more AI capacity, more data center capacity, more networking capacity. The sky may not literally be the limit, but for now it is hard to see the ceiling.