Pure Storage Inc (PSTG) 2025 Q2 法說會逐字稿

(audio in progress) to the Pure Storage second-quarter fiscal 2025 financial results conference call.

Today's conference is being recorded.

(Operator Instructions)

At this time, I'd like to turn the call over to Paul Ziots, Vice President of Investor Relations.

Please go ahead.

Thank you.

Good afternoon, everyone, and welcome to Pure's second-quarter fiscal year 2025 earnings conference call.

On the call, we have Charlie Giancarlo, Chief Executive Officer; Kevan Krysler, Chief Financial Officer; and Rob Lee, Chief Technology Officer.

Following Charlie's and Kevan's prepared remarks, we will take questions.

Our press release was issued after close of market and is posted on our website where this call is being simultaneously webcast.

The slides that accompany this webcast can be downloaded at investor.purestorage.com.

On this call today, we will make forward-looking statements, which are subject to various risks and uncertainties.

These include statements regarding our financial outlook and operations, our strategy, technology and its advantages, our current and new product offerings, and competitive industry and economic trends.

Any forward-looking statements that we make are based on facts and assumptions as of today, and we undertake no obligation to update them.

Our actual results may differ materially from the results forecasted and reported results should not be considered as an indication of future performance.

A discussion of some of the risks and uncertainties related to our business is contained in our filings with the SEC, and we refer you to these public filings during this call.

All financial metrics and associated growth rates are non-GAAP measures other than revenue remaining performance obligations, or RPO, and cash and investments.

Reconciliations to the most directly comparable GAAP measures are provided in our earnings press release and slides.

This call is being broadcast live on the Pure Storage Investor Relations website and is being recorded for playback purposes.

An archive of the webcast will be available on the IR website, and is the property of Pure Storage.

Our third-quarter fiscal 2025 quiet period begins at the close of business, Friday, October 18, 2024.

With that, I'll turn it over to Charlie.

Thank you, Paul.

Good afternoon, everyone, and welcome to our Q2 FY25 earnings call.

We were pleased with our Q2 revenue growth of 11% year over year.

Pure continues to pick up market share and outpace the industry, both in innovation and in growth.

During the quarter, we hosted customers and partners at our Annual //Accelerate Conference.

Our Las Vegas //Accelerate held in June, picked off a series of local events across the Americas, Europe, and Asia Pacific, where thousands of customers and partners learn about our platform strategy and Fusion vision.

As well as Pure's offerings in AI, virtualization, and application modernization.

Specifically, we introduced the next generation of Fusion, a first-of-its-kind storage cloud architecture soon to be available as a non-disruptive upgrade to all of our global customers.

Fusion allows businesses to transform their Pure Storage systems into an automated data storage cloud that eliminates the data silos of existing enterprise data storage systems.

We also unveiled the industry's first AI Storage as a Service for GPU clouds.

Growing GPU and AI clouds need flexibility in their infrastructure, as they are uncertain of their future growth and the type of workloads that they will need to address.

Pure's Evergreen//One for AI and provides them flexibility in both their consumption and price performance needs and matches cost to their revenue growth.

Our Evergreen//One service offering remains strong.

Evergreen//One removes the hard work, expense, and risk of operating a storage environment from enterprise IT organizations.

It provides flexibility, it avoids overprovisioning and rigid planning, and it simplifies customers' operations with solid and guaranteed SLAs.

It also significantly boosts efficiency in terms of capital costs, energy, and labor.

Options Technology, a financial technology company, started with one small Evergreen//One subscription back in 2019 and has grown over the last five years to 18 petabytes of storage across multiple sites globally.

Through Evergreen//One, Pure regularly enhances the delivery of their SaaS services, improving resiliency, efficiency, and overall performance.

While the lengthening of large enterprise deal times impacted Evergreen//One growth in the first half, we continue to see strong deal activity.

Artificial intelligence continues to be of great interest to our customers.

Specifically, customers continue to study both the potential areas for AI use, as well as how to accommodate AI in their infrastructure.

We were pleased to have had NVIDIA join us at //Accelerate to announce our expected NVIDIA DGX SuperPOD certification by the end of this year.

The AI market for data storage has progressed, as we have consistently predicted.

Pure sees three separate AI opportunities for our solutions.

First, storage for machine learning and training environments, where Pure provides high-performance storage for public and private GPU farms.

This quarter, we signed a deal with SoftBank Corporation, one of the big four telecommunication services in Japan.

