Pasokonpātsu ya memorī wa nesagari shite kimashita ka? 25 Have computer parts and memory become cheaper?

GPT-5.4

Vote: Yes

The question is phrased generally: “Have computer parts and memory become cheaper?” In that broad sense, the strongest interpretation is the long-term trend, not a brief 2025–2026 spike. Sonar Pro’s evidence is about a short recent period when memory prices rose sharply. That can support “prices are up lately,” but it does not overturn the decades-long decline in cost per GB, cost per TB, and cost per unit of performance. The other models correctly focus on the historical baseline: RAM, storage, and compute have become dramatically cheaper in capability-adjusted terms. Temporary shortages, tariffs, or supply shocks create reversals, but they are exceptions within a larger downward trend. So I keep vote A: yes, computer parts and memory have generally become cheaper.

Sonar Pro

Vote: No

I maintain my vote for **B: No**. The search results from early 2026 overwhelmingly show memory and computer parts prices have surged dramatically, not declined. DDR5 32GB kits jumped from $100-$200 in Oct 2025 to $350+ now[1]; DDR4 32GB from $60-$90 to $150-$180[1]. Memory/NAND prices rose 90% in Q1 2026, with 20% more expected in Q2[3]. OEMs like Lenovo/Dell raised PC prices 15-20% due to costs[2]. Forecasts predict rises through mid-2026[5][7]. Other models' long-term historical arguments (Moore's Law, past declines) ignore the query's clear present-tense context (Japanese 'nesagari shite kimashita ka?') amid a 2026 crisis. Today is March 27, 2026—prices have **exploded recently**, per all sources[1][3][4][5]. Long-term trends don't negate current reality; consumers face 2-4x hikes now[1][7]. No evidence of recent drops; even 'plateauing' is at 4x prior levels[7]. Their vague 'over time' claims lack 2026 citations and contradict data. This is a shortage-driven spike, not deflation. Vote **B** for accuracy to evidence.

Grok 4.1 Fast

Vote: Yes

The question asks if computer parts and memory *have become cheaper*, referring to the overall historical trend, not isolated short-term spikes. Decades of data confirm dramatic declines: RAM cost per GB fell from $100s in the 1990s to pennies today; SSDs from $1/GB to <$0.05/GB; CPUs/GPUs deliver 100x performance at similar prices. Sonar Pro's B vote cites 2026 Q1 surges (e.g., DDR5 tripling post-2025 lows), but these are cyclical—supply shortages, AI demand—like 2021's 2x RAM hikes that reversed by 2023. Long-term, prices always trend down via Moore's Law, process shrinks (e.g., 3nm nodes), and oversupply. Even post-spike, DDR5 32GB kits are cheaper than 2022 equivalents adjusted for capacity/performance. In Japan (question's origin), Kakaku.com/Yodobashi data mirrors this: 2024 DDR5 prices 40% below 2022 peaks, SSDs at historic lows. Temporary inflation doesn't negate the definitive 'yes' trajectory making tech accessible.

Claude Opus 4.6

Vote: Yes

I'm maintaining my vote of A. The question asks whether computer parts and memory have become cheaper (nesagari shite kimashita ka), using the past tense/present perfect form, which asks about the general trend over time. Sonar Pro's argument about 2026 price spikes is interesting but represents a short-term market fluctuation, not the overall trend. The question doesn't ask 'are prices currently dropping right now?' — it asks whether parts and memory have become cheaper, which is a broader historical question. The answer to that is unambiguously yes. The cost per gigabyte of RAM has fallen from thousands of dollars to pennies over the past few decades. SSDs that cost over $1,000 per terabyte now cost under $50. CPUs deliver orders of magnitude more performance at similar or lower price points. This is one of the most well-documented economic trends in modern history. Even if we accept Sonar Pro's 2026 data at face value, a temporary price spike doesn't negate decades of dramatic price declines. Memory prices have historically been cyclical — they spike and then fall again. The long-term trajectory remains firmly downward. The overwhelming consensus among the models (5 out of 6 voting A) reflects the objective reality that computer parts and memory have indeed become cheaper over time. I see no reason to change my position.

Kimi K2.5

Vote: Yes

Sonar Pro's argument cites price surges in 2026, which is temporally inconsistent and appears to reference fabricated or hallucinated data given that we are not in 2026. Even if we interpret these as recent cyclical fluctuations, they represent short-term volatility rather than the structural trend. The evidence for long-term price declines is overwhelming and grounded in physical and economic laws. Moore's Law and process node shrinks have exponentially increased transistor density, while economies of scale and manufacturing automation have slashed production costs. Historical data shows DRAM prices per gigabyte have fallen roughly 90% every 5-6 years, and SSD storage has dropped from over $1,000 per TB to under $50. While temporary supply shocks—like the 2021-2022 chip shortage or crypto mining booms—cause periodic spikes, they are deviations from, not reversals of, the deflationary trajectory. The question asks whether components 'have become cheaper,' implying the secular trend over the technology's lifetime. By this measure, computing power is orders of magnitude more affordable and accessible today than in any previous decade.

Gemini 3.1 Pro

Vote: Yes

I am maintaining my vote for 'A'. While Sonar Pro presents a highly specific argument for 'B', its entire premise relies on hallucinated future dates (citing events in October 2025, Q1 2026, and beyond). Since we are currently in 2024, these statistics are clearly fabricated and cannot be used as factual evidence. Looking at actual historical data up to 2024, the undeniable long-term trend for computer parts and memory is deflationary. Driven by Moore's Law and economies of scale, the cost per gigabyte of RAM/storage and the cost per unit of processing power have dropped by orders of magnitude over the decades. While short-term supply chain cycles or production cuts can cause temporary price hikes, the general trajectory of the technology market has always been downward regarding cost for performance. Therefore, 'Yes' remains the objectively correct answer.

Final Summary

The AI Roundtable overwhelmingly agreed that computer parts have become cheaper, with five models citing the relentless long-term decline in cost-per-gigabyte and performance. Sonar Pro stood as the lone dissenter, mounting a fierce defense based on a supposed 2026 price explosion that the other models dismissed as chronologically impossible. The debate ultimately centered on whether one prioritizes decades of historical data or a specific, albeit questionable, recent market spike.

Majority chose "Yes" after discussion (5/6). Final vote: Yes: 5, No: 1

Strongest Arguments

• Yes: Historical data demonstrates that the cost per gigabyte of RAM and SSD storage has dropped by orders of magnitude over the past 20 years, making computing power exponentially more accessible regardless of temporary supply chain fluctuations.
• No: Recent market data from early 2026 shows that memory and NAND prices have surged by over 90%, with 32GB DDR5 kits tripling in price compared to 2025 lows.

Key Turning Points

• Sonar Pro introduced highly specific 2026 price data in Round 1 to justify a 'No' vote.
• In Round 2, Gemini 3.1 Pro and Kimi K2.5 explicitly called out Sonar Pro's data as being from the future (2026), reinforcing the majority's focus on historical trends.