DruckFin

Samsung Electronics Says Memory Shortage Will Persist Through 2028 as Multiyear Contracts Lock Up 60-70% of Capacity

Q2 2026 earnings call, July 29, 2026

Samsung Electronics delivered another record quarter, but the more consequential news for investors came from management's unusually specific multiyear framing of the memory cycle. The company now believes the DRAM and NAND supply shortage will not meaningfully ease until at least 2028, and it is restructuring its memory business around long-term, prepaid supply agreements that could fundamentally change the volatility profile of the stock.

A Structurally Tighter Memory Market Through 2028

Jaejune Kim, head of Memory Global Sales and Marketing, laid out the clearest articulation yet of Samsung's supply-demand thesis. He said fab lead times exceeding three years mean "it will be unlikely to see any significant increase in incremental supply through 2028," and that unmet 2026 demand is carrying into 2027, when constraints are expected to intensify further. Beyond 2029, visibility drops off, but Kim said AI token growth is expected to drive exponential demand over the medium term regardless.

The dynamic he described is unusual for the industry: AI frontier model developers unable to secure enough hyperscaler cloud capacity are turning to neo-cloud providers, and when those providers also can't get enough memory, the model developers are now approaching Samsung directly for multiyear supply agreements. That has effectively created a third layer of demand beyond the traditional hyperscaler-OEM relationship, and it is why Samsung is now willing to sign five-year rolling contracts with substantial upfront prepayments.

Multiyear Contracts Now Covering 60-70% of Capacity

This is arguably the single most important disclosure on the call. Samsung has already finalized multiyear agreements with the top five global data center customers and is in final-stage talks with five more. Originally the company planned to cap long-term contracts at 60% to 70% of capacity to preserve flexibility, but Kim said that ceiling is "increasingly becoming constrained" given how many customers, including existing contract holders, are now asking for more.

The contracts run on five-year terms with annual renegotiation options, structured to roll forward indefinitely. Samsung has built in minimum price floors for mainstream products to protect against future investment risk, and has already collected roughly a quarter of the total prepayments due under these deals, though the company declined to disclose dollar figures citing NDA restrictions. Kim was direct about the strategic intent: "we hope to significantly enhance the stability and visibility of our business" and shift memory away from its historical boom-bust cycle toward a "backlog-driven" model. If this holds, it would be a genuine structural change to how investors should value Samsung's memory earnings power, reducing cyclicality in exchange for locked-in margins.

HBM4 Ramp and a More Balanced Product Mix

Samsung shipped industry-first HBM4E samples to major customers in the quarter and expects HBM4 revenue to grow more than threefold quarter-on-quarter in the third quarter, with HBM4 comprising over 60% of total HBM revenue mix in the second half. Management reiterated it intends to keep HBM market share roughly in line with overall DRAM share, rather than chasing HBM's higher margins at the expense of conventional DRAM, because server DRAM and general-purpose computing demand are growing just as fast. This balanced-mix strategy, held consistently across two quarters now, suggests Samsung is prioritizing capacity discipline over near-term margin maximization in HBM specifically.

Record Quarter, But Component Costs Squeezing Devices

Revenue reached KRW 171.5 trillion, up 28% quarter-on-quarter, with operating profit up 56% to KRW 89.5 trillion and operating margin expanding to 52% from 43%. DRAM ASPs rose in the mid-40% range sequentially and NAND ASPs jumped roughly 60-70%, underscoring just how tight the memory market has become. CFO Soon-Cheol Park noted a favorable currency swing added roughly KRW 3.1 trillion to operating profit from the weaker won.

But the flip side of tight memory is that Samsung's own device businesses are getting squeezed by the same component inflation. The Mobile Experience division swung to a combined operating loss of KRW 0.7 trillion with Networks, and Daniel Araujo acknowledged rising mobile memory costs will continue pressuring profitability through the second half. Visual Display and System LSI both cited similar component cost headwinds. This is a notable internal tension: Samsung's memory division is the primary beneficiary of the shortage it is helping create for its own downstream hardware units.

Foundry Turnaround Approaching, 2-Nanometer Orders Doubling

Foundry appears to be nearing an inflection point. Utilization at advanced nodes of 8-nanometer and below has reached full capacity, and Sukchae Kang said 2-nanometer order wins are expected to more than double year-over-year in 2026, with new project commitments from major cloud service providers and AI/HPC customers, alongside ongoing talks with Broadcom. Advanced nodes are projected to exceed 50% of foundry revenue, and AI/HPC applications are expected to jump from high-teens percentage of the mix in 2025 to over 30% in 2026. Kang said a profitability turnaround "is possible in the near term," though he stopped short of giving a specific quarter given the order-driven nature of the business. Taylor Fab 1 in the U.S. remains on track for 2026 startup, and construction on Taylor Fab 2 will begin by year-end, targeting 2030 production, with the company also fielding rising customer interest in 1.4-nanometer capacity.

Shareholder Return Policy Still Unresolved, ADR Not Currently Under Review

Investors have been waiting for clarity on the back half of Samsung's 2024-2026 shareholder return policy, and the company again deferred specifics, saying the board is still finalizing details including a potential special dividend and is simultaneously working on the next multiyear policy. Park acknowledged that customer prepayments under long-term supply agreements and share buybacks tied to employee compensation could both affect the free cash flow base used to calculate returns, adding a layer of complexity investors should watch for when the next policy is unveiled.

On the ADR question that has gained traction following a competitor's U.S. listing, Park was unusually candid: "we're not currently reviewing an ADR issuance," citing concerns about liquidity fragmentation and potential valuation gaps between ADR and domestic shares. He left the door open only as a longer-term option, not an active initiative.

Robotics Build-Out and HARMAN's ADAS Push

Samsung disclosed meaningful new detail on its robotics ambitions, having consolidated capabilities into a new RX business office reporting directly to the CEO. The company is building a robotics pilot line and data factory in Gumi, and has hired Dongkun Lee, formerly of Hyundai's Boston Dynamics strategy team, to lead robotics strategy, along with two academic recruits specializing in autonomous robotics and robotic hands. The stated focus is B2B manufacturing and logistics applications first, with humanoid robots and B2C expansion as longer-term goals, alongside potential M&A and startup investments.

Separately, HARMAN's acquisitions of ZF's ADAS business and Sound United mark a deliberate push beyond digital cockpits into autonomous driving hardware and premium audio branding, positioning HARMAN, in Park's words, to become "a leading global player in auto electronics" through centralized vehicle controller integration.

Disclaimer: This article is for informational purposes only and does not constitute investment advice or a recommendation to buy, sell, or hold any security. Our analysts provide detailed coverage of corporate events but can make mistakes, always conduct your own due diligence. The views and opinions expressed do not necessarily reflect those of DruckFin. We have not independently verified all information used herein, and it may contain errors or omissions. Before making any investment decision, consult a qualified financial advisor. DruckFin and its affiliates disclaim any liability for any losses arising from reliance on this content. For full terms, see our Terms of Use.