Pure is providing the storage layer behind many of SoftBank's cutting-edge services, including their new generative AI platform, created specifically to develop the market-leading large language model for the Japanese language.

The second AI opportunity we foresee focuses on tailored storage for enterprise inference or RAG environments.

Many, if not most, enterprises will use commercial LLMs, or other models, to operate on their own proprietary data in-house.

These systems will use relatively small GPU environments to provide AI insight from their data.

Pure is working closely with NVIDIA on a number of vertical market offerings to satisfy this market.

We continue to believe that our largest opportunity opened by AI is to address the siloed nature of Enterprise's existing data storage architectures.

Current data stores sit behind application stacks, and generally have neither the performance nor the connectivity has served data directly for AI engines and analytics.

Customers that are the most advanced in their AI investigations all acknowledge that data access and preparation are major barriers to AI deployment.

Pure Fusion will allow customers to upgrade their enterprise storage to function as a storage cloud, simplifying data access and management and eliminating data silos to enable easier access for AI.

The focus and uncertainty around AI has caused customers to begin to reevaluate their planning on how they will invest their IT dollars.

We're also seeing large organizations increase their focus on managing escalating costs from software, cloud, and SaaS services.

Our Pure Cloud Block Store, or Microsoft Azure VMware Solution, is helping enterprises contain cloud storage costs generally by more than 50%.

But simply, Cloud Block Store provides a more resilient and performing public cloud storage infrastructure for large enterprise cloud application deployments that is dramatically less expensive than cloud native services.

Furthermore, Cloud Block Store is fully compatible with enterprise storage interfaces and services, including disaster recovery and data protection.

One case in point is a Fortune Global 500 food and beverage customer that face a growing hyperscaler data footprint and accelerating costs with limited visibility into its overall workload performance.

By leveraging Cloud Block Store, reducing thousands of cloud-managed discs to just dozens of Cloud Block store volumes, equipped with data protection, ransomware remediation features and advanced workload performance reporting.

The company is looking to save 50% of its total storage bill.

Our discussions with hyperscalers to replace their core storage with Pure technology continues to progress positively.

Our lead prospect has advanced from extensive evaluation of our core technology to testing and integrated solution and we have been engaged in detailed contractual negotiations for many months.

We remain confident that we will secure our first hyperscaler design win by year-end.

The longer-term opportunity for Pure with hyperscalers is significant.

To provide a sense of scale, the top 10 hyperscalers are projected to buy almost 70% of all disc drives over 600 exabytes this year alone.

Because of Pure's unique direct to flash technology, we can offer hyperscalers better performance, reliability, and power and space savings in hard discs at a similar or better total cost of ownership.

With nearly 15 years of experience with software and hardware flash management, we continue to far outdistance the industry in energy efficiency, density and performance.

Pure holds key intellectual property and unmatched multivendor, multi-process flash expertise that no other vendor can match and cannot be replicated with standard SSDs.

Our latest 150-terabyte DirectFlash module shipping later this year is, but the next stop on our robust industry-leading flash road map.

Energy and space savings generated by our direct-to-flash advantage are significant.

We reduced space, power, and cooling requirements by a factor of 5 to 10 compared to hard discs.

In a world of greater power demands and limited electrical supply, the savings on electricity alone provides a compelling incentive to switch from hard discs in both hyperscaler, as well as enterprise data centers.

Our //E family of products focused on replacing enterprise hard disc systems with more efficient and higher performance Pure technology continues to grow strongly.

Enterprises increasingly recognize that Pure DirectFlash technology has reached the price level where they can eliminate the last mechanical component from their data centers.

As highlighted in our latest ESG report, power reduction on storage from Pure's DirectFlash technology can reduce total power usage in existing data centers by approximately 20%.

Businesses are facing higher energy costs and greater power constraints, while committing to higher sustainability goals.

BT, the British multinational service provider, has set a target to achieve net zero carbon emissions in its operations by the end of March 2031.

As a foundational storage provider to BT, we directly support their data center energy reduction program.

We have enabled BT to grow its data storage, while reducing its energy usage.

BT has measured your storage to be about 18 times more efficient than their legacy storage benchmarks.

Looking back over this past quarter, we have not seen a significant change in the overall macro environment or our customers' intentions to buy.

We have, however, seen customers look to manage increasing costs in cloud, software, and SaaS.

