Middle Eastern sovereign AI funds may soon enter the memory procurement market, with a single order large enough to instantly upend the supply-demand balance.
According to the latest report from Meritz Securities in Korea, countries such as Saudi Arabia have already initiated medium- to long-term procurement negotiations with Samsung and SK hynix. The report also notes that the contract price increase for server DRAM in the third quarter will exceed expectations. Additionally, the launch of Kimi K3 is interpreted by the firm as a positive signal for AI hardware demand rather than a disruptive threat.
Middle Eastern sovereign AI funds have officially entered the memory procurement market, reshaping supply-demand dynamics and accelerating price increases. Meanwhile, the recent release of the Kimi K3 model—widely discussed in AI circles—has been interpreted by Korean brokerages as a medium-term positive for AI hardware, rather than the demand shock feared by the market.
According to a research report published by Meritz Securities on July 19, Middle Eastern sovereign AI investment entities, including those from Saudi Arabia, have recently begun engaging with$Samsung Electronics (005930.KR)$、$SK Hynix (000660.KR)$major Korean memory manufacturers to discuss medium- to long-term procurement plans, inquiring about their supply capacity and volume over the coming years. The report forecasts that the quarter-over-quarter increase in server DRAM contract prices in Q3 2026 will surpass the market’s previous expectation of approximately 15%. Suppliers that offered relatively lenient pricing in Q2 are expected to implement particularly pronounced price hikes in the second half of the year.
Meanwhile, the spot market has already shown clear signs of price surges. Since mid-July, the spot price for 64GB DDR5 server DRAM has risen sharply, with recent quotations ranging from USD 3,100 to USD 3,400—approximately 146% higher than the contract price of around USD 1,380 at the end of June. High-end products with a bus speed of 6,400 Mbps have seen especially pronounced gains.
Amid these dramatic shifts in the memory market’s supply-demand landscape, Meritz Securities’ investment strategy team, in another report released the same day, characterized the launch of the Kimi K3 model by Moonshot AI as a medium-term catalyst for AI hardware demand. The firm recommended investors buy semiconductor and memory component companies while reducing holdings in large-cap tech stocks such as Google and Microsoft. Meritz argued that the market’s view—that 'Kimi K3 would reduce AI hardware demand'—rests on two key misjudgments, and that K3 will, in fact, generate incremental demand for hardware vendors over the medium term.
Entry of Middle Eastern Sovereign AI Capital: A Single Buyer Can Tilt the Supply-Demand Balance
Analysts at Meritz Securities noted in their report that the core rationale behind Middle Eastern sovereign AI investments lies in national security and data sovereignty—countries aim to build data centers within their own territories and operate AI systems using their native languages and domestic data. As these investment plans become increasingly concrete, supply tightness anticipated for 2027 is now emerging earlier than expected.
The current structural shift in the memory market means that the entry of a single large buyer can exert substantial price impact. AI data center procurement is highly concentrated; the arrival of just one new buyer can instantly shift the demand fulfillment rate—the ratio of actual supply to required demand—by 3 to 5 percentage points. Even minor fluctuations in this metric can trigger sharp price spikes. Today, large customers account for over 70% of memory manufacturers’ revenues, concentrating pricing power among a handful of buyers.
This stands in stark contrast to the fragmented procurement structure of the smartphone and PC eras, when numerous manufacturers placed small-volume orders, and individual customer order fluctuations had limited impact on the overall market. In today’s AI data center landscape, centralized procurement means that every new sovereign-scale buyer entering the market could become a decisive variable in pricing.
Supply Shortages Have Spread to End Products: Availability, Not Price, Is Now the Core Constraint
Meritz Securities noted that the impact of memory shortages has already spread from data centers to consumer electronics endpoints.$Apple (AAPL.US)$Both Chinese smartphone makers have encountered production shortfalls in Q4 due to insufficient memory procurement ahead of the peak season in the second half of the year.
He emphasized that the fundamental issue currently facing the market is one of "physical allocation" rather than merely a "price" issue—"In short, there simply isn't enough supply, and all parties are scrambling to source goods." Regarding cyclical concerns circulating in the market, he believes this assessment is somewhat misguided; the current tightness stems from structural supply shortages, not short-term demand fluctuations.
The report also noted that some suppliers adopted relatively flexible low-price strategies in the second quarter to accommodate customer needs, implying these manufacturers have greater room to raise contract prices in the third and fourth quarters, with pricing elasticity becoming more pronounced.
The launch of Kimi K3 is interpreted as positive for AI hardware demand, rather than a disruptive shock.
On the same day, analysts from Meritz Securities’ investment strategy team released a report characterizing the launch of the Kimi K3 model by Moonshot AI as a medium-term positive catalyst for AI hardware demand. They recommended investors buy shares in semiconductor and memory companies that supply AI components, while reducing holdings in large-cap tech stocks such as Google and Microsoft.
Meritz clearly distinguished this event from last year’s DeepSeek disruption. The core narrative of DeepSeek was achieving AI training at an extremely low cost of USD 6 million, whereas Kimi K3 did not disclose its training costs or GPU usage. Moreover, the official statement explicitly noted that running K3 requires a large-scale cluster composed of at least 64 high-performance chips—a model that cannot be supported by low-cost hardware. In terms of operational cost, according to data from AI performance evaluation firm Artificial Analysis, K3’s per-task processing cost stands at USD 0.95, placing it in the same range as GPT-5.6 Sol (USD 1.04) and Claude Fable 5 (USD 2.75), but an order of magnitude higher than DeepSeek V4 Pro (USD 0.04).
The analysis further pointed out that the logic behind claims that "K3 will reduce AI hardware demand" rests on two misjudgments: First, AI infrastructure investment is not solely driven by large technology firms—for instance, data on NVIDIA’s data center chip sales show that the share held by major cloud providers is declining, while the proportion accounted for by smaller AI cloud service providers and ordinary enterprises continues to rise. Second, inference services for Chinese AI models are in fact largely run on U.S.-made hardware; inference service providers such as Fireworks AI and Together AI both rely on American chips to deliver services to enterprise clients.
He concluded, "Regardless of whether models are closed-source or open-source, all participants in the AI ecosystem ultimately depend on hardware vendors, who will benefit from increased medium-term demand following the launch of K3."
Looking to select stocks or analyze your holdings? Want to understand the opportunities and risks in your portfolio? For any investment-related questions,just ask Futubull AI!
Editor/melody