This was a longer than normal Q&A which I understand since Intel in in transition. Here are the pertinent technology questions:
Questions & Answers:
Ross Seymore -- Analyst
Hi, guys, thanks for letting me ask a question, and congrats on solid results. Pat, I want to talk about the 18A transition. Can you just talk about what are the metrics we're going to be able to see externally on this to give people confidence in the ability to ramp it? You've said it's healthy. You've talked about new design wins. But when are some of the true metrics going to come, either internally, or perhaps, even more importantly, externally, as your Intel Foundry revenue from an external customer base grows?
Patrick P. Gelsinger -- Chief Executive Officer
Yeah. Thanks, Ross. And obviously, this was a good quarter on the progress that we had, three new customers, Amazon, which we were public on a few weeks ago, Panther Lake and Clearwater Forest, internally. Two new external customers added to that. So, solid progress on it. Clearly, next year, there's not a lot of financial benefit from it because we're only ramping late in the year. Thus, we'll be giving more qualitative metrics on progress as we go through the year, Ross. We'll continue to update, give LDV updates, lifetime deal value for foundry updates as we go through the year.
Clearly, as we get more customers, we'll be updating on that progress. It has much larger impacts on '26 financials as we ramp and bring wafers home as well as move into the better margin structure that we'll have on Panther Lake with 18A and across the product line. So, we'll be giving clarity in that way, but it will be hard to tie it to specific financials next year, but this is super important for us, for our foundry business, for the industry, so we'll be giving plenty of color as we proceed.
D.A.N. Two new external foundry customers added is great progress in my book. I believe one is Microsoft, I'm still looking for the other one. Amazon AWS Annapurna labs is a big TSMC customer so that is significant. It looks like Amazon is playing both sides of the foundry coin, which is what expanding the NOT TSMC market is all about. On the other side of the coin, here is a quote from the TSMC event in September:
“Our collaboration with TSMC on advanced silicon solutions for our AWS-designed Nitro, Graviton, Trainium, and Inferentia chips enables us to push the boundaries of advanced process and packaging technologies, providing our customers with the best price performance for virtually any workload running on AWS.” – Gary Szilagyi, vice president, Annapurna Labs at AWS.
Timothy Arcuri -- Analyst
So, Pat, I think you mentioned at a conference maybe in September that defect density on 18A is you said sub-0.4, I think. Can you just talk about how that translates to yields? Is that sort of an enough number to translate into high volume? And what sort of number for, say, defect density does a customer want to see when they're looking at your foundry?
Patrick P. Gelsinger -- Chief Executive Officer
Yeah. And defect density, it's a complex conversation, Tim, because that big die sizes have lower percent yields, right, even at the same defect density. So, it's very variable depending on the particular die size of the product. So, when we said the D0 of less than 0.4, that was a healthy yield number at this phase of the process development.
That's not yet a high-volume production yield level, but we're not at that phase of the process development yet for 18A. So, it's a number that says we are at where we'd want to be at this point in the process development life cycle. Clearly, as you bring it to high-volume production, which we'll be doing in the second half of next year, we have to be markedly lower than that in terms of defect density. But we believe that we see all the signs, the signals, and we're managing this very carefully to accomplish that as we get to next year.
Similarly, as we look at today's Intel 3, we're accomplishing the defect densities with the maturity levels that we'd expect on that. But again, yield in the process technology is something you're never done with, right? You hit key milestones on quality, yield, and then you go into high-volume production, and you continue to work on that going forward. I'd also emphasize that the Arizona ramp is important for us as we move to 18A. That comes online in volume in the second half of next year.
And that's all on track as well as we have tool move-ins, EUV tools, and qualification now, first wafers coming out of the volume fab in Arizona in Q1 of next year. So, overall, we're progressing well, and we'll be giving you updates as we proceed.
D.A.N. Personally I am not concerned about 18A yield. I have heard from the ecosystem that the test chips were good. It was a PDK problem that delayed things.
Timothy Arcuri -- Analyst
I have a question on Panther Lake, Pat. So, I know that it's coming back in-house, but we do still hear that most tiles are still being outsourced. Can you speak to that? Is that a change? Or was that always the plan? And I guess when you look at Nova Lake on the desktop side, is it still being dual tracked, meaning that there is still an option that it could be outsourced? Or is it guaranteed to be brought back in-house? Thanks.
Patrick P. Gelsinger -- Chief Executive Officer
Yeah. Thank you. Panther Lake, some tiles would be external, but the majority of the millimeter square in the package are back internal. It's more 70-plus percent of the silicon area is back in-house. So, the majority of Panther Lake wafer capacity by a good margin is coming back inside for Intel. Nova Lake, we definitely have some SKUs that we're looking at continuing to leverage externally, but the large majority of Nova Lake and more of the additional tiles have come back in-house as well. So, we still have some flexibility in the Nova Lake product, but the large majority of that is committed to the Intel Products -- or Intel Foundry. So, overall, we are absolutely executing on the bringing wafers home strategy that we've laid out.