We believe that the storage market will be resilient in this IT economy but we have yet to see a positive inflection.

Overall, we are well positioned in all of the segments in which we compete and believe we will continue to gain share in our market.

We know we are gaining ground, as our growing strength has forced competitors to intensify their efforts and mimic our messaging.

It is clear now that legacy competitors in our markets see Pure as the alpha competitor, and have focused their messaging and strategies on us.

We appreciate the attention and look forward to the competition.

We remain confident in our ability to expand our market share and maintain our strong leadership in storage.

With that, I'll turn it over to Kevan.

Thank you, Charlie.

We are pleased to have delivered double-digit revenue growth during the first half of our fiscal year, and we continue to see strong sales performance for both our FlashArray//E and FlashBlade//E offerings.

Revenue of $764 million in Q2 grew 11% year over year and both revenue and operating profit of $139 million exceeded our guidance.

Subscription services, annual recurring revenue, or ARR, grew 24% to over $1.5 billion, which continues to be driven by our Evergreen//One service offering, in particular for a higher velocity business.

As a reminder, subscription services ARR excludes non-cancelable Evergreen subscription contracts where the effective service date has not started.

Including non-cancelable subscription contracts where the effective service date has not started, subscription services ARR grew 25%.

Total RPO and exiting Q2, which includes both subscription services and product orders grew 24% year over year to $2.3 billion.

As we have shared in previous quarters, product orders within RPO include a noncancelable telco order from Q3 FY24 and orders relating to a Fortune 500 financial services company from Q4 FY24.

At the end of Q2, RPO associated exclusively with our subscription service offerings grew by 21%.

Additionally, total contract value sales for Storage as a Service offerings during Q2 reached $101 million, bringing TCV sales in the first half of FY25 to $157 million.

Our Evergreen//One as a Service business is strong, demonstrating robust pipeline growth and consistent success in converting opportunities valued at $5 million or less.

This continues to underpin our confidence in the growth potential of our Storage as a Service offerings.

Consistent with last quarter, we continued to experience extended closing time lines for larger Evergreen//One opportunities.

Last year, we closed several large Evergreen//One deals in the first half compared to [three] in the first half of this year.

This impacts both year-over-year RPO growth and forecasted FY25 TCV sales for our Storage as a Service offerings, which we now expect to be $500 million, reflecting a growth rate of approximately 25%.

US revenue for Q2 was $538 million and international revenue was $226 million.

Our new customer acquisition grew by 261 customers during Q2, and now we serve 62% of the Fortune 500.

Total gross margins of 72.8% in Q2 continues to be very healthy and comparable year over year.

Subscription services gross margin strengthened to 76.4%, as we leverage increased automation of our service logistics workflows, supporting delivery of our Evergreen subscription services.

Our product gross margin of 69.5% in Q2 and underscores the strong sales growth of our FlashBlade//E, FlashArray//E and FlashArray//C solutions, driven by customers increasingly shifting their cost-sensitive workloads to all-flash.

As we aggressively pursue our efforts to help customers transition their workloads to our all-flash solutions, we anticipate a modest strategic decline in product gross margins during the second half of the fiscal year.

Operating profit and margin strength of approximately 18% were positively impacted by revenue overachievement, strong gross margin performance, and operating expense discipline.

Our head count increased sequentially by nearly 250 to approximately 5,700 employees at the end of the quarter.

Here's balance sheet and liquidity remains very strong, including $1.8 billion in cash and investments at the end of Q2.

Cash flow from operations during the quarter was $227 million, and capital expenditures were $60 million.

Our most significant capital expenditures during the quarter were concentrated in engineering for new test equipment supporting key strategic growth initiatives, including our pursuit of hyperscaler infrastructure opportunities.

As part of our objective of partially offsetting dilution, we began paying withholding taxes due on employee equity awards.

In Q2, withholding taxes on equity awards was $76 million, which offsets dilution by approximately 1.1 million shares.

We have approximately $395 million remaining on our existing repurchase authorization.

Now turning to guidance.

For Q3, we anticipate revenue of $815 million, with an expected operating profit of $140 million, resulting in an operating margin of 17.2%.

Projected operating profit takes into account in modest sequential decline in product gross margins that we expect during the second half of the fiscal year, driven by our expectations of continued sales growth of our //E family solutions, which are successfully targeting cost-sensitive workloads.

