Project Chintan

Sridhar Vembu flags memory price surge raising business costs

Zoho co-founder Sridhar Vembu warned that rising memory and AI token prices are making operations harder, referencing a report that memory costs have surged dramatically. He argues that memory-efficient programming and smarter compilers will be needed as the era of “free” memory ends.

· 2 min read

Key takeaways

  • Sridhar Vembu linked rising memory and AI token costs to increasing pressure on business operations.
  • Tom’s Hardware reported memory price surges, with specific 64GB DDR5-5600 kits rising from $191 to $1,118 within a year.
  • A 128GB DDR5-6400 kit rose to $3,399 from a low of $329, illustrating dramatic memory price inflation.
  • DDR4 prices also increased significantly year over year, by 120% to 177%.
  • Vembu argued for more memory‑efficient programming languages and smarter compilers as memory becomes costlier.

What Happened

Zoho co-founder Sridhar Vembu addressed the spike in computer memory prices, linking it with higher AI token costs and saying these trends are making it harder for businesses to operate. He commented in a post on X, responding to a Tom’s Hardware report that documented sharp price increases in memory modules over the past year.

The Tom’s Hardware piece cited a 12‑month period in which memory prices rose substantially. It noted that a 64GB DDR5-5600 kit averaged about $191 in August 2025 and was around $1,118 in August 2026, roughly a 485% increase. It also highlighted an example where a 128GB DDR5-6400 kit was listed at $3,399, compared with a previously tracked low of $329. The report additionally showed year‑on‑year rises for DDR4 memory ranging from about 120% to 177% across the tracked kits.

Vembu indicated that companies have tried to avoid passing these costs to customers, but maintaining price levels is becoming increasingly difficult. He remarked that the era of memory being effectively free has ended and expanded on a broader point about software development in the context of higher memory costs.

In his posts, he argued that programming languages and software tools will need to become more memory‑efficient and that smarter compilers should be developed, even for AI applications. He also stated that the safety of code and productivity in writing code should not come at the expense of memory efficiency, noting this as part of his research focus.

Why It Matters

The reported price dynamics suggest that both hardware costs and the cost of AI‑related processing could influence operating expenses for technology‑driven businesses. If memory remains expensive, firms may prioritize memory‑efficient software design and compiler innovations, potentially affecting how software is developed and deployed at scale. Vembu’s emphasis on efficiency links hardware cost pressures to broader software engineering practices and tooling choices in an era of AI expansion.

Background

The discussion centers on a Tom’s Hardware analysis of memory pricing over the prior year, which documented sizable increases in DDR5 and DDR4 memory prices. The contemporaneous commentary from Vembu connects these hardware market movements to business costs and software development methodologies, framing it as a broader shift in the economics of memory usage in computing.

Key Facts

  • Sridhar Vembu commented on memory and AI token price increases in response to a Tom’s Hardware report.
  • The Tom’s Hardware article reported memory price increases up to about 500% over 12 months.
  • A 64GB DDR5-5600 kit was about $191 in August 2025 and about $1,118 in August 2026 (roughly 485% rise).
  • A 128GB DDR5-6400 kit was listed at $3,399, compared with a prior low of $329.
  • Vembu argued that the era of memory being effectively free has ended and called for more memory‑efficient languages and smarter compilers, tying this to AI usage and code safety/productivity.

Sources reviewed

Project Chintan independently synthesized and analyzed information cross-checked across the sources listed above.

Related stories