SanDisk Earnings Call: Secures $42 Billion Long-Term Orders, High Margins Remain Sustainable – “At this industry turning point, I can hardly think of anyone who is a loser!”

On April 30 local time, Sandisk held its third-quarter earnings call for the fiscal year 2026. Against the backdrop of a more than 6% drop in after-hours stock price, CEO David Goeckeler and CFO Luis Visoso spent considerable time explaining to analysts that the company is undergoing structural changes, rather than merely benefiting from a cyclical upturn.

Goeckeler stated that five multi-year cooperation agreements have been signed so far, with negotiations ongoing with several other customers. 'We are transforming this business into a recurring revenue model.'

Sandisk remains highly confident in storage. When responding to analyst Jake Wilhelm’s question about long-term agreements, CEO Goeckeler went so far as to say, 'I can hardly think of anyone who would be a loser.'

Customers are very satisfied with these agreements, and we ourselves are pleased as well. As I said, I believe everyone is a winner. I can hardly think of anyone who would lose in this model. So far, every agreement we’ve signed has left customers thrilled at the final signing.

We are not reinventing the wheel; we are simply introducing a recurring revenue model that has proven effective in other industries into our brand.

In summary, we are at a pivotal moment in this business, and we continue to execute with confidence. NAND remains a foundational technology underpinning the world's leading semiconductor storage solutions, driving the most significant waves of technological transformation — from PCs, mobile internet, and cloud computing, to today’s artificial intelligence.

42 billion in long-term orders, 11 billion in guarantees: Management strongly affirms 'it’s not just verbal commitments.'

The central topic of this conference call was Sandisk’s push for a 'New Business Model' (NBMs) — namely, multi-year supply cooperation agreements with customers. The focus was also on whether these contracts were binding.

CEO Goeckeler proactively disclosed progress at the outset: The company has signed multi-year supply agreements with five customers, three of which were finalized in the third quarter and the other two completed at the start of the fourth quarter. He noted that discussions with several other customers remain active.

CFO Luis Visoso provided specific figures: The three contracts signed within the quarter generated approximately 42 billion in minimum contractual revenue (RPO, Remaining Performance Obligations), which will be disclosed in the company’s 10-Q filing. The total financial guarantees across the five agreements exceed 11 billion, with 400 million in prepayments already reflected in the third-quarter balance sheet, while the remainder is secured through financial instruments managed by third-party financial institutions.

Visoso explained the operational logic of the guarantee mechanism:

If the client fails to fulfill their procurement obligations on a quarterly basis, these financial commitments will immediately be transferred to us as compensation.

In response to analysts’ skepticism about whether the contract is enforceable, Goeckeler directly replied: 'Some have told me that these agreements would not hold or carry no binding effect. I can tell you, it’s quite the opposite. Our clients are putting up tens of billions of dollars in collateral through various financial instruments, which will remain valid throughout the contract period.'

He also revealed that shortly after signing the contracts, clients typically begin discussing how to increase their procurement volumes rather than reduce them.

Can profit margins be sustained? Management responded affirmatively but refrained from providing a target range for now.

This was the most frequently asked question during the conference call. This quarter, SanDisk's non-GAAP gross margin reached 78.4%, significantly surpassing earlier guidance (65%-67%). This figure has raised questions in the market: Is it sustainable?

To this, Goeckeler expressed a clear stance:

We believe our gross margin is sustainable.

However, when pressed by Ben Reitzes, an analyst at Melius Research, for a specific target range, Goeckeler indicated that the timing was not yet right:

I don’t think we’re at a point where we can discuss that yet. Once we make more progress, we’ll consolidate all of this into a new business model and present it to everyone. But we are extremely proud of our technology… Frankly, the market has always recognized the value of our technology. It’s just that in the past, others captured that value instead of the producers. Now, we are distributing this value more equitably.

He directly addressed market concerns about SanDisk's valuation — implying that the stock price reflects a return to gross margin in the 40% range — and stated:

We are extremely focused on removing cyclicality from this business. Cyclicality is corrosive — it’s corrosive to how we allocate CapEx, and it’s corrosive to our customers getting enough product to drive their fantastic businesses.

Data center surges 233%: TLC dominates this quarter, QLC enters next quarter

The data center was the most prominent growth engine this quarter. Goeckeler stated that data center revenue grew 233% quarter-over-quarter to reach $1.467 billion, accounting for approximately 25% of total revenue, with expectations for this proportion to continue rising. He attributed the growth to structural demand for high-performance NAND driven by AI inference architectures:

Inference optimization (such as KV caching) and workloads like RAG require large amounts of high-performance, low-latency flash memory to deliver real-time responsiveness and quality user experience. These workloads significantly expand the volume of data that needs to be stored on low-latency flash memory — far exceeding the model itself, as systems must retain context, intermediate data, and large external datasets.

Currently, nearly all data center revenue comes from the TLC product line. Next quarter, Sandisk will begin shipping its QLC Stargate solutions, adding another layer to data center revenue growth. Goeckeler commented on this: 'This product has completed over a year of certification with several major customers, and we have great confidence in it.'

Debt eliminated, $6 billion buyback, and disciplined capital expenditure

During the earnings call, the company announced that the board had approved a $6 billion stock repurchase program, effective immediately, with no expiration date.

Visoso stated that the company had repaid the remaining $650 million term loan (TLB), achieving its net cash target, with a cash and equivalents balance of $3.735 billion at the end of the quarter.

