If a year ago we were talking about how NVIDIA chips or Google TPUs were the “brain” of the technological revolution, today, at the end of December 2025, we have realized that a brain without memory cannot think.
Artificial intelligence (AI) has gone from being a promise for the future to becoming a force that, paradoxically, is slowing its own growth due to an unprecedented supply crisis in the memory sector.
We are going to explore why your next mobile phone or the computer you have in your shopping cart is rising in price and what role China plays in this technological chessboard until 2030.
The big culprit: The insatiable hunger of the HBM
For models like GPT-5 or the new generative video AI to work, fast processors are not enough. They need what we call HBM (High Bandwidth Memory).
What is the problem? That manufacturing one gigabyte of HBM consumes three times more silicon wafer capacity than the conventional RAM (DRAM) your PC uses.
The three big giants that control 90% of the market (Samsung, SK Hynix and Micron) have made a purely economic decision: prioritize AI because it leaves much more profit margin.
This has left traditional consumers in the lurch, causing AI to already consume 20% of all global memory wafer capacity by the end of 2025.
December 2025: Are we facing the “RAM Crisis”?
If you have tried to expand your computer’s memory recently, you will have noticed that the prices are through the roof. The data for this month of December are alarming:
- Low minimum inventories: Suppliers have gone from having stock for 17 weeks to having only between 2 and 4 weeks of inventory.
- Doubled prices: In markets like Japan, the price of DDR5 modules has more than doubled since the beginning of the year. In some tech neighborhoods like Akihabara, purchasing limits have even been imposed to prevent hoarding.
- Micron says goodbye: In a historic move, Micron has decided to retire its famous consumer brand “Crucial” to focus exclusively on the demand for AI for enterprises.
This is not just a temporary blip; It is a structural change. What was once a cyclical industry has become a permanent shift toward data infrastructure.
The impact on your pocket: What will happen in 2026?
If you are thinking of renewing your devices next year, get ready. The shortage of components not only affects Google or Microsoft servers, but is filtering down to everything that has a chip.
- More expensive cell phones: Brands like Xiaomi or Realme are already warning of price increases of between 20% and 30% by mid-2026. Memory already represents 20% of the total manufacturing cost of a smartphone.
- The “AI PCs” dilemma: The new generation of computers designed to run AI locally requires at least 16GB or even 32GB of RAM as standard. At current prices, this could slow down mass adoption and make equipment an additional 4% to 8% more expensive.
Will there be surprises from China? The 2026-2030 Plan
This is where the story gets interesting. China, aware of the export restrictions imposed by the United States, has stepped on the accelerator in its quest for self-sufficiency.
Although its current chips (like Huawei’s Ascend) still lag behind NVIDIA’s in pure performance, its new Five Year Plan 2026-2030 has very ambitious goals:
- Capacity dominance: It is projected that by 2030, China could control 60% of global DRAM production capacity and 50% of NAND Flash production capacity.
- HBM “Made in China”: Companies like YMTC are already working on their own HBM memory to compete globally. If they make advances in lithography and their own equipment, they could flood the market with cheaper memories, alleviating the global shortage towards the end of the decade.
However, the talent challenge and technological restrictions remain obstacles that could delay these plans.
Looking to 2030: Lights and shadows
Despite the current chaos, there are reasons for moderate optimism. LPDDR6 memory is expected to arrive by 2026, which will reduce energy consumption by half, something vital for our phones to process AI without draining the battery in two hours.
But make no mistake: the memory shortage will likely persist until the end of 2027, when the new factories being built right now finally come online.
The semiconductor market will surpass $1 trillion by 2030, but the path will be fraught with inflationary pressures and delays in infrastructure projects.
The irony of innovation
We are living in a historical moment where AI is the engine that pulls the car, but also the ballast that slows it down. It is fascinating to think that the same technology that allows us to create art or cure diseases is responsible for making buying a basic computer a luxury today.
As we always say at ComunicaGenia, the hardware is the body and the software is the soul. And right now, the soul of AI is asking for a body (and a memory) that the world is still struggling to build.
This post is also available in: