THE FULL TEXT AND ANALYSIS OF THE SPEECH AT BERNSTEIN ' S 42ND ANNUAL STRATEGIC DECISION-MAKING SESSION
by invest wallstreet
Sandisk CEO David Goeckler28 May42nd annual strategic decision-making session in BernsteinThe presentation centred around five key strategic themes:FOR THE FIRST TIME, THE DATA CENTRE WILL BE THE LARGEST MARKET FOR NAND, THE RESHAPING OF THE NEW LTA BUSINESS MODEL, RESTRAINED SUPPLY DISCIPLINE, CONTINUOUS UPGRADING OF THE TECHNOLOGY MIX AND THE “DE-CYCLICAL” INDUSTRY NARRATIVE OF NAND。
I. STRENGTHENING OF THE CORE ELEMENTS OF THE STATEMENT
1. AI-DRIVEN DATA CENTRE MARKET RISE
Goeckeler clearly described the structural changes that are taking place in the NAND flash market. The core judgment is:THE DATA CENTRE IS EXPECTED TO MOVE BEYOND THE MOBILE END MARKET FOR THE FIRST TIME IN 2026 (CALENDAR YEAR) TO BECOME THE LARGEST APPLICATION MARKET FOR NAND FLASHTHIS MARKS THE END OF THE LAST DECADE OR SO OF THE STORAGE CYCLE LED BY SMARTPHONES AND PCS, AND AI INFRASTRUCTURE IS BECOMING THE NEW CORE GROWTH ENGINEI don't know. The unprecedented visibility of client demand, which has been extended to 2028, has led to in-depth discussions between companies and clients around supply agreements for one to five yearsI don't know。
2. Adaptation of business models in long-term supply agreements
The most strategic part of this speech is:Sandisk is moving from the traditional quarterly pricing negotiation model in the NAND industry toFULL TRANSFORMATION OF THE NEW BUSINESS MODEL WITH LONG-TERM SUPPLY AGREEMENTS (LTA) AT ITS COREI don't know。Goeckeler states that the company is working to establish multi-year contracts that “can strike a balance between predictable demand and attractive economic returns”, which will transform long-term agreements from a price negotiation tool to a key supply security mechanismI don't know. He made it clear that the core goal of the company was “to maximize the two things — to obtain an attractive financial return and make it sustainable”I don't know。
CFO Luis Visoso adds:Clients willing to make long-term commitments will be given priority in an environment of tight supply, while customers insisting on quarterly negotiations may face supply constraintsI don't know。This strategy clearly demonstrates the strategic intent of companies to use shortfalls as leverage and push customers to embrace new business models. According to the financial statements, Sandisk had cumulatively signed five long-term supply agreements in fiscal year 2026, three of which had been signed in the third quarter, the other two having been added at the beginning of the fourth quarter and the partial duration of the agreements extending beyond 2030。
3. Supply strategy of restraint: rejection of blind expansion
Against a backdrop of extremely high demand, Goeckeler has demonstrated a significant strategic restraint from industry giant behaviour in previous storage cycles. He saidAny substantial increase in capital expenditure requires a high level of confidence in multi-year demand at attractive price levelsI don't know. The company's adherence to its existing medium- and high-digit-digit capital growth capital expenditure plan did not result in a surge in short-term prices, while rejecting market speculation about the early deployment of BiCS10 technology, indicating that existing capacity programmes were sufficient to meet demandI don't know. Such disciplined supply management essentially draws on historical lessons from the “blind spread of post-partum price wars” in the storage industry。
4. Technology and product mix layout
On the product and technology level, Goeckeler disclosed a number of key developments: the company is expected to use BiCS8 as the lead technology node by the end of the 2026 fiscal year; new products are being certified in cooperation with mega-clients, including a 128 TB QLC enterprise-level SSD, known as “Stargate”, which has not yet started operations but is in the certification stage; THE COMPANY ALSO WORKED WITH SK HERCULES TO DEVELOP HIGH-BANDWIDTH FLASHING TECHNOLOGY FOR AI REASONING LOADS, WHICH ARE SAID TO BE 10 TIMES MORE DENSE THAN THE ALTERNATIVE, WITH PROTOTYPE CHIPS EXPECTED TO BE LAUNCHED BY THE END OF 2026 AND TESTED FOR CUSTOMERS ABOUT A YEAR LATERI don't know. In addition, Sandisk has extended the joint venture agreement with the Zhang Man for five years to ensure that manufacturing capacity continues until 2034, enabling companies to invest in research and development at a level comparable to that of the highest market share competitors and to maintain cost leadershipI don't know。
5. Supply side discipline is reshaping the industry Bureau
In his speech, Goeckeler made it clear that:THE CURRENT TENSION IN THE NAND MARKET ISStructural change driven by real demand, not short periodic fluctuationsI don't know。This judgement is the basis for a series of strategic adjustments in Sandisk. AI-driven storage demand is reshaping the NAND value chain, driving demand significantly above supply, and this imbalance is expected to continue beyond calendar year 2026I DON'T KNOW. AT THE SAME TIME, OTHER KEY DATA SUPPORT THE CONCLUSION THAT NO NEW NAND CRYSTAL ROUND CAPACITY INCREASED BETWEEN 2026 AND 2027 AND THAT DEMAND WILL EXPAND BY MORE THAN 20 PER CENT BY 2026, AND THAT THIS ONGOING SUPPLY AND DEMAND MISMATCH IS PUSHING NAND PRICING TO UNPRECEDENTED LEVELS- I don't know。
Notably, the price of 128 Gb MLC NAND jumped from $2.18 in early 2025 to $17.73 in March 2026In the second quarter, the NAND flash contract price increased by about 70 to 75 per cent, exceeding the DRAM increase of 58 to 63 per cent over the same period. Against this background, Goeckeler emphasized that the strategic focus of the company was to move from the pursuit of output to the creation of sustainable profits, and to move low-profit consumer-level markets forward and focus heavily on high-end enterprise-level SSD in data centres。
II. In-depth strategic intent interpretation
1. Strategic positioning regression after separation
To understand the strategic significance of this speech, it is important to review Sandisk ' s recent history. Sandisk officially split independence from Western Digital on 24 February 2025, driven by external pressures such as radical investor Elliott Management — the core logic is to eliminate the “comprehensive enterprise discount” that had previously covered the business potential of Flash, each becoming a more pure standard in the AI supercycleI don't know. Goeckler himself was the centerpiece of Sandisk's break-up and subsequently his independent director and CEO, and his presentation of the company's future strategic direction, which naturally carries with it the important mission of laying the foundations for Sandisk's positioning and reshaping market expectations after independenceI don't know。
NAND “DE-CYCLICAL” NARRATIVE CONSTRUCTION
The most controversial but central strategic narrative of Goeckeler's speech is his report on his workNAND IS MOVING AWAY FROM ITS TRADITIONAL STRONG CYCLE ATTRIBUTES INTO AN INDUSTRY THAT IS “MORE RESILIENT, MORE STRUCTURALLY ATTRACTIVE AND WITH HIGHER AVERAGE LONG-TERM RETURNS”THE JUDGMENT. MANAGEMENT ALSO STATED IN ITS PRESENTATION THAT NAND WAS BEING REDEFINED AS A “KEY COMPONENT OF AI INFRASTRUCTURE” RATHER THAN A TRADITIONAL CYCLICAL BULK COMMODITYI don't know。
The narrative triggered strong reactions in the capital markets. The market, based on the logic of “NAND de-cyclicalization through a very large cloud manufacturer's long-term contract”, pushed Sandisk's return from the beginning of the year to date to over 356 per cent, and the stock price jumped from about $33 to $1406 per share at break-off (as of early May 2026)I don't know. Evercore ISI's first coverage of Sandisk gives a big run-win rating of $2,600 for the cattle market target; Bernstein himself gives an extreme target price of $3,000 and increases the price target from $1250 to $1,700I don't know。
However, there remains a considerable challenge to the authenticity of this “decyclical narrative”. The market generally considers that Sandisk's non-GAAP market share has reached 40.36, far above the level of US$ 26.30, with a significant margin in valuation premiums. HighI DON'T KNOW. MORE NOTABLY, THE CORE WEAKNESS OF THESE NASAS IS THAT THE LONG-TERM COMPONENT USES FLOATING PRICING MECHANISMS — WHICH MEANS THAT WHILE SOME $42 BILLION IN FINANCIAL GUARANTEES (RPOS) CAN PREVENT CUSTOMERS FROM BREAKING UP, THEY CANNOT RESIST FALLING PRICESI DON'T KNOW. ONCE THE NEW CAPACITY OF SAMSUNG AND KK HERCULES (E.G. 321 FLOOR NAND) IS RELEASED BETWEEN 2026 AND 2027, FLOATING PRICING WILL BE RE-PRICING WITH THE SPOT MARKET, AND THE SUSTAINABILITY OF LONG-TERM DEMAND IS FAR FROM BEING CONCLUSIVELY PROVEN。
III. Financial and performance context
The recent explosion of Sandisk's financial data is a key context for understanding CEO confidence and capital alignment. The company received $5.95 billion in the third quarter of fiscal year (as at the end of March 2026), a significant increase of 252.1 per cent over the same period, and the non-GAAP gains of $23.41 per share significantly exceeded market expectations. However, it is even more important to note that the underlying driving force of this over-expected performance is mainly the optimization of pricing rights and the product mix, rather than the growth in the volume of deliveries — the volume of bits is the same and the ring is fallingI don't know。
THE MORE STRIKING NON-GAAP MĀORI RATE OF 78.4 PER CENT, WELL ABOVE THE HISTORICAL AVERAGE OF 30 TO 40 PER CENT IN THE INDUSTRY, IS ONE OF THE STEEPEST CLIMBS IN THE HISTORY OF THE SEMICONDUCTOR INDUSTRY. HOWEVER, THE MAINTENANCE OF THIS UNPRECEDENTED LEVEL OF GROSS DOMESTIC PRODUCT IS ALMOST ENTIRELY DEPENDENT ON A 645 PER CENT INCREASE IN DATA CENTRE OPERATIONS OVER THE SAME YEAR, AND MANAGEMENT HAS MADE EXPLICIT REFERENCE TO “UNABLE TO MEET THE FULL NEEDS OF CLIENTS” AT THIS QUARTERLY FINANCIAL CONFERENCEI don't know. The statement of supply shortfalls conveyed in the speech is highly consistent with these alarming financial figures。
IV. Far-reaching implications for storage industry patterns
The industry model moves from “price warfare” to “supply discipline”
Sandisk's supply discipline — refusing blind expansion and recasting customer relationships with LTA — is reshaping the competitive paradigm across the NAND industry. If this strategy is followed and emulated by other major NAND producers (including Samsung, American Lights, SK Hercules, etc.), it will be a good idea to be able to use the new technology in the futureIT MAY MEAN THAT NAND WILL MOVE BEYOND A NEW GROWTH CURVE THAT IS STABLE IN DEMAND AND MANAGEABLE IN CAPACITY, IN SHARP CONTRAST TO THE INTENSE CYCLICAL FLUCTUATIONS THAT NAND HAS ALWAYS HAD SINCE ITS BIRTHI don't know。
2. Transfer of trade voice to suppliers
Against the backdrop of the expansion of AI infrastructure, NAND ' s strategic position as the base of AI storage has increased significantly, and there has been a reconciliation of bargaining positions between suppliers and mega-clients. The willingness of clients to provide demand projections extending through 2028 and supply commitments up to five years reflects the real nature of this structural change. The central industry signal from Goeckeler is here: storage is no longer a cost item but a key strategic component of AI infrastructureI don't know。
V. Outlook
On a positive note, Goeckeler did capture a real, AI-driven structural demand shift and decisively transform Sandisk from an integral part of an integrated storage company to a purely business company with a high focus on AI storage requirements. Dismantling itself, signing long-term contracts, explosive growth in data centre operations and clear technology road maps — These are objective changes in reality, not purely conceptual。
In general, the core value of Goeckeler's speech isIT PROVIDES A HIGHLY TRANSPARENT AND LOGICAL NARRATIVE FRAMEWORK FOR BUSINESS TRANSFORMATION. FOR INVESTORS AND INDUSTRY OBSERVERS, HOWEVER, THE DISTINCTION BETWEEN REAL NARRATIVES OF BUSINESS CHANGE IN SPEECHES (DATA CENTRE RISE, SUPPLY DISCIPLINE, LTA REMODELLING) AND UNTESTED VALUATION ASSUMPTIONS (“DE-CYCLICALIZATION”, LONG-TERMIZATION OF 78 PER CENT MĀORI) REMAINS A KEY PREREQUISITE FOR RATIONAL JUDGEMENT。
Sandisk Corporation (SNDK) Bernstein 42nd Annual Strategic Decision-Making Meeting
Company Partners
David V. Goeckeler - Chairman and Chief Executive Officers
Access Call Partners
Mark Newman - Bernstein Industrial Services LLC, Research Department
Provision
Mark Newman, Bernstein Industrial Services LLC, Research Department