Turning to our annual guidance for FY25, we are reaffirming our FY25 revenue target of $3.1 billion, representing growth of 10.5%, and our operating profit guidance of $532 million, with an operating margin of 17%.

The anticipated modest decline in product gross margins during the second half of the fiscal year validates our successful strategy of expanding into cost-sensitive workloads with our all-flash solutions and has been contemplated in our annual guidance.

In closing, we are pleased to deliver strong financial performance, which reaffirms the effectiveness of our strategic initiatives.

Our focus on innovation and customer-centric solutions underscores our commitment to be a leader in the data storage industry.

While we remain mindful of the broader macroeconomic environment, we are confident in our ability to capitalize on the growing demand for high-performance, sustainable data storage solutions.

With that, I will turn it back to Paul for Q&A.

Thanks, Kevan.

(Event Instructions) Operator, let's get started.

(Operator Instructions) Amit Daryanani, Evercore ISI.

Good afternoon, everyone.

I guess my question really is, and I think one of the things I think folks are trying to square away is the Evergreen//One TCV target is getting lower from $600 million to $500 million.

I would have thought that a downtick here would have meant perhaps better CapEx trends for your customers, such that it would actually help your fiscal year revenue growth profile.

Clearly, you're not seeing that happen based on the fiscal year guide.

So hoping you just unpack what's driving the downtick on the TCV expectations?

And how do you see that flowing to your revenue guide really?

Thank you.

Absolutely a bit.

Well, I think your supposition would be correct if customers, the same customer, was switching from an Evergreen//One deal to a CapEx deal.

What we've seen instead is large opportunities, large Evergreen//One opportunity staying opportunities longer than we expected, and therefore, stretching out.

A little bit whether that's based on caution or whether that's based on by the customer or whether that's based on other -- we had indicated that customers are looking very closely at their subscription expenses now given increases in software and SaaS expenses that were raised over the year.

We've yet to really fully diagnose that.

But what we're seeing is a lengthening of large deal size Evergreen//One opportunities.

And I'll just add on to that.

We did see three opportunities closed this quarter that were larger.

And we're defining larger as greater than $5 million.

And just to reiterate, the question was really focused on why aren't we seeing an increase to our annual revenue guide.

And I will point out, to Charlie's point, these larger deals that are Evergreen//One are still being actively worked and they're just taking longer to close.

If we saw those opportunities flip to a traditional product sale or CapEx, that's when we would be looking at an upward view of our annual guidance for revenue.

Thank you, Amit.

Next question, please.

Aaron Rakers, Wells Fargo.

Hi, guys.

Thank you.

This is Michael Senoff on behalf of Aaron.

I just wanted to see if I can get any more color just on the hyperscale opportunity you guys kind of mentioned.

It sounds like you're continuing discussions with the customer you expect by the end of the year.

I'm curious if anything has changed or maybe ask it another way, like what's the biggest hurdle you kind of need to overcome to get the deal done?

And then separately, just on other opportunities you're pursuing, if there's any kind of anything to note there for progress?

You bet.

So the lead horse, I wouldn't say there's the largest hurdle.

There's just lots of little hurdles.

A lot of that is just aligning business models, economic improvements to them to pricing and economics for us.

And there's a lot of logistical elements that go into this as well, especially when you're speaking with large hyperscalers, with large orders and large data centers and complex -- their own complex supply chains.

So I would say just lots of little hurdles right now, testing is going well, conversations are going well, but a lot of detail that has to be worked out.

Yes, this is Rob.

Just to add to that.

I would say, overall, our engagements with our lead prospects are progressing very well, as Charlie mentioned in his prepared remarks.

What we've done over the last many months is really moved forward on our testing in phases from initial proof of concept to testing of that core IP to now extensive testing of really an integrated -- think of it as a co-engineered solution.

And as you'd imagine, this involves detailed performance, operational testing, so on and so forth.

And with that, as Charlie mentioned as well, accompanied with a detailed contractual discussions around the commercial package.

And so overall, engagement goes well.

And to the original question, I would say it's lots of little things as opposed to one big hurdle.

Thank you, Michael.

Next question, please.

Howard Ma, Guggenheim Securities.

Great.

Thank you, and good afternoon, everyone.

My question is can you tell us who is your lead horses?

I'm joking.

That's not my question.

(laughter)

The question is it's a variation of the question that Amit asked by lowering your as-a-service TCV sales estimate, but keeping the total revenue outlook and change, that obviously means you're getting less contribution from product sales.