That said, TSMC has been a great partner. Clearly, Lunar Lake has demonstrated the strength of the partnership and one that we'll use selectively in our product lines for the future. But a large percentage of wafers coming home. That meaningfully fills our factories.
That also meaningfully improves the margin structure of Intel and Intel Products.
D.A.N. As I suspected, Intel will be moving away from TSMC as soon as possible. I just don't see how Intel Foundry can succeed without 100% support from Intel product groups. And as I have heard, Intel chip designers had better have a compelling reason to use TSMC moving forward.
C.J. Muse -- Analyst
In terms of the optionality in terms of securing leading-edge capacity, obviously, you're all in on 18A. But if that were to get delayed, what do you have in terms of negotiated capabilities in terms of securing capacity in '26, '27? And on the other hand, if 18A proves to be more successful, how are you thinking about time frame being more aggressive in terms of adding capacity in Arizona?
Patrick P. Gelsinger -- Chief Executive Officer
Yeah. Thank you, C.J. I'll say we have optionality in the product portfolio that way. We have more resilience in our supply chain than competition does as a result.
So, we feel quite good about that. We continue to really value the relationship, right, with TSMC. And our product portfolio is set up very nicely with respect to 18A capacity. Given our shell ahead strategy and the investments that we've made over the last several years, we have a lot of flexibility to scale up if market conditions require for our products, but also if market conditions require for our foundry customers as well.
And given the margin stacking nature that we uniquely are able to benefit from, every wafer we bring home adds to the margin structure of Intel in a meaningful way. So, we really are setting ourselves up very nicely for the future. And as we make progress on 18A,18AP, 14A, an aggressive road map, advanced packaging, which is uniquely based on Intel technologies and finding more momentum in our product line, but also in our foundry customers, we really are starting to see the benefits of the long-term strategy that we've put in place with Intel Products, with Intel Foundry.
D.A.N. Good point, AMD is exclusive to TSMC, for now. I do not however ever see AMD using Intel Foundry. The "build shell ahead" strategy is a nice way of saying you can't fill the fabs you had announced?
Aaron Rakers -- Analyst
So, first on the foundry business. As you guys ramp that, we start to focus on the ramp of 18A looking forward given the pipeline of the design wins that you guys have talked about. I'm curious, how large is your external piece of that, call it, $4.4 billion revenue in this most recent quarter? I think it was about $77 million last quarter. And how do we think about the pace of that kind of inflecting higher as we look out into '25, '26? And I'm just curious how you would define success in that external business.
Patrick P. Gelsinger -- Chief Executive Officer
Yeah. So, overall, the external foundry business will be a modest portion, right, of the foundry business for the next couple of years. The foundry revenues will be dominated by the internal products, as we've been building that portfolio for many years. So, we'll ramp as we go through the rest of the decade.
We said our financial objectives by the end of the decade are $15 billion plus of external foundry revenue. We'll be giving periodic updates on LDV, lifetime deal value, of external foundry customers as we go forward. And indicators like we did this quarter of new design wins, new customers are coming on online. But the reported numbers for Intel Foundry will be substantially dominated by the Intel Products for the next couple of years. That said, we saw a nice growth, right, quarter on quarter in the external foundry business, and we did break profitable for the advanced packaging portion of that business in Q3. So, we are seeing nice growth characteristics, nice business characteristics, and we'll be giving more updates as we go forward. Dave, anything to add on that?
David A. Zinsner -- Executive Vice President, Chief Financial Officer
No. I mean, other than -- it was up a bit this quarter but still dominated by internal business within the Intel Foundry business.
D.A.N. There is no such thing as LVD in the foundry business or any other business I am familiar with. Good luck with that one. Someone needs to get an official response from TSMC on LVD. They should have LVDs in the Trillions of dollars!
Srini Pajjuri -- Analyst
Thank you. Pat, I have a question on the foundry side, in particular, the packaging business you have. I know, even though there's a lot of skepticism about 18A, but I think you guys have pretty good packaging technology. And it's proven, I believe, if I look at EMIB and Foveros, etc.
So, my question is, given the tightness in the packaging industry right now, I would have thought we would see more interest in your packaging services. So, I'm just curious as to why we are not seeing that. Is it because it's -- I guess there must be some nuances of using your packaging with the TSM wafers. I'm just curious as to why we are not seeing more interest.