Good afternoon. I'm Mark Newman, Bernstein's American IT hardware analyst. It is a pleasure to welcome David Goeckler again today, Chairman and CEO of Sandisk, who had previously served as Chief Executive Officer in Western Digital and dominated the split in Sandisk. So thank you very much for coming here again today, David。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Nice to be here, Mark. Thank you for inviting us。
Mark Newman, Bernstein Industrial Services LLC, Research Department
Thank you。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Can I start with a safe harbour? It seems I have to。
Mark Newman, Bernstein Industrial Services LLC, Research Department
Okay, go ahead。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Obviously only I can do this. In today's discussion, I will make forward-looking presentations based on current management assumptions and expectations, including on our technology and product mix, our business plan and performance, our capital allocation priorities, market trends and opportunities, and our future financial performance。
These forward-looking statements are influenced by risks and uncertainties. We have no obligation to update these statements. For more information on risks and uncertainties that may lead to significant differences between actual results and expectations, see our annual report form 10-K, the quarterly report forform 10-Q and other documents submitted to the SEC。
WE WILL ALSO REFER TO NON-GAAP FINANCIAL DATA, AND RECONCILIATION BETWEEN GAAP AND NON-GAAP PERFORMANCE CAN BE FOUND IN THE INVESTOR RELATIONS SECTION OF OUR WEBSITE。
Other Organiser
Mark Newman Bernstein Industrial Services LLC, Research Division
Thank you very much. Well, since that's all said, I'll start the question-and-answer session. I have a bunch of questions to ask. I remind you that you should all have a Pigeonhole link that you can submit your own questions in there. I've got an iPad here, and I'll check in on the questions. If I may, after I have asked myself some questions, I will try to pick up a few questions from the audience。
I WANT TO START WITH THE NEED. IF YOU CAN TALK ABOUT NEEDS FIRST, THEN WE CAN DISCUSS OTHER MATTERS. FIRST, IN TERMS OF DEMAND, IN VIEW OF CURRENT DEVELOPMENTS IN AI AND OTHER AREAS SUCH AS MOBILITY AND CONSUMPTION, CAN YOU DESCRIBE THE ENVIRONMENT OF DEMAND THAT YOU SEE NOW? WHAT ARE THE CHANGES IN DEMAND COMPARED TO THE LAST TIME WE DISCUSSED IT, PARTICULARLY WITH REGARD TO AI
David V. Goeckeler Chairman and Chief Executive Officer Officers
AT THE OUTSET, I WOULD LIKE TO SAY THAT THIS IS ONE OF THE REASONS WHY I VERY MUCH LIKE THE MARKET AND I VERY MUCH LIKE THE BUSINESS: THERE ARE MANY DEMAND DRIVERS. I MEAN, NAND IS USED IN ALMOST ALL THE INTERESTING TECHNOLOGIES IN THE WORLD. TRADITIONAL SMART MARKETS, SUCH AS SMART PHONES, PERSONAL COMPUTERS, DATA CENTRES (WHICH ARE CLEARLY GROWING SIGNIFICANTLY NOW), HAVE ALSO EXPANDED TO INCLUDE GOODS-NETWORKING EQUIPMENT, CARS AND ROBOTS。
That's how it continues. It is a very diversified market with many demand drivers. These drivers change at different rates. I think it really makes this a very interesting place for business development like ours. So what happens now? I don't think that's a secret to anyone. The data centre is indeed growing very rapidly。
If we go back to about three projection cycles, we initially thought that the data centre would grow by 20 per cent this year. We then moved the forecast upwards to 40 per cent and then to 60 per cent, and now we have a slight further upwards. If you look at the growth of the data centre in the year 2026。
IT DID HAPPEN. IT IS CLEARLY A MAJOR DRIVING FORCE IN THE MARKET, AND MUCH HAS HAPPENED. OTHER MARKETS, HOWEVER, REMAINED — DEMAND REMAINED STRONG ACROSS MARKETS. WHETHER IT BE PC OR SMARTPHONES, WE ARE STILL MAINTAINING GOOD COMMUNICATION WITH THESE CLIENTS, INCLUDING IN THE AREA OF AUTOMOBILES AND GOODS NETWORKING. I THINK THE OVERALL DEMAND ENVIRONMENT IS VERY ROBUST。
Mark Newman, Bernstein Industrial Services LLC, Research Department
I THINK, I MEAN, GIVEN THAT AI IS SO STRONG, DATA CENTRES ARE GROWING BY OVER 60 PERCENT, BUT SOME OF THE OTHER PARTS OF THE MARKET ARE BEING SQUEEZED OUT. THIS MAY BE PART OF WHAT WE SEE. HOW DO YOU EXPRESS THAT
David V. Goeckeler Chairman and Chief Executive Officer Officers
Look, I mean, it's a market, right? I believe that markets always rationalize supply and demand. They are always balanced to some extent. Clearly, some would be liquidated through prices. In this market, a large and very attractive demand environment is emerging. This is very, very exciting。
This will always affect other parts of the market that may be less attractive from an economic point of view. This happens at any time in any market. We happen to be -- it's a big market. It's a very liquid market. We know the price all the time。
IN FACT, THE MARKET HAS BECOME ACCUSTOMED TO CONSTANT PRICING TRANSACTIONS, EVEN IN THE CONTRACTUAL SEGMENT OF THE MARKET, PRICES ARE TRADITIONALLY SET QUARTERLY, AND THIS DOES BRING ABOUT SIGNIFICANT VOLATILITY. THAT IS ONE OF THE THINGS WE FRANKLY WANT TO GET RID OF. I THINK YOU WILL SEE THIS DYNAMIC IN ANY MARKET THAT CONTINUES TO CREATE A LARGE NUMBER OF NEW, ATTRACTIVE AND ACCESSIBLE MARKETS (TAM)。
You see, we came in to talk about this a little bit. Just a short time ago, I came to the stage in New York, when we completed the dismantling and roll-out of the company, and I said that we would invest to achieve a growth of up to 10 percentage points。
We believe our view — that was early 2025. By the end of 2025, we expect market pricing to be at a turning point. Until the end of the summer last year, the mainstream view considered this to be a mistake。
I saw reports until later (e.g. late summer) that, oh, Sandisk would fail in December because prices would fall. That has not been the case. I don't think anyone can predict the demand for data centres so intensely。
BUT WE BELIEVE THAT THIS HAS BEEN A VERY GOOD MARKET. WE HAVE BEEN INVESTING IN THE GROWTH OF THIS MARKET. WE MUST MAKE INVESTMENT DECISIONS MANY YEARS AHEAD OF ACTUAL SUPPLY. WE ARE INVESTING HEAVILY: BILLIONS OF DOLLARS IN CAPITAL SPENDING AND HUNDREDS OF MILLIONS OF DOLLARS IN IMPROVING R & D EFFICIENCY. BY THE WAY, WITH REGARD TO R & D EFFICIENCY AND NAND, THIS IS IN ITSELF A COMPLETE TOPIC。
It's amazing. For each node to be delivered, we can achieve substantial growth through additional capacity. But we have long been confident that markets can be scaled up, and we have committed ourselves to allowing them to grow at a rate of a few dozen per cent。