He described the capital allocation priorities as a three-step process: investing in the business, achieving net cash, and returning value to shareholders — 'We are now at step three.'

CFO Visoso added that the absolute amount of capital expenditures would slightly increase over the next few quarters, but the overall philosophy remains unchanged, maintaining a mid-to-high single-digit annualized Bit capacity growth target.

In other words, large-scale capacity expansion is unnecessary, as node migration drives growth. Goeckeler stated that the company achieves high-teens percentage shipment growth through BiCS node migration (rather than new capacity construction), while the ratio of capital expenditures to revenue continues to decline.

Full Summary of Sandisk's Q3 2026 Earnings Call

Meeting Date: April 30, 2026 Company Name: Sandisk Meeting Description: Q3 2026 Earnings Call Source: Sandisk Version: Final

Presentation Segment
Moderator: Good afternoon, and welcome to Sandisk's Q3 2026 Earnings Call. All participants will be in listen-only mode. (Operator instructions) After today’s presentation, there will be an opportunity for questions. (Operator instructions) Please note that this event is being recorded.

Now, I will hand over the call to Ivan Donaldson, Vice President of Investor Relations. Please proceed.

Ivan Donaldson, Vice President of Investor Relations: Before we begin, please note that today’s discussion will include forward-looking statements based on management’s current assumptions and expectations. These statements are subject to various risks and uncertainties. These forward-looking statements include expectations regarding our technology and product portfolio, business plans and performance, capital allocation priorities, market trends and opportunities, and future financial results. We do not undertake any obligation to update these statements. For more information on risks and uncertainties that could cause actual results to differ materially from expectations, please refer to our Annual Report on Form 10-K filed with the SEC and other documents.

Today, we will also refer to non-GAAP financial measures. A reconciliation table between non-GAAP and comparable GAAP financial metrics is included in the written materials available on the investor relations section of our website.

Next, I will hand over the call to David.

David Goeckeler, Chairman and Chief Executive Officer:

Thank you, Ivan. Good afternoon, and thank you all for joining Sandisk's third-quarter earnings call for the fiscal year 2026. Once again, we delivered outstanding results this quarter, with all key metrics performing exceptionally well, fully demonstrating the competitive strength of the Sandisk brand. Before discussing the performance across our end markets, I would like to provide an update on a strategic priority we previously highlighted.

Last quarter, we were in discussions with customers regarding multi-year supply cooperation agreements, which we refer to as the 'New Business Model' (NBM). I am pleased to announce that these negotiations have successfully advanced, resulting in the signing of five multi-year agreements. These agreements aim to secure committed supply for customers while locking in guaranteed revenue for Sandisk, backed by clear financial assurances from the clients. These partnerships support Sandisk in achieving sustained, structurally higher profitability and a more predictable, less cyclical business model. We believe this marks a fundamental evolution in our business—centered on deeper customer alignment, enhanced visibility, and long-term value creation.

These new business models reflect the strategic value of our world-class NAND technology, which is built upon decades of continuous innovation. Our cumulative capital expenditures and intellectual property investments in R&D and manufacturing total hundreds of billions of dollars, establishing a robust foundation for this new business model—spanning full-stack vertical integration from front-end manufacturing, chip and system-level design, to final back-end packaging and testing.

The extension of our joint venture with Kioxia and the DRAM supply agreement reached following our investment in Nanya Technology further enhance our supply chain resilience. This leverage is driving deeper collaborations with customers and fostering long-term dialogues with partners who prioritize technical performance and long-term supply assurance.

With increased customer engagement and flexible configuration capabilities across our product portfolio, we are better positioned to optimize our end-market structure. Collectively, these transformations have driven a step-change improvement in our sustainable gross margin, free cash flow generation, and overall profitability—all within a market we expect to sustain double-digit growth in the foreseeable future.

The data center segment serves as a prime example of this strategy in action—revenue from this business grew 233% quarter-over-quarter. Behind this milestone lies years of meticulous preparation and our proactive transition toward the most strategically significant and fastest-growing end markets. Despite substantial progress, the growth potential driven by the fundamental shift in AI infrastructure demand remains vast.

We are witnessing extraordinary growth—not only in the expansion of model sizes but also in the resulting increase in token generation, prolonged model runtimes and complexity, and the growing importance of contextual relevance. As AI models scale from billions to trillions of parameters, deployment scenarios evolve from simple inference to deep inference, and increasingly autonomous intelligent agent systems emerge, NAND has become a critical component of the underlying infrastructure.

Inference optimization techniques such as KV caching, along with workloads like retrieval-augmented generation (RAG), require substantial high-performance, low-latency flash storage to enable real-time responsiveness and high-quality user experiences. These workloads significantly expand the volume of data that must be stored on low-latency flash—far exceeding the models themselves—as systems must retain contextual information, intermediate data, and large external datasets.

Consequently, NAND flash is emerging as the only economically viable solution capable of delivering the capacity, performance, and efficiency required to maintain model accessibility in large-scale real-time inference. This recognition of the criticality of our technology coincides with the peak differentiation of our product offerings—anchored by our industry-recognized gold-standard BiCS8 NAND technology and complemented by a broad leading portfolio encompassing TLC and QLC products.

We firmly believe that, with our world-class product portfolio and technological leadership, we will continue to position Sandisk as the preferred partner for data center customers over the long term. We are already seeing this advantage translate into tangible performance outcomes.

In the third quarter of fiscal year 2026, our revenue was strongly supported by robust demand for our enterprise TLC-based SSD product portfolio, which is primarily used in high-performance computing workloads where speed and latency are critical.