But Charlie, you just said that lower as-a-service sales is not because more customers were opting it by CapEx instead.

So does that mean, if we look at the product line, does that mean there's increased risk in the product line?

Or can you point to certain demand drivers that give you more confidence that, that guidance is appropriately set?

Thanks, guys.

Yeah.

As we look at our sales and the pipeline, et cetera, we're seeing CapEx sales continue as we had expected.

We're seeing the base, if you will, what we call the velocity sales of Evergreen//One progress as expected.

But we've specifically seen deals that we've been tracking all along, large deals for Evergreen//One, just length -- just not coming in when we expected and lengthening out.

They haven't changed in character.

In other words, they have not switched those same accounts, those same opportunities, haven't switched from an Evergreen//One intention to a CapEx intention, but they've also not closed.

So that's specifically what we're seeing.

So we're seeing good growth in the velocity business.

And so the view there is that Evergreen//One continues to be of great interest.

And the activity is good, it's just that the larger deals taking longer to close.

Yeah.

I think it's really important, this is Kevan, to decouple what we're seeing with our larger Evergreen//One deals.

And the longer period of time from the demand overall that we're seeing, which, look, we're not seeing any significant change in demand that informs our annual guide, which is, I think, what you're asking as well, Howard.

Thank you, Howard.

Next question, please.

Pinjalim Bora, JPMorgan.

Great.

This is Jason on for Pinjalim.

How has the customer interest been on the new Fusion offering?

And do you see customers looking at it to unify their storage space as they prepare for AI imprinting?

Yeah.

This is Rob.

I'll take that one.

Look, I mean, early interest has been great.

As you know, we have been out talking to customers, partners alike about the Fusion vision for the last year, I think what we're really looking forward to and really what we've seen coming off of our //Accelerate user conference is the demand.

And I think the ability for customers in the existing installed base with existing arrays to go and take advantage of that fusion technology later this year.

With our latest release of Fusion coming later this year, we'll be able to deliver all of these capabilities that Fusion is designed for being able to fully automate the management multiple environments allow customers to manage their Pure Storage estate through policy declaration as opposed to individual operational steps and really step back and unify that as one pool of resources, one storage cloud, if you will.

Later this year, customers will be able to take advantage of all these capabilities on all of their existing arrays and data storage states.

So early interest has been great.

I think as we roll into the later -- the release coming out later this year, it will make it that much easier for customers to take advantage of those capabilities.

Yeah.

I'll also say that the beta users that have been involved in this have been very pleased.

And we've been very pleased by the fact that we see a strong interest and utility infusion by both large customers and small.

So even small customers are seeing a great benefit in being able to manage even their relatively smaller fleets through policy and further reduce the amount of labor that it takes to operate at almost every level.

Thank you, Jason.

Next question, please.

Mike Cikos, Needham.

Hey.

Thanks for taking the question, guys.

I just want to circle up because it seems like customers are continuing to choose maybe more of a CapEx type purchase, which is a little bit counterintuitive in this environment given the macro.

And I'm just trying to get a better understanding on the customer preference.

Does it tie in any way to potentially customers thinking about data repatriation to their on-prem environments to help handle some of the ballooning costs behind these gen AI workloads as they move into production environments?

Yeah.

I'll start this, and then have Charlie come off, it's Kevan.

Look, what we're seeing with our Evergreen//One demand is actually quite strong.

And we're really seeing that in our velocity opportunity.

So these are opportunities that are less than $5 million.

And they are tracking strongly and tracking with our expectations that we set at the beginning of the year.

So I really think the dynamic comes back to these larger Evergreen//One arrangements.

And then there's probably a few dynamics that are probably a play there but it's less about demand signals of preference to an as-a-service offering versus a traditional offering or CapEx.

And look, customers are continuing to evaluate managing their cost in the cloud, software and SaaS.

And I think when customers are evaluating larger as-a-service offerings, that certainly could be part of the consideration as well and why we're seeing more time being taken as that evaluation is taking place.

Charlie, any other thoughts on that?

Yeah.

No, Kevan, I think you answered it well.

Thanks.

Thank you, Mike.

Next question, please.

Jason Ader, William Blair.

Yeah, thank you.

Hey, guys.

I just wanted to ask about the competitive environment.

I know we've seen an increase in QLC-based arrays from some of your competitors.

Can you just comment on how that has impacted or not some of the deals and some of the opportunities?