Mark Newman Bernstein Industrial Services LLC, Research Department
Back to the demand side. In addition to the main figures published, what are the leading indicators that you are looking at, such as order book depth, customer forecast revisions, accreditation activities? What are you looking at best to convince you that the current demand cycle is sustainable? It is clear that, given current pricing trends, demand far exceeds supply。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Yeah. What we want to see is what that environment will look like in the coming period — I mean, for example, we just invested $1 billion a few months ago to extend the agreement with Kioxia from 2030 to 2034, which is a very good one. So it's clear that we start thinking about what needs are in the long run。
We have many different ways to answer your question. First, we do a lot of bottom-up research, communicate with clients and understand what they are doing. It's like smartphones, personal computers, and so on. We are close to our clients and know what equipment they want to launch in the future。
We will have our own judgement on how the composition will be. We have undertaken various bottom-up analyses of major markets to assess what bits of growth will drive. We are clearly focusing on capital expenditure. I mean, the rise in capital spending per financial cycle is the reason for the increase in data centre-related figures。
We know that we have a fairly clear understanding of how this relates to our growth in the technological field. So these are our medium- and long-term concerns on the demand side. And then we're on the market every day. I mean, we're talking to clients. They have called us to talk about what they need now and in the future — obviously we are also constantly discussing pricing。
Thus, the convergence of all this allows us to see where the market is going。
Mark Newman Bernstein Industrial Services LLC, Research Division
IN PARTICULAR, AI, WE WENT THROUGH DIFFERENT STAGES OF AI TRAINING: THE EARLY IMPACT OF CHAT ROBOTS, THE LATER IMPACT OF MORE ADVANCED STAGES, AND NOW ENTERING THE AGE OF INTELLIGENCE. HOW DO YOU THINK THESE CHANGES WILL AFFECT NAND'S NEEDS OVER TIME AS WE MOVE FROM AN EARLY STAGE TO A LATER STAGE
David V. Goeckeler Chairman and Chief Executive Officer Officers
WE HAVE ALWAYS THOUGHT THAT REASONING IS THE REAL USE OF NAND. SO WE HAVE TO DO THIS. I WANT TO SAY THAT IN THE FIRST TWO OR THREE YEARS OF AI, I WAS OFTEN ASKED: WILL THE DEVELOPMENT AND DEPLOYMENT OF AI AFFECT NAND OPERATIONS? THE ANSWER IS ALWAYS YES, BUT WE HAVE TO DO IT FIRST, RIGHT
You have to build models, deploy them, replicate them, and get users to use the technology. There must be valuable examples of usage that can be promoted. I think we've gone through these stages. Now we are moving forward quickly, and you have begun to see its impact over the past year。
THIS HAS INDEED BEEN THE CASE. YOU'RE STARTING TO SEE THE IMPACT OF NAND ON THE REASONING STRUCTURE. I THINK AS OUR CLIENTS BEGIN TO BUILD THESE STRUCTURES, YOU'RE THINKING ABOUT HOW TO EXPAND REASONING GLOBALLY, RIGHT
In training, you do not need a truly global expansion. You're training a lot of very smart people and a lot of infrastructure to support training. But in terms of reasoning, you have to extend it to the public, if you like. Billions of people will use reasoning to some extent。
It must therefore be economic when you want to go through that process and expand it globally. Right? In the early stages of any technology, you'll naturally want to -- when you're a technician for the first time. These things, or when a market is set up, tend to use all the resources that may be needed to fill your architecture。
Give me all the computing skills, give me all the memory, give me all the power, give me all the network resources, give me everything I need, and then I'm going to build a system, but when you're going to expand it, you need to really figure out what to expand, and that must be economically viable。
Because if it was not economic, it would obviously be too expensive. That means you have to raise the charges. That means you're gonna give people the opportunity to come in and make it more economical and squeeze you out. So these decisions are very, very important and very difficult。
What I believe has happened over the past year is that those responsible for doing this — the amazing technology companies — have a wealth of expertise in expanding technology globally. I think it is indeed the story of the past 20 years。
I MEAN, THE DISTRIBUTION OF TECHNOLOGY IS ALMOST COMPLETELY OUT OF TOUCH, RIGHT? YOU CAN GET THE MOST AMAZING TECHNOLOGY IN THE WORLD IF YOU POINT THE DEVICE AT A URL. IT WASN'T LIKE THAT 10 OR 20 YEARS AGO, WHEN WE HAD TO SEND YOU SOMETHING, OR YOU HAD TO UPGRADE THE SOFTWARE, AND THERE WAS ALL KINDS OF FRICTION IN THE SYSTEM。
All those frictions have been eliminated, which means that we can deploy technology on a very rapid scale, which is amazing, isn't it? We are witnessing this. But those responsible for this work face very, very difficult tasks, because of the high cost, which you must do in the most economical way。
SO THOSE PEOPLE HAVE BEEN GOING THROUGH A PROCESS: HOW DO I BUILD THAT ARCHITECTURE? AND THAT'S WHERE NAND'S GETTING TO THE POINT OF VIEW. WHY? NAND IS VERY SCALABLE. IT IS THE MOST SCALABLE SEMICONDUCTOR TECHNOLOGY IN THE WORLD. WE CAN SUPPLY, RIGHT? SO, AS THE MODEL GETS BIGGER AND THE CONTEXT BECOMES LONGER, ALL THESE FACTORS ARE DRIVING YOU - IF THIS IS TO BE DONE ECONOMICALLY, MORE SCALABLE TECHNOLOGIES MUST BE USED. OR EVEN IF I WANTED TO DO IT, THERE WASN'T ENOUGH, WAS THERE? DRAM IS EXCELLENT TECHNOLOGY. HBM IS EXCELLENT TECHNOLOGY. IT HAS INCREDIBLE CHARACTERISTICS。
IT IS NOT OF SUFFICIENT SIZE TO SOLVE THE GLOBAL PROBLEM OF REASONING. SO I THINK THE COMPANIES HAVE BEEN TRYING TO FIGURE OUT WHAT STRUCTURE THIS IS, AND WE'RE STARTING TO SCALE UP AND LOOK AT HOW WE'RE GOING TO EXPAND IT. THIS IS WHY THE DEMAND FOR NAND HAS INCREASED BEHIND THE SCENES。
AND I FOUND OUT THAT WHEN I LOOKED AT THAT STRUCTURE AND PRECISELY CALIBRATED WHAT I WAS GOING TO BUILD, MY NEED FOR NAND WOULD INCREASE AND DECREASE, AND THE CONCLUSION WAS THAT WE NEEDED MORE. SO IT'S DRIVING THE MARKET, AND IT'S FORCING THE CUSTOMERS TO COME TO US AND SAY, "HEY, LOOK, WE'RE PLANNING FOR THE NEXT FEW YEARS — THAT'S OUR BUSINESS."