Looking ahead to the fourth fiscal quarter, we anticipate beginning to generate revenue from our QLC Stargate solutions, adding a new layer of revenue growth. TLC and QLC each serve distinct but complementary application scenarios, demonstrating how we carefully architect our product lines with a broad AI-oriented data center portfolio to meet evolving customer needs.

In the edge computing space, both the PC and smartphone markets are showing a continued trend toward higher-end devices. These platforms are increasingly integrating local capabilities, driving up storage requirements and fueling demand for high-performance solutions. As a result, our product mix continues to shift toward higher-value configurations and customers who fully recognize the value of our technology.

Despite ongoing shifts in consumer market dynamics, our consumer business achieved strong year-over-year revenue growth across all major storage categories and regions. This performance was driven by our strong brand recognition and channel coverage, as well as our strategy of focusing on the most financially attractive demand segments.

In February this year, we launched our next-generation portable SSD portfolio, designed to support faster and more demanding workflows and AI-enabled content creation. This launch reinforced our leading position in innovation within the SSD category and garnered extensive coverage from global media, significantly enhancing our brand's external visibility.

At the same time, we have deepened global consumer engagement through new brand-led marketing campaigns, such as the 'Space to Hold More' initiative, leveraging localized global storytelling and deep connections with diverse global communities to further enhance customer loyalty.

In summary, these efforts exemplify our continued focus on end markets and our commitment to driving demand through brand awareness, product innovation, and strong market execution, while steering our portfolio toward higher-value opportunities and gradually moving away from traditional upselling models. Our broad end-market coverage remains a core differentiator, and we remain dedicated to serving various customers across these markets.

Next, I will hand over the meeting to Luis for a detailed discussion of our financial results and guidance.

Luis Visoso, Executive Vice President and Chief Financial Officer:

Thank you, David. I will first provide an update on the progress of the New Business Model (NBM) – these models are designed to offer us demand certainty while providing supply assurance to our customers.

In the third quarter, we signed three agreements, and two additional contracts have been executed in the fourth quarter to date, with ongoing active negotiations underway with multiple clients. These agreements are tailored to each client’s specific needs and collectively provide us with demand certainty and financial security aligned with our expectations for the fourth fiscal quarter guidance.

The durations of the agreements vary, with the longest extending up to five years. Collectively, the delivery commitments under these contracts increase progressively over their terms, incorporating quarterly commitments combined with both fixed and floating pricing mechanisms. The inclusion of floating pricing allows us to capture upside benefits during price increases while enabling customers to enjoy certain price discounts if prices decline over time.

As you will see in our 10-Q quarterly report, the three contracts signed this quarter guarantee a minimum contractual revenue of approximately $42 billion. We will provide updates as progress continues. Each contract includes financial guarantees that protect us in the event that customers do not fully meet their purchasing obligations. The total amount of financial guarantees under the five signed agreements exceeds $11 billion, including prepayments and other financial instruments managed by third-party financial institutions. Of this, $400 million in prepayments has been recorded on our balance sheet for the third quarter.

These five New Business Model agreements account for more than one-third of our bidding share for fiscal year 2027, and we expect this proportion to increase further as more agreements are signed in the coming months.

We anticipate that these new business models will reshape the historical cyclicality of our business, improve visibility, and deliver returns in pricing and margins commensurate with the value of our technology and investments, ultimately creating higher, more stable, and sustainable returns for shareholders.

Regarding quarterly performance: Third-quarter revenue reached $5.95 billion, representing a sequential growth of 97% and a year-over-year increase of 251%. This performance significantly exceeded our guidance range of $4.4 billion to $4.8 billion, primarily driven by a structural shift towards high-value customers and an overall improvement in pricing levels.

BiCS shipments were flat year-over-year but declined by a high single-digit percentage sequentially this quarter due to inventory accumulation efforts, mainly to support robust BiCS8 QLC demand in the fourth quarter, the ramp-up of Stargate product mass production, and preparations for recently signed New Business Model agreements. Consistent with our mid-to-high single-digit growth model, BiCS shipments have increased by 18% year-to-date in fiscal year 2026.

By end market: Sequentially, data center revenue grew 233% to $1.467 billion; edge computing increased 118% to $3.663 billion; and consumer revenue was $820 million, down 10%, consistent with our historical seasonality patterns. Our portfolio planning strategy remains focused on achieving long-term sustainable economic benefits, with diversification continuing to be our core competitive advantage. We remain committed to serving all three end markets to maximize long-term value.

In terms of profitability: Non-GAAP gross margin for the third quarter was 78.4%, a significant increase from 51.1% in the previous quarter and well above the guidance range of 65% to 67%, primarily driven by a shift towards higher-value products and an improved overall pricing environment. Non-GAAP operating expenses for the third quarter were $448 million, representing 7.5% of revenue, a substantial decrease from 13.7% in the prior quarter, reflecting continued improvements in scale efficiency and outperforming the guidance range of $450 million to $470 million.

As a result, the non-GAAP operating profit margin reached 70.9%, significantly higher than the previous quarter's 37.5%. The non-GAAP earnings per share (EPS) were $23.41, a substantial increase from $6.20 in the previous quarter, and clearly outperforming the guidance range of $12 to $14 per share.

Key adjustments between GAAP and non-GAAP include post-tax equity incentive expenses of $20 million, accounting for 0.3% of revenue; and a $46 million write-off of unamortized issuance costs due to the repayment of the remaining balance of the $6.5 billion Term Loan B (TLB). At the end of the quarter, our balance sheet reflected cash and cash equivalents of $3.735 billion.