And then also, can you comment on the NAND pricing environment right now, which I know has been rising?

Yeah.

Let me start on that.

I would say that our lead in QLC remains as strong as it ever has been.

So I wouldn't say -- and that's, I think, identified by the fact that our //E family of products that is really targeting the low end has been very, very successful.

So I wouldn't say that QLC has changed the competitive environment very much at all.

I would say that competition is as tough as it's ever been.

I think we are -- clearly, everybody's top competitor right now that's in this market.

We really do feel that the sites are squarely on us and we are competing appropriately.

But I don't think it's a QLC activity at all.

I think it's a size and scale and effectiveness situation with respect to us.

I mean, if you look at our and our progress now, we're now the number two vendor of all flash systems into the enterprise, firmly in that spot and only a few points behind the number one in that area.

And I think that -- I think everybody is feeling it at the moment.

So competition is tough, but our lead in QLC remains.

And then I'll just touch on the NAND pricing, which is consistent with our previous commentary.

And flash pricing from our lens really is affecting top line.

And the volatility again, is highlighting the differentiated advantages we're seeing with our Purity software and our DirectFlash technology, and that becomes more evident when we see the volatility in NAND pricing, such as the current environment.

And look, we're having a lot of success in winning workloads across price-sensitive workflows for our customers, and that would be our //E family as well as C. And that will have an impact on gross margins as well.

Thank you, Jason.

Next question, please.

Asiya Merchant, Citigroup.

Great.

Thank you for the question.

The Evergreen ramp expected in the second half, maybe you can just begin dive back into some confidence that you have recognizing that these deals are taking longer to close, given macro and other dynamics that you talked about.

So what gives you the confidence that we ramp from $157 million here in the first half to $500 million for the full year?

Thank you.

Yeah, I appreciate the question.

This is Kevan.

Look, we certainly believe that the adjusted forecast for TCV sales of our as-a-service offerings at $500 million growing 25% is achievable and considers the dynamic that we're seeing with larger deals.

Our current forecast of TCV sales does assume less contribution throughout the year from larger deals, but also assumes at the continued higher velocity business continuing to track strongly similar to what we've seen in the first half.

And so the implied second half ramp of TCV sales is only slightly higher than what we've typically seen from our traditional seasonality with our CapEx sales or traditional sales.

So look, there's obviously work to do and we need to execute in the back half, but we absolutely expect to achieve this forecast.

Thank you, Asiya.

Next question, please.

Jim Fish, Piper Sandler.

Hey, guys.

This is Quinton on for Jim Fish.

Thanks for taking our question.

I understand it's still pretty early, but as you think about the backlog or your order book for the 150-terabyte flash module how is that compared to kind of what you were seeing from the last upgrade cycle at this point in time?

And I know it's a little bit apples and oranges here, but it seems like we're facing scrutiny in transformational budgets you're seeing in Q1.

Is there any concern that, that would impact customers' willingness or ability to kind of transform and move to this new module?

Or are people sell focused on performance that's not really in your kind of concern list here?

Yes.

Thanks, Quinton.

I think it's an interesting question.

I don't think it would be very much in our concern list for the following reason.

What the 150 really does is open up new opportunity for us at lower price points and to reduce our cost, to reduce Pure's cost at similar price points for existing E series transaction.

So to put it another way, we're satisfying a particular performance point with 75 today that we could also satisfy with 150s, the cost to us will be less even if the price is the same.

So I think it's a margin enhancement for us as well as allowing us then to get into even less expensive hard disc environments.

And so I think we see it as a positive, and it's not something that if we don't ship is going to hurt us in any way.

It's not something customers would wait for.

So I don't see it as, if you will, endangering any revenue opportunity.

Yeah.

And this is Rob.

Just to add to that, I think it's important to step back and look at the 150-terabyte drive is just, the next step, and as Charlie mentioned in his prepared remarks, a robust road map that we have on the flash -- the flash end of the portfolio. each step on the way, each density improvement that we make allows us to do two things.

One is more aggressively compete for lower-cost disc-based systems on an acquisition cost basis.

But also remember, it reduces the power and space requirements and obviously, the associated operating costs with those for our customers.

The drivers for this are quite simple, the denser modules that we ship require less common equipment to support.

Removing that common equipment obviously reduces the cost structure associated with that, but also reduces the space and power associated with it.

And so executing on this density and efficiency road map, is what has allowed us to go aggressively -- so aggressively to date after the most cost-sensitive disc-based systems in the enterprise.