We would like to know your plans to supply us with this critical technology after many years. We do not want to come here on a quarterly basis to talk about prices and see if there is enough supply. Now we need to know: can you supply me in 2028? Can you supply me in 2029? That is why we are going through this comprehensive transformation。
Mark Newman Bernstein Research Services LLC, Research Department
Thanks. The needs we see now are amazing. Jensen Huang set out this KV Cache vision at CES early in the year, probably equivalent to an extra 17 TB per GPU. Have you seen this? Is this reflected in your needs data? Do you think this would have a significant additional impact on NAND's needs
David V. Goeckeler Chairman and Chief Executive Officer Officers
YEAH. I MEAN, THAT'S EXACTLY THE PROCESS I WAS TALKING ABOUT -- PEOPLE ARE DESIGNING AND CONFIGURATIONING SYSTEMS. KV CACHE IS MOVING TO NAND BECAUSE IT HAS TO BE EXTENDED. I NEED AN EXTENDED STORAGE TECHNIQUE, THAT'S NAND. DEPENDING ON THE EXAMPLE YOU'RE GOING TO BUILD, I KNOW PEOPLE WANT A VERY CLEAR MEASURE OF "IF I DO THIS, I CAN GET THAT MUCH" BUT THINGS ARE NOT THAT SIMPLE。
For example, you need to figure out which model you want to build in the future and what scale you want to expand; once you get it, you can design the architecture. When you walk through this process, you have a lot of variables. How big are your models, token, KV, cache size? Or do you have a cache somewhere. What's the chance of that cache
YOU HAVE TO GO THROUGH THIS VERY COMPLEX CALCULATION, AND WE HAVE DONE SOMETHING THAT WE HAVE SHARED. AND EVENTUALLY YOU'LL FIND OUT HOW MUCH NAND YOU'RE GOING TO NEED, AND THEN YOU'LL COME TO US OR OUR COLLEAGUES AND ASK ME HOW I'M GOING TO GET SO MUCH NAND IN THE NEXT FEW YEARS. SO WE BELIEVE THAT VERY MUCH。
I think this is far from just a vision. This is what is happening in reality, and it has been going on for quite some time now, because companies need to land and scale up this excellent artificial intelligence technology so that we can all use it。
Mark Newman, Bernstein Industrial Services LLC, Research Department
YEAH. I THINK YOU'RE SAYING IT'S ABOUT DENSITY, RIGHT? NAND FLASH AND DRAM DENSITY. NANDS ARE MUCH LARGER IN TERMS OF THE AMOUNT OF GB PER DOLLAR AND THE AMOUNT OF GB PER UNIT AREA。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Yeah. Just -- we can provide more -- we can provide more capacity, yes, more density. Exactly. It's a different technology. It solves different uses. It's not a substitute, and it doesn't mean either good or bad. This is not the problem at all. You need both。
And you have to use this highly scalable storage technology as part of the architecture, which is why the number of data centres is rising: As people repeatedly explore how to determine that architecture and how it needs to be expanded, the demands of scale continue to grow and the numbers continue to climb, pushing up demand。
Mark Newman, Bernstein Industrial Services LLC, Research Department
A slight shift to pricing, focusing on ASPs rather than LTAs, is simply looking at the pricing environment in the industry. For those who did not pay close attention to Sandisk, the average selling price per GB increased by about 140 per cent in Sandisk last quarter。
But that's my estimate. I don't think you really gave that exact number, but it's pretty amazing around here. My question is, how do you describe the current pricing environment? I mean, it is clearly impossible to sustain 100 per cent ring growth, which is not sustainable. But in different sub-markets, do you still see the strength continuing? Are prices still rising? Or what do you think
President and Chief Executive Officer of David V. Goeckeler Officers
See, I mean, we've got our own projections. I'm not going to discuss what future pricing will be. See, I mean, what we have to do is — most important in our business is to build very valuable technologies. Everything starts with technology. It's always about technology。
If you find yourself producing good products that can address real needs, we're on this journey to explore the value of these technologies. And our job is to do it, and we will continue to do it。
Mark Newman Bernstein Industrial Services LLC, Research Division
GOT IT. OKAY. SURPRISE IS A LITTLE TOO STRONG. OKAY. GOT IT. HISTORICALLY, NAND PRICING HAS BEEN QUITE CYCLICAL, AND WE'LL TALK ABOUT IT A LITTLE LATER. BESIDES LTA, WHAT GIVES YOU CONFIDENCE IN THE SUSTAINABILITY OF THIS PRICE LEVEL, BECAUSE WE'RE GOING TO TALK TO LTA。
David V. Goeckeler Chairman and Chief Executive Officer Officers
I mean, I think it's really... I mean, to a large extent, it's the way I see my work: make it sustainable. I think it's really cyclical — I mean, I've used that word many times, and it's really corrosive. We are either in such a situation — it seems so, at the supply end, we are struggling to survive。
I was in that situation in '23. We started a company a year ago, and everyone gave us a valuation — which I felt very low at the time — which proved to be true later. Now that we are in the situation that we have just discussed, you have just asked me the kind of situation where no one can get what they want。
For me, that shows that our incentives are inconsistent. Our business model is not uniform. So I think what I'm doing, what our team is doing, if there are about three big things we need to do, is in this technical business — indeed, in any technical business。
THAT IS WHAT I AM CONSTANTLY TRYING TO BALANCE. FIRST OF ALL, THERE MUST BE A FAIR RETURN FOR WHAT YOU HAVE CREATED, RIGHT? WE ARE VERY PROUD OF OUR TECHNOLOGY, BUT THAT IS DIFFICULT TO ACHIEVE. NOT ONLY HAVE WE INVESTED HEAVILY IN THE CONSTRUCTION OF NAND, BUT ALSO IN THE CONSTRUCTION OF SYSTEMS. WE DON'T HAVE A SINGLE RESEARCH AND DEVELOPMENT TEAM, BUT WE HAVE TWO: ONE FOR NAND AND THE OTHER FOR SSD AND ALL OF ITS PRODUCTS — AND WE'RE DOING IT。
We must invest all capital in manufacturing. Backend? Yes, we do backends. We actually have a back end, so we do the whole process. Clearly, we have many excellent suppliers that provide us with many important technologies that enable us to do so. But above all, there is a fair return on the investments we have made, for which we have invested a long time。
So that's the first thing. I would say that we are doing well in that regard. Frankly, we have been doing poorly for a long time. I mean, when you go back a year ago, people basically tell us that you're not doing enough in this area because they don't want to invest in your companies。
So that's the first thing you have to do. The second thing that really concerns me is that we need to do something about periodicity, right? It's too bad because everyone is waiting for the recession to come. You have a good quarter. Oh, you're just a quarter closer than a quarter, this crazy psycho. People either do not have what they need or have too much. It doesn't help me at all。
That is why we want to do it, and we are doing it through business practice. That is why we call these new business models. How do we change the way we interact with clients? And then, in any technology enterprise, the third thing you have to do is grow, right
You understand the right economics. You strip off the cyclical factor or deal with cyclicality in different ways, and then you have to grow. And in every technology business I've managed in my career, the third is the hardest. Growth is hard, isn't it? I mean, especially for big, profitable businesses。
But we've taken care of that, right? We say we're going to grow a little bit higher, and people say, "Can you not grow faster?" I would say, first let me settle the first two objectives, then we can start talking about it. It is therefore very difficult to balance this equation。
If you start to change — you can talk about one of the three alone, but you have to discuss the three together, because if you start adjusting one, the other will change in the opposite direction. So we've been trying to balance the equation and focus on the three。
I'd say the hardest part is growth, which is a huge advantage for us, right? People always argue: Should you grow faster? Well, maybe we can grow faster, but that's at the expense of economic efficiency, which is not a good trade-off from a valuation point of view。
Should we focus more on economics, pursue this more and accept greater periodicity? That does not seem to be a good trade-off. So you have to do all three. We've been -- that's how we think. At least that's what I thought. That is what we are trying to balance. It's interesting。
