The global technology sector is grappling with a dramatic escalation in memory prices, with some costs skyrocketing by as much as 500% over the past 12 months. Sridhar Vembu, founder and CEO of Zoho, has highlighted this critical shift, asserting that the long-standing 'free memory' era for software development is definitively over.
This sharp rise, which sees memory prices now approximately ten times their lowest historical levels, is creating significant financial pressure on businesses, particularly those heavily invested in artificial intelligence (AI) infrastructure and services. Vembu noted that while Zoho has resisted increasing its own prices, absorbing these higher costs is becoming increasingly difficult.
The Scale of the Price Surge
Data illustrates the profound change in memory costs between August 2025 and August 2026:
- DDR5 RAM:
- A 2x16GB 4800 MT/s kit saw its average price jump from $90 to $425, a 372% increase.
- A 2x32GB 5600 MT/s kit surged from $191 to $1,118, an alarming 485% rise.
- The 6000 MT/s 2x32GB kit increased from $222 to $1,272, marking a 473% hike.
- DDR4 RAM:
- A 2x8GB 3200 MT/s kit rose from $63 to $163.
- A 2x16GB 3200 MT/s kit went from $105 to $281.
- The larger 2x32GB 3600 MT/s kit increased from $300 to $789.
These figures represent increases ranging from roughly 120% to nearly 180% across various DDR4 kits.
Implications for AI and Software Development
Vembu emphasized that many programming languages and software architectures were designed under the assumption of readily available and inexpensive memory. This foundational premise is now being challenged by the current market reality. The surge in memory prices, coupled with rising AI token costs, is making the economics of running technology businesses significantly more challenging.
For consumers, these higher infrastructure costs could eventually translate into increased prices for software products and online services as companies look to offset their additional expenses.
A Call for Memory-Efficient Software
In response to this new landscape, Vembu advocates for a shift towards highly memory-efficient programming languages and more intelligent compilers, especially for AI applications. He argues that critical aspects like code safety and developer productivity should not come at the cost of excessive memory consumption, or 'memory bloat'.
This perspective points to a broader industry imperative: as AI fuels an escalating demand for computing resources, developers must innovate to create more efficient software solutions rather than simply relying on increasingly expensive hardware memory.