How concentrated is the market behind Micron mania?

The mad dash for GPUs gets lots of attention, but DRAM shouldn’t be overlooked. Dynamic random-access memory chips power everything from datacentres to smartphones. And right now, DRAM is HOT.

The past month has been a particularly big one for DRAM. Two weeks ago, Micron — responsible for the majority of DRAM chips made in the US — posted a 15-foldincrease in earnings. Bulls are gleefully declaring Micron the new Nvidia, but the memory mania is spreading. One South Korean memory maker is planning a US listing next week.

A single ETF that invests in DRAM makers (yes, its ticker is in fact DRAM) has accumulated over $24bn in assets since launching in February, with nearly $10bn coming in the last month alone. Apple has hiked MacBook and iPad prices, blaming high chip costs, and last week, the FTrevealed that Apple was lobbying the Trump administration to remove Chinese DRAM-maker ChangXin Memory Technologies from a US trade blacklist.

The result of all this is that analysts have in recent weeks jacked up earnings forecasts to wild levels.

All this could point to the conclusion that the DRAM market is a bit oligopolistic. After all, a diverse, competitive industry doesn’t typically enjoy these kinds of earnings upgrades. But just how concentrated is the memory business?

The answer — depending on your interpretation of South Korean industrial policy — is somewhere from “highly concentrated” to “VERRRY highly concentrated” (admittedly not a technical antitrust term, yet).

To calculate market concentration, Alphaville is going to use the measure favoured by US antitrust regulators at the Department of Justice and Federal Trade Commission: the Herfindahl–Hirschman Index.

HHI is typically used to measure the concentration of producers in a single national market, but for our purposes, we’ll measure the international market to make an important point about how concentrated DRAM is today.

The measure takes the summed-squares of each producer’s market share to estimate the concentration of a market. A market in which a monopolist owns 100 per cent gets an HHI of 10,000, a duopoly scores above 5,000, and a perfectly competitive market approaches zero. In other words, the higher the HHI, the more highly concentrated the market. The US DoJ considers anything between 1,000 and 1,800 points to be “moderately concentrated”.

Per Counterpoint Research, as of the first quarter of 2026 the memory market is 38 per cent Samsung (South Korea), 29 per cent SK Hynix (South Korea), 22 per cent Micron (US), 8 per cent CXMT (China), 2 per cent Nanya (Taiwan), and 1 per cent everyone else.

This gives us an HHI of 2,838. That’s comfortably above the 1,800 level where the US Department of Justice considers a market “highly concentrated” and subjects any acquisitions to extra antitrust scrutiny.

But even this, of course, assumes that each of the listed companies operate as independent, profit-seeking entities. The industrial policy of South Korea might beg to differ.

Much has been written on different forms of corporate competition and collaboration, such as the vast literature of export cartels, the influential model developed by James Brander and Barbara Spencer on the use of state export subsidies to gain international market share, and an Asia-focused literature spearheaded by Chalmers Johnson on what he calls the “developmental state” model.

State subsidies and a competitive, export-driven approach mean that South Korea fits squarely in the latter category. Both Samsung and SK are two of Korea’s largest conglomerates (the so-called chaebols) which benefit from cosy relations with the state, so viewing them as fierce competitors in the same memory market might be wrong-minded in this case.

Just this week, Samsung and SK announced a joint $590bn investment plan alongside the Korean government to build chips. This initiative is part of President Lee Jae Myung’s signature initiative to build the country’s semiconductor, data centre and robotics industry — which he termed a “Great Leap Forward”. We didn’t pick the branding.

In other words, South Korean industrial policy means that South Korean companies aren’t really competing with each other to reach the most customers — they’re competing with Chinese and US companies to establish national market dominance.

And if the two companies function as a single economic entity in the global DRAM market, we should probably count them together for HHI purposes. And doing this we get a much higher HHI reading of 5,042. That’s above 5,000 —the HHI of a perfect duopoly.

What does this all mean in practice? The FTC and DoJ obviously can’t break up foreign enterprises, nor can it do much in the near term to make the domestic US memory market more competitive. As a result, DRAM today looks a lot like an OPEC for memory — a tight-knit cartel with American tech companies such as Apple at their mercy.

In recent years, the largest change has been the growing share of the Chinese companies. But import restrictions on China exacerbate this market concentration for US tech companies — if Chinese memory makers are excluded, HHI goes even higher.

There are of course legitimate national security reasons for US policymakers to insist on supply chains that are invulnerable to China. But as the famous Foxconn factories demonstrate, the iPhone supply chain has long been controlled by the People’s Republic. It will be some time before the US can make its own iPhones.

Micron and the Korean companies clearly hold tremendous pricing power at the moment. The market might get slightly looser if the Trump administration relaxes import restrictions on DRAM as they relaxed export restrictions on the Nvidia H100. But even if its shares have dipped lately, don’t expect Micron to go anywhere.