Patrick P. Gelsinger -- Chief Executive Officer
Well, I'd say we do see significant interest in this area. And moving your supply chain, though, is complicated and takes qualification time. And many of these designs were first designed on CoWoS, and now they're looking at Foveros, and then they're looking at the unique technologies that we have like EMIB as well. So, we actually see very good momentum in that area.
Now, those aren't as revenue-producing as wafer design, so they're not nearly, right, as large in size. But the pipeline of activities is very strong. And the quarter-on-quarter improvements we saw were largely driven by advanced packaging. Most of the external revenue that we'll see, which will be nicely up year on year as we go into '25, in the external foundry will be advanced packaging.
And we also see healthy margins for that technology as well. So, moving supply chain is always hard and slow, but the progress that we've seen already, the design wins that we already include in our LDV and then the pipeline of additional designs that we are engaging in, all of it give us great optimism that this becomes a foundational piece of our foundry business for the long term. And as I noted in an earlier comment, we did see our advanced packaging now as a profitable business to stand alone by itself as well.
D.A.N. I do think packaging is critical in the foundry business now that TSMC is working closely their customers but I do not see packaging as a big differentiator for choosing a foundry for wafers. TSMC and Intel both have good packaging solutions so it is not a going to change a wafer decision. Will TSMC customers go to Intel for just packaging? I highly doubt it and I think Pat's answer supports that. He was much more enthusiastic about packaging winning wafer business. TSMC is rapidly expanding CoWos capacity so the NOT TSMC packaging market is shriveling.
Chris Caso -- Analyst
And I think you were a bit clear about what the plans are and what the resulting free cash flow is. I guess, Dave, the question is how much flex maybe in those numbers. I guess as you go through '25 and '26 as well, given the fact that you do have to invest in new technology nodes, how much flex is there for changes in market conditions. Also recognize that you don't have pretty much revenue growth in the plan for that as well.
Patrick P. Gelsinger -- Chief Executive Officer
Chris, yeah, the thing I would -- just to add a little bit to that, Chris, would be that now that we've gotten an EUV fleet, right, into our capex base, we have a lot more flexibility across that fleet. Largely, the capex used for Intel 3, 18A and 14A highly leveraged across it. So, we're much more building to overall capacity requirements and not this very rapid five nodes, four years, get back to a modern fleet of capability. Obviously, the other point to add to that is now that we have finished these five nodes, four years, we're progressing to a more normal cadence of new technologies, as opposed to this racing through capital.
So, that will also give us additional flexibility. And I'll say largely, with the singular exception of High NA EUV, the equipment bases are almost entirely the same across the 14A node as well. And even at High NA, we've built flexibility into the TD development that we have optionality to include or not include that as a central part of it. So, we've built a lot of capacity. Now, we're going to leverage that capacity in much more efficient ways for both an overall high-volume manufacturing and a TD leverage going forward.
D.A.N. A fleet of EUV? That is a new one to me. TSMC has many more EUV systems so what would that be? A butt load of EUV? Actually a butt is a measurement of liquid so that is not it.
Translation: Intel CAPEX will be very conservative moving forward., as it should be.Chris Caso -- Analyst
I do. And I mean, it's a good segue in the next question, Pat, and it's -- I guess a question about your comments on Better Together and if you could kind of explain the rationale of why you think that's the case. Obviously, there's various opinions on that. And if Intel would be willing or has looked at anything strategic beyond what the current plan is right now.
Patrick P. Gelsinger -- Chief Executive Officer
Yeah. And I'll just say we're very comfortable that moving to the subsidiary model, as we outlined in the last earnings call, gives us, I'll say, three that we're aiming for: greater operational rate integrity, right, as we create that clear separation; the opportunity to be able to communicate that more clearly and definitively to our external customers; and then the potential to fund and manage the capital requirements of foundry. That said, the vast majority of volumes through the decade come from Intel Products. The synergies of that co-development and Customer Zero aspect is very substantial as we see it.
And the benefits that we get from the overall cash flows and managing the balance sheet of the company as we go forward are highly beneficial as well. So, for all those reasons, our simple view is distinct but better together.
D.A.N. Translation: Intel Foundry is not going anywhere. From what I heard Pat and the Intel BoD support keep it. Samsung did the same a few years ago, making the foundry business a subsidiary without too much notice. I do not recall people telling Samsung to sell it. Personally I think Intel can shed Altera and other acquisitions but certainly not the foundry business. It would be the death of Intel as we know it, absolutely.
I was hoping to hear more about the Tower Semiconductor and UMC deals. I really thing Intel should do more of those at the N-2 nodes. Globalfoundries is a likely candidate for FinFet nodes below 14nm and Samsung for GAA.
Just my opinions of course.
Intel (INTC) Q3 2024 Earnings Call Transcript | The Motley Fool
INTC earnings call for the period ending September 30, 2024.
www.fool.com