In this environment, I believe, our operations are undergoing very significant changes that can really address the first two issues。
Mark Newman Bernstein Industrial Services LLC, Research Department
It's very helpful. Then we go further into the second point, the long-term agreement — the new business model you call it. Would you now like to comment on how these agreements are broadly based on continuity, quantitative commitments and pricing structures
If you could explain what you can do, like how you see those agreements and how you're doing。
I know you said at the last telephone conference that more than one third of the long-term agreements that you call new business models come from here. What do you expect this to be? It would be helpful if we could explain。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Yeah. So let's talk — you just mentioned a little bit, and I think we all understand. I mean, it's always been a very volatile business, right? In fact, prices change almost every quarter. This is an operation that is difficult to plan and unpredictable。
And there has traditionally been something about agreements — by the way, we have very remarkable clients. I mean -- that's one of the reasons this business is so attractive. Our clients are the most enviable company in the world. Indeed, they did very well, whether in personal computers, smartphones, data centres or any other field, and in general, our clients did incredible things。
But the conventional view of LTAs is that I will commit to supply and then we will discuss prices. That means it's better than nothing, right? At least we understand that if prices are agreed, we know how much to allocate to everyone。
But we want to get rid of this volatility. So how do we think about this differently? That's why we, the “long-term” idea, have put out a lot of terms in the industry: LTAs, NCNRs, take-or-pay, and all sorts of things。
And, in my experience over the past two years, as long as you mention any of those statements, the person who sits opposite you will immediately begin to tell you the reasons why they are not working. So we studied it in depth。
WE SAID, "LISTEN, WHAT WE WANT TO DO IS ALIGN OUR BUSINESS MODEL WITH THAT OF OUR CLIENTS, RIGHT? AND MORE AND MORE CLIENTS ARE COMING TO US, ESPECIALLY AFTER WE HAVE PASSED THESE DATA CENTRE ACCREDITATION. THINK ABOUT THE DATA CENTRE BUSINESS WE'VE BEEN DEVELOPING, WE'RE DOING ENTERPRISE-LEVEL SSD. IT WILL TAKE YEARS, AND IT WILL BE A VERY DIFFICULT PROCESS. IT MAY TAKE THEM TWO YEARS TO UNDERSTAND WHAT YOU'RE DOING, TO GIVE THEM SAMPLES, TO PUT THOUSANDS OF DEVICES IN THE LAB, TO LET THEM RUN FOR A YEAR TO COMPLETE CERTIFICATION, WHICH IS A VERY DIFFICULT PROCESS。
So at the end of that process, the client would say, well, you made a great product, right? We've invested a lot in this. We made great products. I want to buy it. Great, right? And I want to buy it for a long time. I want to buy it for the next five years, because mine -- back to what we said at the beginning. I'm doing all this, building this new technology。
I have a great need for your products. I don't just want to buy something this quarter. I want you to tell me that you can supply me for the next five years. So, well, then tell me what you need and then we'll talk. It's kind of new, isn't it? Because it's usually like this: I'll tell you what I need for the next 12 months, and then we'll talk about it four times a year。
IT'S LIKE, NO, NO, NO, NO. I DON'T WANT TO TALK ABOUT THESE 12 MONTHS. I NEED TO KNOW IF I CAN GET WHAT I NEED FROM YOU TWO, THREE, FOUR YEARS LATER. BECAUSE WHEN I WAS BUILDING, WHAT THEY WERE BUILDING WAS AMAZING, INCREDIBLE TECHNOLOGY. SO WE'RE GONNA START TALKING ABOUT HOW WE CAN ALIGN BUSINESS MODELS, RIGHT? YOU WANT TO CONSUME NAND, I WANT TO PRODUCE NAND。
NOW I'M PRODUCING NAND IN A WAY THAT'S PROBABLY A DIFFERENT BUSINESS MODEL THAN YOURS, RIGHT? I HAVE TO INVEST 10 YEARS AHEAD. I'M GOING TO BUILD A GIANT ROUND FACTORY THAT CAN BE SEEN FROM SPACE. IT'S HUGE. I HAVE TO PLAN MY CAPACITY YEARS AHEAD。
The good news is I've already done it. We have the round factory. We have research and development. We know what the technology road map is for the coming years. But now I'm done, and I'm gonna start work on the crystal mill, right? Now I'm investing in growth, right? I'm going to achieve a moderate to a dozen percentage points higher. So my crystal mill is running, and tomorrow there are more crystals than yesterday. Every day it's like this, every day it comes out of the crystal mill。
And I have to sell them. They cannot be placed in stockpiles or dropped on the ground, and someone has to take over. For ordinary consumers, this business model is a little unnatural. They have big, growing businesses, but, guys, do I have to buy things every month? Do you have to buy more than last month? The answer is yes。
So how do we agree — you will need these supplies, and I will provide them. How do we adapt our business model so that I am confident that you will be my strategic partners, and you are confident that I will deliver to you? How do we put this in the contract? That's why we came up with these new business models. So what do we think of this? First, we need partners that consume a lot of products, right? This would be a major contractual arrangement. Secondly, we need your demand to grow at least as fast as our supply, preferably faster than our supply。
So if I'm investing in a modest to a dozen percent growth rate, and you're telling me that four years in a row for the same amount, that doesn't really help me, does it? The rate of consumption you need must be faster than I can provide。
What you need to do next is have predictable consumption. Remember, the round factory is running every day and the round keeps coming out. If you are my strategic partner, you need to be predictable on demand weekly, um — or monthly, quarterly。
The more you make me understand that need, the better we are. What would your product mix look like? What is the share of this product and that product? We have to figure this out. And then there's an incentive, because you're a listed company and I'm a listed company. Something might happen to you to have to quit the contract。
I know, right? Things happen. When the black swan happens, the economy will rise and fall. For example, we are facing a global pandemic. Assuming an event that could affect the whole world. So at that moment, I needed an incentive to give you the incentive to continue with the contract。
If you don't fulfil your contract, I'll get paid, okay? So I want you to leave some money. We'll let the third party keep it for us, right? We won't -- I won't -- we won't argue. I won't sue you. This will never happen, will it
So... you're not gonna sue your client, are you? We're partners, right? If anything happens, you have to quit the contract. So the third party is supposed to trust a sum of money you have in advance — you can use the word if you like. In any case, a third party would be hosting a sum of money. But the simplest thing is that you give me all the money in advance。
That's kind of unrealistic, right? I mean, it's a five-year relationship, you don't -- it's a big check for anyone. That is unrealistic for various reasons. So we have to figure out another way. By allowing third parties to keep the money, they will hold the contract and will be able to judge whether or not you have breached the contract. In case of default, the third party will give me the money. That's gonna happen on my balance sheet, so we get together。
By that moment, the contract was over. I've kept everything since then, and you kept everything you paid for, and then we lived together, right? So we think this will match our incentives. Now you have the incentive to continue the contract. You might think, God, I'm gonna quit the contract. Do you really want to quit? You'll have to give up a sum of money, possibly billions. So you better think about it. If I had to quit, I could at least have a buffer. I get a cash deposit, which is very helpful in cases like the Black Swan incident. For example, I would like to eliminate cyclical fluctuations, assuming that there is a huge down cycle, and what does it take to get through it? Need cash。