In terms of free cash flow: This quarter, we generated $2.955 billion in adjusted free cash flow, with a margin of 49.7%. Operating cash flow amounted to $3.038 billion, partially offset by net capital expenditures of $83 million. Total capital expenditures were $240 million, representing 4% of revenue.

Our capital planning aims to strike a balance between growth opportunities and meaningful returns while supporting the ongoing BiCS8 technology transition. We maintain strict discipline in evaluating such investments to ensure the long-term financial sustainability of the business.

Fourth Quarter Guidance: We project revenue to be between $7.75 billion and $8.25 billion, driven by increased BiCS production and higher pricing. The non-GAAP gross margin is expected to range between 79% and 81%. As we continue to ramp up innovation and R&D investment, non-GAAP operating expenses are forecasted to be between $480 million and $500 million. Non-GAAP interest and other income is projected to be between $10 million and $30 million, while non-GAAP tax expense is estimated at $775 million to $875 million. Based on fully diluted shares of 158 million, non-GAAP EPS is forecasted to be between $30 and $33.

Regarding capital allocation: Our priorities outlined in February last year were to invest in business growth, achieve a net cash position, and then return cash to shareholders. In line with these priorities, over the past two quarters, we have taken several measures to strengthen the supply chain, including extending the joint venture agreement with Kioxia until December 2034 and investing approximately $1 billion in Nanya Technology to secure long-term DRAM supply.

At the same time, we have taken steps to move toward a robust net cash position by repaying the remaining balance of the TLB. Given these positive developments, today we announced that the board has authorized a $6 billion common stock repurchase program, effective immediately with no expiration date.

Next, I will hand the meeting back to David for closing remarks.

David Goeckeler, Chairman and Chief Executive Officer:

Thank you, Luis. In summary, we remain steadfast in our execution at this pivotal juncture. NAND continues to be a foundational technology underpinning the world’s leading semiconductor memory solutions, driving the most significant waves of technological transformation—from PCs, mobile internet, cloud computing, to today’s artificial intelligence. The data center has become our fastest-growing market, and the workloads driving this demand—including inference, reasoning-based systems, and intelligent agent systems—represent a structural and enduring shift in how the world’s most critical technologies are built and deployed.

Our new business model is a reflection of this transformation: five signed agreements to date, financial guarantees exceeding 11 billion US dollars, and more than one-third of the bidding share for fiscal year 2027 under clear customer commitments. This marks a fundamental reshaping of our business, providing the company with greater visibility, more effective pricing protection, and more stable, long-term returns.

Our technology and product portfolio have converged perfectly with this extraordinary wave of demand at the right time, and we now possess a complete portfolio that includes a scaled and rapidly growing enterprise SSD business. We are aligning supply with the highest-value opportunities and establishing a new growth pillar for Sandisk.

These developments converge at a historic moment: we believe current profit margins are sustainable; we have achieved our net cash target; and we have announced capital returns to shareholders through stock repurchases—all while further solidifying our operational foundation. Combined with multi-year new business model agreements and accelerating growth in the data center end market, this gives us financial strength and structural resilience.

The result is a durable growth model, a highly valuable brand, and a business system capable of consistently generating substantial cash flow.

Ivan, let’s see if there are any questions.

Q&A Session
Host: Thank you. We will now begin the Q&A session. (Operator instructions) The first question comes from Mark Newman of Bernstein. Please go ahead.

Mark Newman, Analyst: Thank you very much for taking my question, and congratulations on another outstanding quarter. I have a few brief questions. Regarding the earnings per share guidance of 30 to 33 US dollars, these numbers are very impressive, but they also imply a slowdown in price increases this quarter. I would like to understand whether this is due to conservative considerations—given that we are still in the early part of the quarter—or is it related to these long-term agreements?

Regarding the long-term agreements, you mentioned that approximately one-third of BiCS production for fiscal year 2027 has been incorporated into long-term agreements. I would like to know the extent of price lock-in over the next few quarters. Thank you very much.

David Goeckeler, Chairman and Chief Executive Officer:

Mark, it’s great to hear your voice, thank you. Regarding pricing for the next quarter, we typically do not provide specific guidance on pricing. However, as you can see, there was significant price acceleration across the entire business in the third fiscal quarter. We are very pleased with this. You are correct that it is still early in the quarter, and the market is extremely dynamic, so it is reasonable to maintain a degree of conservatism in our forecasts. Nevertheless, we are highly confident in these numbers.

Regarding the price-locking issue in the agreements, let me first make a few points, and Luis will add more. These agreements are tailored to different customers with varying structural arrangements depending on the contract duration. The core of these agreements is to provide us with guaranteed demand consistency, which is exactly what we need most. We operate factories with highly stable output, so we require very steady consumption. This is precisely the role these agreements play — our clients also clearly understand this logic. This is why these agreements were not achieved overnight — they are not merely advance payments for supply over a few quarters but rather long-term agreements for continuous, stable, quarterly deliveries spanning five years.

As mentioned earlier, these agreements include financial instrument arrangements: if a customer fails to complete purchases according to the agreed schedule, the financial compensation obligation will immediately transfer to us. These agreements are strongly backed and highly effective. In terms of pricing, there are both fixed and floating elements.

Luis will further elaborate on the details.

