Mexico Surges Into Elite Trillion-Dollar Ranks Riding Artificial Intelligence Demand Wave

SK Hynix and Micron Technology both crossed the $1 trillion market capitalisation threshold this week, making them the latest entrants into a valuation club that now spans ten companies globally and signals that AI infrastructure spending has structurally repriced the memory chip sector, not just the logic chip sector.

SK Hynix shares surged 10% on Wednesday, capping a rally that has more than tripled its stock price since the start of 2026. Micron's shares rose nearly 20% on Tuesday after UBS tripled its price target for the company. The two moves, separated by 24 hours, mark a milestone: memory, long treated as a commodity business subject to brutal cyclical swings, now commands the same valuation territory as consumer platform monopolies.

UBS's price target revision on Micron was the proximate trigger for that company's single-day move. The decision by one of the Street's largest research operations to triple a price target in a single revision, rather than ratchet it upward in increments, signals that sell-side models had systematically underestimated the duration of AI-driven memory demand. The implied message: this is not a cyclical spike. It is a structural repricing.

The so-what is this: the $1 trillion threshold is not merely a symbolic milestone. It changes the institutional ownership universe for both companies, forces index rebalancing across funds benchmarked to large-cap indices, and repositions memory chips as a critical infrastructure asset class rather than a discretionary technology holding.

AI Data Centres Have Created a Memory Shortage, Not Just a Logic Chip Shortage

For three decades, the semiconductor investment narrative centred on logic chips: microprocessors, graphics processors, application-specific silicon. Memory was the unglamorous substrate, priced per gigabyte, hammered by oversupply whenever fabs ran hot. That framing is now obsolete.

The surge in global demand for advanced chips powering AI tools has created a memory chip shortage that is lifting revenue for manufacturers including SK Hynix and Micron. The mechanism is straightforward: large language models and AI inference workloads require massive, high-bandwidth memory capacity co-located with compute. Every Nvidia GPU shipped into a data centre requires accompanying high-bandwidth memory (HBM) stacks. SK Hynix is a key supplier to Nvidia. That supply relationship transforms SK Hynix from a cyclical commodity producer into a critical node in the AI supply chain.

Our view: the shortage dynamic is unlikely to resolve quickly. HBM production requires converting existing DRAM capacity, which tightens conventional memory supply simultaneously. Both markets tighten at once. Manufacturers with HBM capability capture margin expansion on two fronts.

The $1 Trillion Club Now Has a Memory Wing

Samsung joined the club earlier in May 2026, becoming only the second Asian firm to reach the milestone after Taiwan's TSMC. Samsung is now valued at approximately $1.34 trillion, with its shares more than doubling since the start of this year. A separate development reinforced Samsung's position: union members voted in favour of a pay deal this week, averting a strike that had threatened production continuity, sending Samsung shares up more than 6% on Wednesday.

The implication for index construction and institutional allocation is material. Three Asian semiconductor companies, Samsung, TSMC, and SK Hynix, now sit above $1 trillion. For investors benchmarked to the MSCI World or equivalent global indices, the weighting of Asian semiconductor names has increased substantially. Passive rebalancing flows will follow.

Geopolitical Risk Has Turbocharged the Demand Signal

The AI chip boom does not exist in a geopolitical vacuum. The Iran war, which began on February 28, 2026, when the United States and Israel launched strikes on Iran, has disrupted global energy markets and supply chains. Oil prices rose to approximately $100 a barrel from a pre-war level of approximately $66 a barrel. The Strait of Hormuz, through which one-fifth of the world's oil and natural gas supplies were shipped before the conflict, remains under Iranian control, with a US naval blockade of Iranian ports in place.

The energy cost impact on semiconductor fabs is direct. Wafer fabrication is energy-intensive. South Korean fabs operated by SK Hynix and Samsung face higher operating costs as energy prices rise. That cost pressure, if sustained, could compress margins even as revenue per unit climbs. Bangladesh's experience illustrates the broader emerging market stress: the country, home to 170 million people, imports 95 percent of its oil and LNG needs, raised fuel prices by 10 to 15 percent in April, and is now seeking IMF assistance. Its ready-made garment industry, which accounts for more than 80 percent of Bangladesh's export earnings, has been disrupted by shipping route changes through the Red Sea and Middle East.

The connection to semiconductor investment is second-order but real: supply chain disruption in Asian manufacturing corridors elevates the strategic premium on companies that control AI-critical components. Memory is one of those components.

SK Hynix's Nvidia Relationship Is the Structural Anchor

The relationship between SK Hynix and Nvidia is the central value driver. SK Hynix supplies high-bandwidth memory to Nvidia, which has become the dominant infrastructure provider for AI data centres globally. Nvidia became the first company to reach a $5 trillion stock market valuation in October, continues to report record sales, and faces growing competition in the sector.

The competitive threat to Nvidia matters for SK Hynix primarily if alternative AI accelerator architectures require less HBM, or source it from competing suppliers. Samsung also counts Nvidia among its customers. That means Nvidia's purchasing decisions effectively determine margin and volume allocation between the two Korean giants.

Our view: SK Hynix's more than tripling in share price since January 2026 likely reflects the market pricing in a multi-year HBM supply agreement duration with Nvidia. The risk is not demand destruction; it is supply normalisation. If Samsung or Micron close the HBM technology gap, pricing power for SK Hynix moderates. The speed of that normalisation is the key variable for investors entering at current valuations.

Investment Positioning: What Institutional Capital Should Model Now

For PE and institutional capital, the memory chip re-rating creates three actionable themes.

First, the supplier ecosystem below SK Hynix and Micron remains fragmented and undervalued. Chemical suppliers, equipment manufacturers, and substrate producers that feed HBM production have not experienced equivalent multiple expansion. The value chain trade is to identify these names before the re-rating propagates downstream.

Second, AI infrastructure spending is creating demand concentration risk. If hyperscaler capital expenditure programmes pause or compress, the entire memory supply chain faces simultaneous inventory correction. The 2022 to 2023 memory downcycle erased billions in market value within months. Current valuations embed a soft-landing assumption for AI capex. Investors should size positions accordingly.

Third, the geopolitical environment, specifically the Iran war's effect on energy prices and supply chains, raises operating cost risk for Asian fabs. A sustained oil price above $100 per barrel compresses fab margins and could force production prioritisation between commodity DRAM and premium HBM. That prioritisation benefits HBM pricing but tightens conventional memory supply further.

Some investors have warned of a potential AI bubble, questioning whether these companies are overvalued. That concern is not unfounded. Nvidia at $5 trillion, two memory makers above $1 trillion each, and an energy-disrupted global economy represent a combination that warrants scenario analysis across multiple demand outcomes.

The Plocamium View

The market is pricing SK Hynix and Micron as if HBM demand is perpetual and supply is permanently constrained. Neither condition is guaranteed. What the market is not pricing adequately is the regime change in how memory is categorised by institutional capital allocators.

For two decades, memory was a cyclical sector. Portfolio managers reduced exposure ahead of supply gluts and re-entered near trough valuations. That mental model is now a liability. HBM is not commodity DRAM. It cannot be commoditised quickly because the manufacturing process, bonding stacks of dies with through-silicon vias at yield rates that matter commercially, is technically differentiated and capacity-constrained by equipment availability, not just capital spending intent.

The Plocamium thesis is this: the correct analogue for HBM today is not DRAM in 2018. It is TSMC in 2020, when the market began to recognise that leading-edge foundry was a structural monopoly, not a cyclical business. TSMC's valuation re-rating from that point was sustained over multiple years. SK Hynix, as the leading HBM supplier to the world's dominant AI chip company, occupies an analogous structural position.

The second-order play that the source coverage misses: if SK Hynix and Micron sustain $1 trillion-plus valuations, South Korea and the United States will face intensifying political pressure to subsidise and protect these companies as national strategic assets. That political dynamic could result in government-backed capex support, export control reciprocity, or preferential procurement, all of which further entrench incumbent positions. The Iran war, by elevating energy price risk and supply chain fragility, accelerates that political logic. Governments watching Bangladesh seek IMF assistance over energy import costs will not miss the lesson that controlling critical technology supply chains is as important as controlling energy supply chains.

Investors who treat SK Hynix and Micron as momentum trades will exit too early. The structural case for sustained premium valuations is intact, provided AI infrastructure capex holds. The risk to monitor is not demand, it is the speed at which Samsung closes the HBM technology gap and introduces pricing competition at the top of the stack.

The Bottom Line

SK Hynix and Micron entered the $1 trillion club this week because AI data centre demand has structurally repriced memory from commodity to critical infrastructure. UBS tripling its Micron price target in a single move, and SK Hynix shares more than tripling since January, are not anomalies. They reflect a market catching up to a regime change that began when hyperscalers committed to multi-year AI infrastructure build-outs. The immediate catalyst is clear; the durable question is whether HBM supply constraints persist long enough to justify terminal values embedded in current multiples. Plocamium's answer is yes, for now, with one caveat: watch Samsung's HBM yield ramp. When Samsung closes the gap with SK Hynix, the premium compresses. Until then, memory is no longer a cyclical trade. It is an infrastructure position.