So we can protect ourselves, both of us — we can move on, everything is fine, we can do business again sometime. So that's the general idea of the contractual structure that we have set。
Mark Newman Bernstein Industrial Services LLC, Research Division
Is that a $12 billion commitment
David V. Goeckeler Chairman and Chief Executive Officer Officers
Yeah. Let's split these numbers. We've discussed this. We have RPOs now. This is one you'll consider. I run a lot of software companies, right? So that's an indicator from over there. But that's an accounting indicator, right? We did not suddenly wake up and say, “Oh, we need to use this indicator”, but when there are contracts and future obligations in the industry, people do。
We signed three contracts at the end of the quarter. The figures that we have mentioned earlier, some $40 billion or $42 billion, represent the remaining procurement obligations, which represent the minimum amount of the procurement obligations of the three contracts for the entire period。
And then we talked about another number, a little different, so it's complicated. We signed two contracts after the end of the quarter, so they were not included in our data. However, of the five contracts that we have signed, a total of $11 billion was set aside to prevent withdrawal. Basically, that's how the two figures relate。
But honestly, we don't expect to see the money at all. I don't want to see the money either. I think we have good partners. I think these contracts will be completed. I think our interests are the same and everything will be fine。
But we live in the real world. There must be some kind of incentive, and I think we have aligned those incentives. I also believe that we have willing partners who are willing to join us in this path because they value supply commitments。
Mark Newman, Bernstein Industrial Services LLC, Research Department
You said there were five clients, those were super-moderate cloud service providers, or not
David V. Goeckeler Chairman and Chief Executive Officer Officers
No, we didn't say that. We wouldn't say that either. What we're looking for is -- you just did it perfectly. I mean, look, we want to diversify, right? What we want is the same as what we treat ... I have talked about the portfolio many times。
WE WANT A DIVERSIFIED PRODUCT MIX THAT HAS A GREAT DEAL OF CHOICE IN THE PRODUCTS WE SELL. THAT'S WHERE I STARTED. WHY DO I LIKE NAND MARKETS? I LIKE THE NAND MARKET FOR MANY REASONS -- IT'S A VERY DIVERSE MARKET. THERE'S A LOT OF GOOD CUSTOMERS YOU CAN SELL YOUR PRODUCTS IN A LOT OF PLACES. BUT YOU HAVE TO HAVE THE TECHNOLOGY TO DO IT. YOU CAN'T JUST SELL THE ORIGINAL ROUND. YOU HAVE TO MAKE THEM A PRODUCT. IF YOU WANT TO ENTER THE CONSUMER MARKET, YOU HAVE TO HAVE A TEAM TO BUILD THESE PRODUCTS。
You gotta have a back end in creating all those things. So you need to diversify the portfolio as much as possible in order to have as many options as possible. The same applies to these new business models. We'd like to have time-frames, right
You don't want them all due on the same day. So if you want some one-year periods, some three-year periods, some five-year periods, and then you need a diversified client, preferably covering as many portfolios as you can, because that keeps the portfolio alive and that is optional, we have taken the first step, right
That's what we announced at the newspaper conference. Again, back to the three things I mentioned: fair returns, procyclical responses and growth. Fair return. I think we're okay, but overall we're good. Now we have five in the middle box to start working on this. More than one third of the portfolio is visible, not only for three months or possibly for twelve months at a time。
Now we're talking about visibility — two, three, five — and it's very different. As for the component of growth, remember, growth has always been. That one was always chosen, right? This has been the most important and the reason why the market is so good: it will grow。
That's why we've got the growth tied up. The first was achieved, and we are in a good position. What we have to do now is keep it, and this is the second part. That is why we call it a new business model, because it is a different business model to achieve that。
Mark Newman Bernstein Research Services LLC, Research Department
That 33%, do you want to raise it to 50%, 60%, 70%, or do you think that's unrealistic
David V. Goeckeler Chairman and Chief Executive Officer Officers
No, that's not unrealistic. Still to be determined, right? Again, we're not done yet. We took the first step — perhaps the first five — perhaps more appropriately. But we are still discussing, and that depends on the portfolio I mentioned earlier。
See, as I said, we have a lot of -- we have very great clients, really great clients. They're great companies, great people. They develop incredible technology. Some of them like our old business model. They like to come quarterly, hey, we'll -- just do it quarterly. Okay, okay. We accept that. We're totally fine. We know what to do. If that's what they want, we'll do our best. So we'll see..
Mark Newman Bernstein Industrial Services LLC, Research Department
But if those clients are not registered ... will they be provided with sufficient supplies
David V. Goeckeler Chairman and Chief Executive Officer Officers
I can't run business for them. They have to run their own business, right? I'm not the only supplier on the market. But what I want to do is to get a set of such agreements that bring me the kind of diversity that I have described and that can cover a significant part of my portfolio. Not all coverage is required. It will never be possible to reach 100 per cent because many of the clients outside are too small to be able to engage in such cooperation, although they are good customers and valuable businesses。
We'll see. I think it's a little uncertain, and maybe it's a little off for you now. The final destination has yet to be determined. But I think we'll keep pushing it up if we get the chance。
Mark Newman Bernstein Industrial Services LLC, Research Department
Great. I mean, this is very clear, much clearer than what we have so far obtained from your competitors in the LTA. So I thank you very much for that。
David V. Goeckeler Chairman and Chief Executive Officer Officers
I like mine. Those companies are great. Indeed。
Mark Newman Bernstein Industrial Services LLC, Research Department
By the way, the supply and capacity issues are very strong, given current prices and demand. Many of your competitors, Samsung, Hynix, Micron, don't have the space to increase their capacity because they use all of the mills for DRAM, which is also very tight。
But Sandisk Kioxia, together, you're one of the few companies that actually can increase capacity. I'm not encouraging you to do that. I'm just asking. I was just asking。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Are you asking for friends? Is that so
Mark Newman, Bernstein Industrial Services LLC, Research Department
HOW DO YOU SEE THIS? WHAT I'D LIKE TO ASK IS, CONSIDERING THAT YOU HAVE THE POTENTIAL TO INCREASE PRODUCTION CAPACITY, AND THAT FEW OTHER PEOPLE CAN DO THAT (WITH THE POSSIBLE EXCEPTION OF YMTCS IN CHINA), CAN YOU DO MORE? OR ARE YOU JUST TRYING TO OPTIMIZE PRICING NOW
David V. Goeckeler Chairman and Chief Executive Officer Officers
Okay. Could be a little more complicated than you think. First, we've been increasing capacity. I think we should start with that. We have been expanding — remember, we are growing at a rate of about a dozen percentage points higher. The market is very large, and the turnover is also growing at a rate of 10 or more percentage points. First of all, it's amazing, right
Second point. So under normal conditions, we'll increase capacity. That's it. That's the business we decided to enter. We must make decisions well in advance. For example, next quarter's demand has no impact on my capacity decisions。
I had to make that decision three years ago. Our Cyclops plan is for years to come. It is very complicated to move equipment between nodes — in a crystal mill you do not run only one node, but many and many different nodes at the same time, and the transition is very complex。
So you have to make those decisions well in advance. And I don't want to get too involved with this, but back up, about 12 months ago, everybody told me we were investing too much, right? And we said, no, no, no. We believe that the medium to high growth rate is the correct figure, but some say, well, that is not the correct figure by price。