Luis Visoso, Executive Vice President and Chief Financial Officer:

I would like to emphasize and supplement what David has said. The objective of these new business models is to deliver more enduring, predictable, attractive, and stable financial results, which is a win-win for both us and our clients — we provide supply, and they provide demand, with visibility extending up to five years. We are extremely satisfied with this arrangement.

As we have previously noted, you may not have heard us discuss RPO (Remaining Performance Obligations) in this industry before, but now we are beginning to do so. As mentioned, in our 10-Q quarterly report, you will see an RPO of approximately $42 billion.

Regarding pricing, as we have stated, it is a combination of fixed and floating components. Shorter contracts have a higher proportion of fixed elements, while longer contracts incorporate more floating elements. Therefore, it is reasonable to assume that pricing is predominantly fixed in the short term, with floating components increasing over time, allowing us to share in the benefits of price increases while also enabling customers to gain certain advantages during price declines.

Mark Newman, Analyst: Thank you very much.

Host: The next question comes from Joe Moore at Morgan Stanley. Please go ahead with your question.

Joseph Moore, Analyst: Thank you. I would like to understand the growth of enterprise SSDs, which has been quite impressive. How much of this growth is driven by overall market expansion, and how much is attributable to your company's improvements in product mix?

David Goeckeler, Chairman and Chief Executive Officer:

The growth of 233% quarter-over-quarter is driven by multiple factors. First, the improvement in product mix – currently, nearly all sales are from our TLC products, and next quarter we will begin generating revenue from QLC products under the Stargate project. Product performance remains exceptionally strong, qualification processes continue to expand, and we are adding new customer accounts consistently.

Second, there is definite market pull, with robust demand for high-performance enterprise SSDs. We have the right product at the right time. This portion of our product mix accounted for 25% of our total portfolio this quarter, and we expect that percentage to continue growing.

Joseph Moore, Analyst: Looking ahead, what proportion of your business do you see enterprise SSDs comprising? It seems hyperscale cloud providers are eager to purchase everything you can produce.

David Goeckeler, Chairman and Chief Executive Officer:

It could be quite significant. Joe, as you know, we always focus on a balanced product mix and adjust our structure to maximize financial returns. However, the key point is that we are now in a position to tilt supply toward the data center market like never before. I anticipate this proportion will continue to rise over the next several quarters and years.

Host: The next question comes from Ben Reitzes of Melius Research. Please proceed.

Ben Reitzes, Analyst:

Hello everyone, it feels a bit like a software company earnings call today, and I'm very pleased to participate. David, you mentioned that about one-third of next year's bit output is already contracted. Where do you expect this percentage to reach eventually? Could it exceed 50%, or even start the year with a substantial portion already locked in?

David Goeckeler, Chairman and Chief Executive Officer:

First, welcome to the team, and it's great to hear your voice again. I would like to make a few points: we are currently engaged in numerous negotiations regarding how to transform this business, and the pace of progress varies by client. Some clients are highly interested in multi-year supply agreements, making the negotiation process relatively smooth; others are accustomed to the traditional quarterly bidding model, which we have no interest in – what we seek are agreements that provide us with economic certainty.

Luis made an important point in his remarks, and I want to ensure everyone is clear: these agreements contain both fixed and floating components, but the financial returns targeted by the five agreements we have signed align with the guidance levels we just provided. This represents a highly attractive business opportunity.

We are still actively negotiating future supply contracts, covering next year and extending up to the next five years. I expect this proportion (we are referring to "at least one-third") to continue rising. Could it exceed 50%? I believe it is entirely possible, and we aim to push this figure significantly higher.

Ben Reitzes, Analyst: Thank you. Additionally, regarding profit margins, Dave, do these types of agreements allow you to lock in profitability? The market does not seem to be valuing your company at an 80% margin but rather anticipates a decline to the mid-40s. Do you have a target range for profitability, and are you willing to share a rough estimate?

David Goeckeler, Chairman and Chief Executive Officer:

Ben, I don't think this is the right time to discuss this issue. When we advance further, we will present a new comprehensive model. However, we are extremely proud of our technology. In my view, this may be the first time in decades that the industry has begun to genuinely recognize the value of technology – at least from our perspective. Frankly, the market has always acknowledged the value of our technology, but in the past, that value was captured by others, not the producers. I believe value distribution is now becoming more equitable.

We do not intend to trade value for certainty; instead, we aim to achieve both value and certainty. We understand how the market evaluates brands, and we are highly focused on eliminating the cyclicality of the business. Cyclicality is corrosive – it undermines our capital expenditure investment decisions and hinders customers' ability to secure sufficient product to drive their impressive operations.

We believe we have taken very meaningful steps forward: substantial commitments from heavyweight clients. As we continue to make progress, the entire business will transition to a new level that benefits everyone – benefiting us, investors, customers, and even customers' customers.

This journey has only just begun. Over the past year, many people told me this would never happen. But it is happening, although we are still in the early stages. We will keep you updated as we progress.

Moderator: The next question comes from C.J. Muse at Cantor Fitzgerald. Please go ahead.

C. J. Muse, Analyst: Good afternoon, and thank you for taking my question. I would like to get your perspective on the future supply and demand dynamics of NAND. Currently, greenfield capacity expansion is quite limited, with bit output growth primarily driven by stacking layers, while new greenfield capacity is largely prioritized for DRAM. Against this backdrop, coupled with incremental growth from smart agent AI, how do you foresee the industry potentially reaching a supply-demand equilibrium?

David Goeckeler, Chairman and Chief Executive Officer:

CJ, as you know, my view is that markets are always in balance, and supply and demand will eventually reach equilibrium. I think your underlying question is whether price reductions will create additional demand? That's the logic we are trying to avoid.