And of course, as we've been discussing with you all, it really opens up the opportunity set as we look to the hyperscaler infrastructure environments.

Thank you, Quinton.

Next question, please.

Meta Marshall, Morgan Stanley.

Great, thanks.

Maybe just diving into these Evergreen//One deals.

I just wanted to get a sense, are some of those kind of with Tier 2 customers or these really kind of enterprise customers that we're talking about?

And is the quantum a handful?

Or is this really kind of just a couple of deals that are hung up?

Thanks.

Yeah, Meta.

It is enterprise customers, because we're talking about large deals.

And it does tend to be just a handful per quarter.

So because it's a handful per quarter and they're not always easily predictable in terms of exactly when to transact, it does make it more challenging to forecast.

Thank you, Meta.

Next question, please.

Mehdi Hosseini, Susquehanna.

Yes, thanks for spending -- excuse me.

Thanks for taking my question.

Charlie, in past earnings conference calls, you have talked about engaging up to 10 data centers, including hyperscalers.

Could it be your traction with these data centers, large and small, be more focused on selling products which you're capturing in your product revenue?

And until enterprise and Evergreen//One and Evergreen//Two were to pick up to hit that $500 million run rate, you're not really going to be able to scale that subscription model but you are penetrating data centers, large and small, and that's captured in product revenue.

Is that the right way of thinking about the current dynamics?

Thank you.

Yeah.

Thanks, Mehdi.

I would say we're speaking about Evergreen//One, Evergreen//One is being sold into both large and small environments; as Kevan put it, large deals and small deals.

Actually, the run rate of -- in what I'd call, the commercial market, small deals of Evergreen//One is very strong and very good.

It's very good as well in large environments and even into cloud environments.

We're selling quite a few into MSPs and other clouds as well.

So that's going well.

All that goes into our Evergreen//One.

When we classify the $500 million, for example, those are all regardless of who they sell into identified as Evergreen//One.

I hope that answers your question.

Thanks, Medhi.

Next question, please.

Jessica Hansen, Raymond James.

Yeah.

Thanks.

This is Jeff Koche in for Simon Leopold.

Maybe you can give us an update on like how -- what percentage of revenue at this point is AI?

And maybe, more importantly, like given all the AI tailwinds, how does the order book?

Like what percentage of the order book is starting to become AI related?

Well, we've not split out AI, but let's give you a little bit of a flavor on AI.

As I identified in my prepared remarks, we really do foresee three major segments.

And it's really just the first one that has become a real -- I would say, real revenue at this point in time, not just for us, but for some other players out there as well.

And that is the training environment.

Now we've been selling into training environments for over five years, probably closer to six years right now.

Of course, the training environment really hit the news once it became large LLMs, right?

But there's been training environments for self-driving cars for drug discovery for various different medical technologies as well as high-speed trading.

That's all been training models, typically what are known as parameter-based models.

More recently, obviously, there are generative AI models that have been getting a lot of the news.

Our view is that in total, the total market or storage for large language-type models, training models, less than $1 billion a year currently.

I would say that we're getting our fair share of that considering where we are.

So that gives you sort of an order of magnitude, but we haven't broken it out specifically.

The other two areas are the inference or rag model inside of enterprises.

That's just starting to be discussed and investigated.

I think not really started yet in terms of revenue, but we expect soon.

And then the third area is a general, we believe, up-leveling of existing storage environments to make data that's sitting currently in data silos inside of organizations much more available for use for inference.

And as you might imagine, that's a third on the list in terms of order of when revenue will hear, but we think it's the largest overall in the enterprise space.

And a lot of what we've been working on has been in preparation to enable organizations to take advantage of that.

Thank you, Jeff.

Next question, please.

Krish Sankar, TD Securities.

Hi.

This is Robert Mertens on for Chris.

Just with the reiterated full-year guide, assuming the midpoint of the October quarter outlook, that would imply the January quarter growth decelerates a bit to mid-single digits year on year.

Could you just speak towards some of the puts and takes in the guide if fans and industry inflection is mid- to high single-digit growth, the new norm?

And is there any sort of seasonality headwinds to expect in the January quarter?

Yeah, I appreciate the question.

This is Kevan.

And look, glad to see that our total revenue is tracking with our expectations for the year, which again, is double digits.

And it's tracking as well in terms of what we saw for the first half of the year, but you got it.

