Most people think AI is driven primarily by processors. In reality, every AI model also depends on enormous amounts of memory, networking capacity, data-centre infrastructure, electricity and cooling. As AI systems become larger and more sophisticated, each of those components is becoming a critical bottleneck. Markets have always rewarded companies that solve bottlenecks. Railways needed steel. The internet needed fibre-optic networks. Electric vehicles need batteries. Artificial intelligence increasingly needs memory.<\/p>\n
Only three companies dominate advanced high-bandwidth memory globally: Micron, SK Hynix and Samsung. As demand has exploded, supply has struggled to keep pace, transforming what was once viewed as a cyclical corner of the semiconductor industry into one of the hottest parts of global markets.<\/p>\n
The shortage has now become so acute that some analysts have labelled it \u201cRAM-ageddon\u201d. The world\u2019s largest technology companies are effectively competing with each other for scarce memory chips, with AI data centres consuming an ever-growing share of global supply.<\/p>\n
Perhaps the strongest evidence of memory\u2019s growing importance came not from Micron itself, but from Apple and Microsoft. Both recently announced price increases on a range of products after unprecedented rises in memory costs, while manufacturers including Dell, HP, Lenovo and Asus have either lifted prices or reduced memory specifications to offset soaring component costs. When even Apple \u2013 widely regarded as having one of the world\u2019s most sophisticated supply chains \u2013 is forced to pass on higher component costs, it tells you something important about where pricing power now sits.<\/p>\n
In other words, the AI boom is no longer simply lifting technology share prices. It is beginning to influence prices paid by consumers and businesses around the world. What was once viewed as a relatively commoditised part of the semiconductor industry has rapidly become a strategic constraint on the global economy.<\/p>\n
History suggests this should not surprise investors.<\/p>\n
During California\u2019s gold rush, some of the greatest fortunes were made not by the prospectors themselves but by businesses supplying the picks, shovels and durable clothing. Levi Strauss became one of the world\u2019s most recognisable brands without mining an ounce of gold. Technology revolutions tend to work the same way. The internet boom did not simply create Microsoft and Cisco. It produced winners across fibre optics, networking equipment, semiconductors, cloud computing and eventually e-commerce. Artificial intelligence increasingly appears to be following the same path.<\/p>\n
In many respects, this is exactly what investors are witnessing today. The extraordinary profitability now being enjoyed by Micron reflects the same shortages forcing Apple and Microsoft to raise prices. Pain for technology manufacturers has become gain for memory producers.<\/p>\n
Yet every time another AI stock surges, comparisons with the dotcom bubble inevitably return. On the surface, the comparison is understandable. Technology companies dominate headlines, a handful of stocks are driving market returns, and artificial intelligence has become one of the world\u2019s biggest investment themes. But beneath the surface, today\u2019s market looks fundamentally different.<\/p>\n
During the late 1990s, valuation discipline largely disappeared. Investors famously stopped talking about price-to-earnings ratios and instead embraced \u201cprice-to-eyeballs\u201d, valuing internet businesses on website traffic rather than profits. Companies with little revenue, no earnings and often nothing more than a \u201c.com\u201d suffix attracted billion-dollar valuations.<\/p>\n
Today\u2019s AI leaders are generating extraordinary profits. Nvidia trades on around 20 times forward earnings despite growing earnings by more than 70% year-on-year. Micron and SK Hynix, both now valued at more than US$1 trillion, trade on fewer than nine times expected earnings over the next 12 months despite being among the biggest beneficiaries of AI infrastructure spending. Those are not valuation multiples normally associated with speculative bubbles. If anything, they suggest investors remain surprisingly disciplined despite extraordinary earnings growth.<\/p>\n
There are signs that valuation discipline has returned more broadly across the technology sector. Reports suggest OpenAI is increasingly pushing any blockbuster public listing towards 2027 rather than rushing to market. That is not because demand for artificial intelligence has faded. Quite the opposite. It reflects a market that has become far more discerning about valuation. Following the more muted post-listing performance of SpaceX, investors appear less willing to pay any price for even the world\u2019s highest-profile growth companies.<\/p>\n
The money is still there. The blind faith is not.<\/p>\n
There is another reason today\u2019s market feels different.<\/p>\n
There is an old story that Joseph Kennedy decided to sell his shares shortly before the Wall Street Crash after receiving stock tips from a shoeshine boy. His conclusion was simple: if everyone was speculating, there were few buyers left.<\/p>\n
I often think about that story because I experienced something remarkably similar while living in Dublin during the technology boom of the late 1990s. Almost without fail, every week taxi drivers would offer unsolicited recommendations for the latest internet stock that \u201ccouldn\u2019t miss\u201d. Investing had become part of everyday conversation, and there was a widespread belief that technology shares could only go higher.<\/p>\n
That is not to say speculation does not exist today. It clearly does. But the conversation feels fundamentally different. Most discussions are about how artificial intelligence will reshape businesses, which companies will benefit most and whether earnings can justify valuations, rather than simply chasing the next stock tip.<\/p>\n
None of this removes the need for caution. The memory industry remains highly cyclical, competition will intensify and today\u2019s shortages could become tomorrow\u2019s oversupply. Leadership within AI will almost certainly continue to rotate.<\/p>\n
But that is precisely the point.<\/p>\n
The biggest winners from transformational technologies rarely remain confined to the first obvious company. As industries mature, opportunities migrate deeper into the supply chain as new bottlenecks emerge.<\/p>\n
Nvidia may have ignited the AI revolution.<\/p>\n
Micron\u2019s latest result suggests the next wave of winners is already emerging.<\/p>\n
The lesson for investors is that the biggest opportunities in a technological revolution rarely remain confined to the first obvious company. Sometimes the real fortunes are made by those quietly supplying the picks and shovels.<\/p>\n
Generate is a New Zealand-owned KiwiSaver and Managed Fund provider managing over $9 billion on behalf of more than 190,000 New Zealanders.<\/p>\n
This article is intended for general information only and should not be considered financial advice. The views expressed are those of the author. All investments carry risk, and past performance is not indicative of future results.<\/p>\n