Starting with the demand side, even before what we observed this week, we had already revised our data center growth forecast for 2026 upwards to the mid-70% range, compared to 60% three months ago, 40% six months prior, and 20% before that. Data center demand remains exceptionally strong. Outside of data centers, we are seeing some market contraction due to declining equipment shipments, which was expected, but anticipate a rebound in 2027.

On the supply side, this is one of the core strengths of this brand: we can increase supply through node transitions without substantial capacity expansion. Our R&D pipeline is highly efficient, with the BiCS roadmap benefiting from decades of deep collaboration with joint venture partners. We can continue to achieve high single-digit BiCS bit output growth through node transitions, where each new node involves more process steps, requiring additional equipment, thus involving some incremental capital expenditure. However, this differs from industries that must build new capacity to achieve output growth.

Frankly, this is why this brand is such an exceptional cash generator: the ratio of capital expenditure required relative to revenue continues to decline. The absolute amount is there, but as a proportion of revenue-generating ability, it keeps decreasing. Our node transition plans have years of runway with a clear roadmap, and we will continue advancing these transitions to expand market supply at a high single-digit rate. This situation is generally similar across participants in the NAND industry.

C. J. Muse, Analyst: Very helpful. Regarding capital structure, your company currently has no debt and holds $3.7 billion in cash. Given the current business outlook and newly signed contracts, how much cash do you believe is necessary to retain? How should we interpret stock repurchases? Thank you.

Luis Visoso, Executive Vice President and Chief Financial Officer:

We announced a $6 billion share repurchase program effective immediately with this earnings call. We will continue to monitor cash flow – we are generating good cash. As conditions evolve and the buyback progresses, we will keep CJ and everyone updated.

Host: The next question comes from Jim Schneider at Goldman Sachs. Please go ahead.

Jim Schneider, Analyst: Good afternoon, and thank you for taking the question. One more on the new business model: Are any of the top five U.S. hyperscale cloud vendors included in the five contracts signed to date?

Additionally, do you plan to provide disclosures in forms such as Annual Contract Value (ACV) or confirmed contract value within the year going forward?

Luis Visoso, Executive Vice President and Chief Financial Officer:

We will not disclose customer names. However, as David mentioned, we have onboarded some very significant customers, with more in progress, but we cannot specify the names.

Regarding the second question, we will disclose the RPO metric on a quarterly basis — the contracted business volume based on minimum commitments, which is a highly informative indicator reflecting the portion of our business that has been contracted. We will continue to provide this information every quarter to allow everyone to track progress.

Jim Schneider, Analyst: Thank you. Additionally, given the enhanced visibility into customer demand brought by the new business model, what discussions are you currently having with Kioxia regarding a potential increase in bit supply? Are you considering plans that exceed the previously discussed growth target of around 20%?

David Goeckeler, Chairman and Chief Executive Officer:

Our discussions with Kioxia remain deep and ongoing, with both teams collaborating daily. Our BiCS8 transition plan is well-defined and being executed smoothly, with no current plans to alter the arrangement.

Host: The next question comes from Blayne Curtis at Jefferies. Please proceed.

Jake Wilhelm, Analyst: Hello, and thank you very much. Congratulations again on another strong performance. Regarding Stargate, which is expected to begin generating revenue in the fourth quarter, could you elaborate on the ramp scale over the coming quarters?

David Goeckeler, Chairman and Chief Executive Officer:

In the data center sector, there are two major product directions that we have mentioned multiple times before: the first is enterprise SSDs for compute-intensive scenarios, which have relatively lower capacity but extremely high interface speeds; the second is the ultra-high-density QLC products represented by Stargate. Progress to date has mainly come from compute-oriented drives primarily using TLC, and now we are bringing QLC products fully to market—these products have completed qualification processes exceeding one year with several major customers. We do not forecast the size of specific market segments individually, but we are very proud of this product and believe it will perform exceptionally well in the market.

Jake Wilhelm, Analyst: Alright, thank you. With the recent release of more powerful LLMs, how do you view the opportunities in KV caching and the growth of intelligent agent AI? Has there been a noticeable shift in customer discussions over the past few quarters?

David Goeckeler, Chairman and Chief Executive Officer:

Over the past one or two quarters, this topic has become a significant part of our conversations, and our understanding of it has deepened considerably—the team even held a dedicated webinar on this subject, and if interested, we would be happy to host it again.

When you truly delve into analyzing this opportunity and attempt to quantify it, it quickly becomes very complex: variables such as the number of concurrent sessions, average input token count, cache hit rate, and storage duration need to be considered. This underscores the necessity of maintaining very close partnerships with our customers because they are the ones who possess all the detailed infrastructure construction knowledge—those customers deploying hyperscale infrastructure have profound insights into how various variables and use cases combine.

This forms the foundation of a new business model: when our customers go through these calculations and truly understand the scale of NAND demand, it provides an excellent basis for reaching large procurement agreements spanning two, three, or even five years—five contracts covering more than a third of the product portfolio exemplify this understanding.

The market changes every day. Even those of us immersed in it and constantly tracking the data can feel its rapid evolution. Our customers are deeply understanding the logic behind the infrastructure they are building and responding enthusiastically to our products. They are willing to make multi-year procurement commitments because it gives them supply assurance while providing us with attractive financial returns. More importantly, they are willing to take on substantial financial guarantees to ensure continued purchases.