Mark Newman Bernstein Industrial Services LLC, Research Department
Less than 12 months ago。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Yeah, less than 12 months ago, right? So we can't take that big swings. What do I think? My idea is that we are investing in a growth rate of up to 10 percentage points, and that we have excellent partners in Kioxia and very good relationships。
This has been going on for a long time and for a good reason, because it is very productive and very valuable. We're good at planning. We are good at planning and ensuring that, at the right time, we have everything necessary to continue to develop our business。
There is one thing I would like to say now that is very important and must be understood. We can achieve growth through a change of pace. That's what I call R & D productivity. The bits of each crystal round continue to rise at a faster rate of compound growth, faster than what I call the mid-to-twenty-percent higher rate。
So if we move on a single node with an equal number of crystal circles, we create oversupply. So we keep adjusting the equation. Remember that more room space is required for each node of the procedure. Each node is more complex, with more steps, more equipment and more clean room space。
But this dynamic is extremely important, isn't it — if you look at the proportion of our capital expenditure in income, it continues to decline as income rises, because our R&D productivity remains high. So this is very important to all the investors here。
It means that I can achieve growth without devoting significant additional capital expenditures. And back to my old model -- yes, economics eliminate cyclicality and expand what you eventually find. After all, what are we doing this for? This is a free cash flow。
You will find that this business is very good at this because we invest very efficiently in capital expenditures and are able to obtain more incremental outputs from them。
Mark Newman, Bernstein Industrial Services LLC, Research Department
I GOT SOME QUESTIONS FROM THE AUDIENCE. CAN I ASK ONE MORE QUESTION BEFORE I ANSWER THE AUDIENCE? HBF, HIGH BANDWIDTH FLASH. WHAT'S THE LATEST ON IT
David V. Goeckeler Chairman and Chief Executive Officer Officers
WE HAVE BEEN VERY EXCITED ABOUT THIS TECHNOLOGY SINCE WE FOUNDED IT LAST FEBRUARY AND ANNOUNCED IT. WE HAVE LONG BELIEVED THAT NAND WOULD BE A VERY IMPORTANT TECHNOLOGY ONCE THE REASONING PHASE HAD ENTERED。
YOU DON'T HAVE TO CONVINCE US THAT MEMORY STRUCTURES NEED TO BE CHANGED TO MAKE REASONING SCALABLE. THIS IS ESSENTIALLY PART OF THE HBF. THIS DOES NOT MEAN THAT HBF WILL REPLACE ENTERPRISE-LEVEL SSD, NOR DOES IT MEAN THAT HBF WILL BECOME AN ALTERNATIVE TO DRAM AND SO ON。
This shows that innovation has great opportunities as the scope of reasoning expands. For those with new ideas, for example, I now see artificial intelligence and the massive expansion that is under way, which I see as a huge green light for innovation. If you have any new ideas, take them out, right
Because the world is looking at how to scale this amazing technology. And the sooner we do that — as I said before — the more we can scale this technology in a way that is completely without friction。
IT IS SURPRISING HOW FAST TECHNOLOGY CAN BE USED BY ALL IF IT IS ECONOMICALLY SOUND. HBF IS A STRATEGY DESIGNED TO PROVIDE HIGH DENSITY TO REASONING, WHICH IS PRIMARILY A READ-ORIENTED AND DEFINITIVE READING ACTIVITY. SO WE'RE VERY EXCITED ABOUT THIS TECHNOLOGY. IT'S BRAND NEW. WE'RE MAKING NAND CRYSTAL PARTICLES NOW, AND THEY'RE EXPECTED BY THE END OF THIS YEAR. SOMETIME NEXT YEAR, WE'LL HAVE THE WHOLE SYSTEM. WE'RE BUILDING THE CONTROLLER ON IT, AND WE HAVE A LOT OF WORK TO DO. WE'RE WORKING WITH CLIENTS ON HOW THEY CAN INTEGRATE THEM INTO THEIR OWN STRUCTURES, RIGHT? BECAUSE IT'S NOT PLUGGED IN。
It's not the simple substitute that we plug components in, you pull something out. You have to-- this is a systematic programme. So you have to let your clients adopt it into what they're building, and we're going through it. We will keep updating as we move forward。
Mark Newman, Bernstein Industrial Services LLC, Research Department
Okay, sounds great. Now, the question from the audience is, first of all, whether the shift to the edge or to the end of the device would be beneficial or risky to Sandisk ' s growth projections
David V. Goeckeler Chairman and Chief Executive Officer Officers
No, I think -- I mean, anywhere you've changed, you say artificial intelligence..
Mark Newman, Bernstein Industrial Services LLC, Research Department
So, AI basically, AI with Edge means it's on your equipment, like your smartphone and your PC。
David V. Goeckeler Chairman and Chief Executive Officer Officers
OKAY. I MEAN, IT'S JUST THE SAME SUBJECT: NAND IS EVERYWHERE. AS YOU START TO SCALE UP, YOU'LL FIND THAT YOU NEED MORE CAPACITY AND THAT WE HAVE SCALABLE TECHNOLOGY。
We see this as very important — and that is why, even last year, we insisted on a growth rate of up to a dozen percentage points, because this market has a permanent nature。
It is a wonderful thing to think of new ways to use our technology, just as the world has been constantly changing。
Mark Newman Bernstein Industrial Services LLC, Research Division
I have a good question from the audience: may I raise the hand of the audience, if you hold the Sandisk stock? Can I ask you to raise your hands if you hold the Sandisk stock
Okay. Okay。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Thank you. We are working very hard for you。
Mark Newman Bernstein Industrial Services LLC, Research Department
Here's the next question from the audience: Can you build on the lessons learned from past booms and busts? In order to avoid a repetition of the past, have management incentives changed throughout the industry, similar to those in the oil and gas sector
David V. Goeckeler Chairman and Chief Executive Officer Officers
I don't know much about oil and gas. But what I'm saying is -- listen, I've been in the technology industry longer than I should have been, about 40 years, building a global technology. I started my career in Bell Labs a long time ago. And I run many different technical lines of business, many different ones covering hardware, software, and SaaS, which are very large。
WHEN I ACTUALLY GOT INTO THIS BUSINESS AS CEO, I WAS A LITTLE SURPRISED HOW IT WORKED. AS YOU SAID, THIS WHOLE CYCLE OF “PROSPERITY AND DEPRESSION” — AND I HAVE SAID IT MANY TIMES — I THINK IT'S HARMFUL, AND PEOPLE ALWAYS FEEL THAT THEY DON'T GET WHAT THEY WANT: EITHER THE SUPPLIERS STRUGGLE TO SURVIVE, AS I DID FOR 23 YEARS; OR WE ARE IN A SITUATION WHERE PEOPLE SAY, “I CAN'T GET WHAT I NEED”. I THINK THIS IS BECAUSE OF THE WAY WE DO THINGS. AS I SAID BEFORE, THERE ARE SOME REASONS, RIGHT
We have to stretch the investment cycle, and the increase in supply is more like a leap function, and demand is more like a curve. So it is not easy to align these factors. I don't think this is something we should give up. And I don't think it's meant to be. I do not think that the mere fact that this has been the case for so long means that the future must be the same. Maybe I'm a little arrogant, but that's what we do. We are innovators. What we mean is that we invent new things that can also be applied to business models。
That is basically what we do with money, and I think the world has done well in that regard. I think if we think about this business model, yes, what have we learned from the boom and the bust? Don't do that again. Really, it would be great if we stopped doing that。
SO HOW DO WE STOP DOING THAT? WHAT ABOUT THE CRASH? HOW CAN WE ACHIEVE A FAIR RETURN ON THE TECHNOLOGY WE HAVE DEVELOPED? IT'S REALLY HARD. TECHNOLOGY IS COMPLICATED AND DIFFICULT. THIS IS A 3D SEMICONDUCTOR TECHNOLOGY FOR WHICH PEOPLE HAVE DEDICATED THEIR LIVES。
The cost of doing so is very high. It requires huge capital expenditures, and it is very difficult to build and operate a crystal mill. Let us get a fair return for this and establish a business model that will make it all smoother. I think it's totally possible. I also believe that we have taken a few steps along that path and will continue to do so。
Mark Newman, Bernstein Industrial Services LLC, Research Department
Great. Okay, time's up. Thank you very much, David。
David V. Goeckeler Chairman and Chief Executive Officer Officers
Thank you。