The primary factor in terms of what you see in terms of growth in first half and second half is largely due to seasonality.

That would be the primary driver and that's been consistent historically for many years with the exception maybe a one year, following COVID.

But another consideration as well is that we are expecting a sales ramp of our as-a-service offerings, especially with our higher velocity business in the second half of the year.

And this would create some headwind as well to our expected total revenue growth.

And that's been considered as well in our annual guide for total revenue.

Thank you, Robert.

Next question, please.

Eric Martinuzzi, Lake Street Capital Markets.

Yeah.

I wanted to see if we could put a finer point on the gross margin commentary for the product.

Kevan, I think you said that the 69.5% was down year on year.

And then for the -- if I can recall, it was the full year or just the back half, but a modest strategic decline in product gross margin.

Would you care to comment on that?

Sure.

And thanks, Eric.

Look, we've been talking for some time about our sweet spot for product gross margins really being in the high-60s.

And it's taken us a long time to achieve that.

And so I actually think this is quite a positive for us in terms of what we're seeing in the high-60s.

Now we are seeing and looking at a modest sequential decline that I talked about in my prepared remarks.

And that's really coming from the strong momentum and demand from customers shifting their cost sensitive workloads to our all-flash solutions, whether that's our family of E solutions or FlashArray//C.

And look, we'll continue to aggressively pursue this transition.

And that's really what's driving our expectation of seeing a modest reduction in our product gross margins in the back half of the year.

And as well this has been considered in our annual guide for operating profit as well.

Thank you, Eric.

We have one more question.

So this will be the last question.

David Vogt, UBS.

Great.

Thanks, guys.

Thanks, Charlie.

Thanks, Kevin, for all the detail.

I just had a thematic question about the hyperscaler opportunity.

We appreciate all the color, but can you maybe talk to how the hyperscalers are viewing your solution?

What I mean by that is what kind of workload, what kind of application, kind of what are the use cases that they're thinking about as you have these extensive discussions, so we could think about sort of how your product would dovetail with their existing architecture.

Thank you.

Yeah.

Let me start, and I think Rob will add some color to this.

Honestly, it's a very broad -- it's a fundamental architectural shift for them.

And the more -- and obviously, we're at different stages in the conversation with different hyperscalers.

But the further we go down the path with any one of them, they're looking to make it an architectural shift in their infrastructure, whereby it would replace not only higher performance workloads, but lower performance workloads, including disc.

And once implemented, no reason for it not to replace, for example, even their use of SSDs, so fairly broad.

Now knock on wood, we haven't -- nothing has been signed yet, but the longer the further in we are in conversations with hyperscaler, the more they test us, the more we discuss terms and conditions and structures, the broader the use cases are in those discussions.

Yeah, David, and this is Rob.

Just to add on that, I really want to emphasize a point that Charlie made, which is really going after this and looking at this as a broad set of workloads targeted at various performance and architectural replacement points.

And what I mean by that is these hyperscalers operate so many workloads, so many different environments, they get economies and simplicity of scale and operations by standardizing the infrastructure largely to serve all of those workloads.

They generally don't build out special purpose infrastructure for each workload one by one.

And so certainly, within and under that umbrella, have got higher performance workloads, typically being served by flash today all the way to lower performance workloads, largely being served by disc.

We think in the long term, we've got opportunity to go provide value across the board with our DirectFlash technology.

But certainly, as an initial step, would be focusing on the replacement of the disc base environments in that infrastructure.

So hopefully, that's helpful for you, David.

Yeah.

Thanks, David, and thanks, Rob.

That's a great question to end on.

Charlie has some final comments.

Yes.

I want to thank you all again, as always, for joining us today on today's earnings call.

Our platform strategy and our innovation is once again transforming the storage industry.

We empower enterprises to address their fragmented data silos, which is going to be a critical step in unlocking their full potential of analytics and artificial intelligence.

And we believe that providing a unified, versatile, and energy-efficient platform, we're going to enable businesses to embrace the technological change that they face and seamlessly integrate their data centers with their cloud environment as well.

I want to thank everyone, particularly our customers, employees partners, our suppliers, and our investors.

Your dedication, collaboration and trust are the driving forces behind our success.

Thank you, all, for your continued support and commitment.

Goodbye.

That concludes the Pure Storage second-quarter fiscal '25 financial results conference call.

Thank you for your participation.

You may now disconnect your lines.