There are sometimes external doubts about whether these agreements can be implemented or honored. I can state unequivocally that the reality is quite the opposite. Our customers are genuinely securing these deals with collateral amounting to billions of dollars—through various financial instruments managed by third-party financial institutions, effective throughout the contract period. If they fail to meet quarterly obligations, corresponding financial compensation will be paid to us immediately.

Of course, I do not anticipate needing to invoke these guarantees because our customers have an urgent demand for these products. Typically, within weeks of signing the agreement, customers begin discussing with us how to increase delivery volumes within the agreed timeframe. This is an extremely dynamic market, evolving very rapidly, and indeed poses challenges for forecasting. Our forecasts from three months ago were far below current actual results precisely because the market changes every day.

We are aware that we possess outstanding technology, have made substantial investments in intellectual property, and spent billions of dollars on factory construction. We co-own one of the largest wafer fabrication complexes in the world with our joint venture partners, and our BiCS roadmap is the result of decades of effort. We aim to fully leverage all these investments to achieve fair returns on our products and eliminate the cyclicality of the business. Cyclicality is detrimental to all participants in the industry—at least from our perspective.

Now, there is already a significant number of high-profile clients who no longer wish to engage in quarterly bidding games. They run excellent businesses and understand that we provide critical components essential to their operations. They want to ensure continuous access to top-quality products—and we believe we offer exactly that. This is driving a fundamental transformation in the business model of this industry—a change that has not occurred in several decades. Frankly, this is an exciting development for us because clients are highly satisfied with these agreements, as are we. With every signed agreement, clients have genuinely been delighted to reach the signing stage.

This may be a longer response than you anticipated, Jake, but we feel very, very confident about where we stand, and we believe we are on a new trajectory. As Ben mentioned, other technology industries have long understood how to achieve this—we are not reinventing the wheel; we are simply introducing a recurring revenue model that has proven effective in other industries into our own brand.

Jake Wilhelm, Analyst: Congratulations once again.

Host: The next question comes from Asiya Merchant at Citigroup. Please go ahead.

Asiya Merchant, Analyst: Thank you for taking my question. A very impressive set of figures, and it’s great to speak with you again. David, I heard you mention some potential rebound in client demand (PC or smartphone-related) next year, and you seem optimistic about it. Could you share the factors supporting this optimism, such as whether AI-driven edge computing trends are playing a role? In this context, given that supply is more oriented toward meeting hyperscale data center demands, how can you ensure that client-side needs will also be met?

Additionally, a question for Luis: Capital expenditure used to account for a mid-teens percentage of revenue, and now, with revenue growing rapidly, we certainly do not expect to maintain the same ratio. How should we forecast capital expenditures going forward? Thank you.

David Goeckeler, Chairman and Chief Executive Officer:

Let me break it down. Regarding the outlook for 2027: Both PC and smartphone shipments are currently declining, which is expected. However, we anticipate that by 2027, both will stabilize, with PCs showing a slight increase. Smartphone content configurations, at least for smartphones this year, will continue to grow, while PCs remain flat. Next year, we expect both categories to rise again, with shipments transitioning from flat to slightly increasing.

Regarding supply allocation, this is at the core of the changes we are implementing. We are not waiting until next year to see what the market gives us—we are also negotiating multi-year new business model agreements with edge clients. These agreements are structured similarly to the previous five: committed, incremental demand. These clients deeply understand their businesses, and negotiations are progressing. I believe there will be results. At that point, our insight into their needs will be very clear because they will explicitly inform us backed by financial commitments, allowing us to better plan supply. We are transitioning from the past "market fluctuation, constant adjustment" model to a cooperative model where clients clearly commit to their needs with financial backing—both parties can truly rely on each other.

Luis Visoso, Executive Vice President and Chief Financial Officer:

Regarding capital expenditure, David has already elaborated on our philosophy: we will continue to invest in capacity growth at a mid-to-high single-digit rate to drive bit production growth. In terms of specific amounts, this should not be considered as a percentage of revenue but rather in absolute terms.

In the next few quarters, the absolute amount of capital expenditure will increase slightly as we have already completed relatively easier node transitions, and the subsequent transitions will require more expensive equipment tools to achieve the same level of output growth. The overall direction remains unchanged; only the absolute amount of capital expenditure will increase slightly during these node transitions. Please take this into account when modeling.

Moderator: The next question comes from Vijay Rakesh of Mizuho Securities. Please proceed with your question.

Vijay Rakesh, Analyst: Thank you, David and Luis, for the impressive results. Regarding RPOs and financial guarantees, I noticed that data center revenue for the quarter reached 1.5 billion dollars, which annualizes to approximately 6 billion dollars. Does this mean that most of the 11-billion-dollar financial guarantee and the 42-billion-dollar RPO are derived from the data center segment? Additionally, regarding the pricing mechanisms in these long-term guarantees, are they dynamically adjusted according to market conditions? I also have a follow-up question.

Luis Visoso, Executive Vice President and Chief Financial Officer:

We do not disclose customer information, so I cannot provide further details on the breakdown by source. However, regarding the 42-billion-dollar RPO, many companies have the scale to enter into such agreements—this represents the minimum contractual revenue confirmed through three contracts signed before the end of the quarter, with two additional contracts to be reflected in next quarter's report, which will result in an even higher RPO figure.

As for the 11-billion-dollar financial guarantee, we utilize various financial instruments: part of it is reflected as prepayments (approximately 400 million dollars have been recorded on the balance sheet), while the remainder is managed through financial arrangements facilitated by third-party financial institutions, which trigger compensation mechanisms if the contract obligations are not fulfilled within the agreed timeframe. This is the general mechanism of operation.

Vijay Rakesh, Analyst:

David, regarding your company’s NAND SSD roadmap, High Bandwidth Flash (HBF) appears to be a highly disruptive technology. Could you provide an update on its progress?

David Goeckeler, Chairman and Chief Executive Officer:

Progress has been steady, and we are in discussions with customers regarding their deployment methods. Development of both the NAND die and controller is advancing, and we remain on schedule: completing the NAND die later this year and launching the full system-level product with the controller in the first half of next year.

Moderator: The next question comes from Blayne Curtis of Jefferies. Please proceed with your question.

Blayne Curtis, Analyst: Good afternoon, and congratulations on the excellent results. I’ve been hearing more discussions about different memory tiering architectures, such as accelerators using more DRAM. Regarding KV caching, there seems to be a general assumption that it will rely heavily on enterprise-grade SSDs. However, as memory tiering architectures and high-bandwidth flash roadmaps evolve, have your views on storage hierarchy architecture changed over the past quarter or so?

David Goeckeler, Chairman and Chief Executive Officer:

In terms of tiered architecture, the mainstream solution that emerged roughly a quarter ago is essentially what is being deployed now. High-bandwidth flash is not necessarily a replacement for enterprise-grade SSDs but serves to provide higher storage density for inference processes in a different way. We can delve into the specifics further if needed.

From our data, the pull for high-performance enterprise-grade SSDs is obvious and substantial — driven by the deployment of various architectures and the scaling up of inference operations.

As mentioned in our prepared remarks, given factors like model size, KV cache size, and context length, NAND has become an indispensable key component of these architectures — NAND is the most scalable semiconductor technology in the world and is now at the forefront. I expect this architecture to continue evolving and refining, which is why we maintain very close collaboration with our customers — those major customers deploying at scale have detailed plans and insights on how to expand global infrastructure. Understanding their needs is the foundation that drives our demand signals for the coming years and supports new business models.

Moderator: The next question comes from Amit Daryanani of Evercore ISI. Please proceed with your question.

Victor Santiago, Analyst: Good afternoon, this is Victor Santiago speaking on behalf of Amit. Thank you for taking the question. Regarding the investment and supply agreement with Nanya announced last month, could you help us better understand its strategic intent? Is it primarily to secure DRAM for high-bandwidth flash or future memory products, or are there other considerations?

Luis Visoso, Executive Vice President and Chief Financial Officer:

As you can see, our data center business has performed quite well this quarter, achieving impressive growth driven solely by TLC products. With the introduction of QLC products and the impetus from new business models, further growth is expected.

DRAM is one of the key foundational materials in our continued expansion of the data center business. The partnership with Nanya Technology not only represents a financial investment in the company but also grants us priority access to DRAM supplies. This is the strategic rationale behind this investment.

Victor Santiago, Analyst: Understood, thank you. A follow-up question: you mentioned the longest contract duration. Could you reveal the average duration of the five agreements?

Luis Visoso, Executive Vice President and Chief Financial Officer:

Frankly speaking, we are currently unable to provide that level of detail or calculate an average. I apologize for this, Victor.

David Goeckeler, Chairman and Chief Executive Officer:

Let me add a point, Victor. We aim to maintain a diversified portfolio of agreements — with contracts spanning one year, two years, three years, and five years, distributed in such a way as to avoid the risk of all contracts expiring simultaneously. We anticipate renewing some of them upon expiration, and certain contracts include renewal options after their initial terms.

This is still a relatively early stage, and there is no fixed template for the overall framework. Each agreement is tailored to specific clients. As we progress further, we may revisit discussing the numbers you're interested in, but now is not the time.

Moderator: The next question comes from Wamsi Mohan at Bank of America. Please proceed with your question.

Ruplu Bhattacharya, Analyst: Thank you for taking my questions. This is Ruplu speaking on behalf of Wamsi. I have two brief questions: The first one is for Luis—Are there any restrictions on price increases within long-term agreements? Are annual adjustments allowed, or are there specific conditions? The second question is for Dave—How do you view the trend of interest in QLC flash memory, and how do you expect the mix between TLC and QLC to evolve over the next few quarters?

Luis Visoso, Executive Vice President and Chief Financial Officer:

Regarding the pricing details of each contract, I can only say this: As mentioned earlier, each agreement contains both fixed and variable pricing components, and there are significant differences between the agreements, so it is not possible to provide a universal, overarching answer.

David Goeckeler, Chairman and Chief Executive Officer:

Regarding TLC and QLC, from the perspective of the overall product portfolio, it is roughly two-thirds TLC and one-third QLC. In terms of data centers, TLC currently dominates, but we will launch a major QLC product next quarter.

The strong demand for TLC in data centers is primarily driven by the high-performance requirements of inference architectures, particularly in enterprise SSD application scenarios, as well as the previously discussed KV caching demands—where storage needs can expand exponentially across different usage scenarios, creating very strong demand for TLC. Of course, we also anticipate that QLC products will perform well.

Host: That concludes the Q&A session. I will now hand the meeting back to Ivan Donaldson for closing remarks.

Ivan Donaldson, Vice President of Investor Relations: Thank you all for joining today's conference call. We appreciate your support and look forward to staying in communication with you throughout the quarter.

Host: This concludes today’s meeting. Thank you for attending today's presentation. You may now disconnect.

Editor/Jeffy