Sean Kenney:
Hi, I'm Sean Kenney, and welcome to the MFS All Angles podcast. Over the past few episode, we've gone deep into industries, from pharma to software and energy, but today we're focused on the hyperscalers and talking about the massive CapEx expansion here in 2026. While this topic is well covered in the media, our goal today is to share the insights and focus areas coming from our research team. To do that, I'm joined by Erica Zieba and John Haddad. Erica's an equity research analyst on our technology team, John covers technology on our investment grade fixed income team. So, I want to thank you both for joining me, this is a rich conversation. There's a lot of places we can go with this, but having both equity and fixed income around the table I think will be great for our listeners to hear how you're thinking about that within your particular coverage, but also as a team across the platform. So, thanks for joining us.
Erica Zieba:
Thanks for having us.
John Haddad:
Happy to be here.
Sean Kenney:
Well, so let's start with Erica, I want to start with the overall theme in the industry, which is this massive CapEx expansion. We've seen 600 plus billion dollars of commitments from just a handful of companies here in 2026. And one of the key themes that we wrote about in our 2026 outlook was is the AI premium too high? And some of the narrative that you hear in the industry is that is there a corollary here in 2026 with all the hype and the massive investment with AI, to the dot com bubble back in the late '90s and early 2000s. And I know we've talked about this and you share that there isn't quite the same backdrop. So, maybe you could spend a little bit of time sharing your perspective on the backdrop today relative to that period in time.
Erica Zieba:
Yeah, sure. Yeah, I think there's so much press around the parallels, but I would actually draw, I guess the attention to what's actually different this cycle than what was happening in the early 2000s, which are a number of things. The first being that late '90s, early 2000s, the build out there was really speculative build out on the demand that was still to come. So, we were building to demand that didn't actually materialize for 10 or 15 or 20 years in some cases. And so, that's a huge difference. I think this time we're actually building and the hyperscalers are building to demand in their cloud that they're actually seeing an underwriting to an RPO that they're actually printing today. So, I think that's the first difference. The second difference is that the funding of the CapEx build is being underwritten by very well capitalized large enterprises with hundreds of billions of dollars actually sitting today on their balance sheets.
And so, when I think back to the early 2000s that was certainly not the case. A lot of it was debt financed, a lot of it was speculative, a lot of it you just couldn't underwrite to a return to. Whereas these companies, their bread and butter is cloud and they're underwriting with strong balance sheets. And then I think the third thing is that the capacity that's going in the ground today is not sitting idle. So, in the early 2000s we put a lot of fiber in the ground that just sat dark for a number of years. This time we're supply constrained. So, every cloud provider would tell you, "Look, we can't actually get enough chips fast enough to actually fulfill the demand we're seeing from our customers on the other side." And I think that continues through '26. And so, I guess it would be the three big differences. I don't know if I'm missing anything.
But when I think back to that time period there are a number of parallels, but I would say that this cycle certainly feels like it's off to a more healthy start from a capitalization perspective than the last.
Sean Kenney:
Yeah, it's interesting. The companies that have massive balance sheets and could fund this through free cash flow are still accessing the debt markets in some cases. So John, from a fixed income perspective, we had Google make a big announcement just this week around the investment or issuing bonds to fund this. It was unique in that it was very global, even a 100-year bond in the UK, which is pretty unique. So, tell me a little bit from the fixed income perspective how you're seeing things.
John Haddad:
Sure. Well, what's interesting is that for the first part of this journey AI was really more of an equity market phenomenon than it was a fixed income market phenomenon. So if you think back to 2024, which was the first year we had these big CapEx step-ups, if you think about the five big investment grade hyperscalers, so Microsoft, Google, Meta, Oracle, and Amazon, collectively, they spent about 240 billion of CapEx in 2024. Net issuance, so what they issue into the market minus what they actually mature from a debt perspective in a given year was de minimis. And so, that was entirely funded with internally generated cash flows. 2025 was the pivot year, that's when we started to see them lean more heavily on their balance sheets. Those same companies spent about 400 billion in CapEx last year, and they issued about 90 billion in net issuance over the course of the year.
And that actually excludes the Meta Blue Owl JV funding that we saw, which we can certainly unpack later on if you want. But that's been a little bit of a wake-up call for our market in particular, especially as you think about a starting point where credit spreads are at multi-decade tights, that's underpinned by a favorable technical backdrop. We've had inflows into the asset class and the supply demand picture has been very favorable. And so, you have investors now beginning to reassess what that multi-year issuance from these players is going to look like. I think looking out over the next few years, 2026 is going to be a year where the CapEx bumps are so large that free cash flow margins are actually thinning for a lot of these companies, and they're actually going to be leaning a lot more heavily on our market. Those same five companies are anticipated to do something north of 700 billion this year, which is a pretty big step up.
And for our market, that means that 90 billion of net issuance could look something like 150 billion this year, conservatively. And so, that's a pretty big chunk for a market that does a trillion and a half of issuance in a given year.
Sean Kenney:
Yeah. Who are buyers of these notes? These are highly rated companies, is there value in some of these issuances?
John Haddad:
Yeah. I mean, I think the ownership in the corporate bond market is pretty diverse between institutions like pension funds, mutual funds like ourselves, and international flows. We've seen fairly strong international flows into the asset class that's helped support and take some of this funding down. I think from our perspective on the fixed income side, and the question that I'm always asking myself is, are we getting paid enough spread to really take on the risk here? Traditionally, companies come to the fixed income market in a certain point in their maturity life cycle, which is they tend to be further along in that and they have more stable business positions with predictable cash flows or asset bases that can be lent against. These are somewhat uncertain investment propositions, and the competitive environment could be shifting. And so, as a fixed income investor I'm constantly assessing whether I'm accidentally taking equity risk or I'm actually getting compensated for that uncertainty in spreads.
Sean Kenney:
Yeah, interesting perspective. And you mentioned the off-balance sheet funding, so maybe we'll spend a minute there before we get back to the equity side, but talk to me about the off-balance sheet financing. We hear a lot about these special purpose vehicles that some of these players have utilized, and even some narrative around circular funding. Any red or yellow flags from your perspective? And Erica, I'd be curious of your take too.
John Haddad:
Yeah, I think you hit on the two main ones, which is when we think about off-balance sheet and worries from a debt perspective, there's two avenues to it. The first one is the vendor financing and circularity, and this idea that we're hearkening back to the late '90s. I think there are important differences. When you look back at that period, you had equipment makers who were lending to their customers so that they could buy their own equipment, and there wasn't really any revenue on the other side. In this case, as Erica alluded to, these structures that NVIDIA is promising or whatever tend to be more equity oriented so they're getting equity back as opposed to lending and debt form, and they're to support revenues and backlog that's here today from a supply constraint position. So, I think that is a key difference that should be highlighted.
That being said, it is a watch item, we're watching it really closely and there could be a tipping point here to the extent that that continues to escalate and could create issues. The other piece of it that you also hit on is just off-balance sheet. And so, what we've noticed is some of these weaker players who have less capacity to fund this on their own balance sheets, the Metas and the Oracles of the world are increasingly looking to off-balance sheet vehicles to get the funding in order to build this out. So, there's a couple avenues that takes, it's in some cases a JV where they're bringing in a private equity partner to actually own the data center. And then there's other instances where it's just off-balance sheet leases or guarantees or whatnot that the company's making over a very long period of time.
In both of those instances as a credit investor, we're thinking about that in leverage and credit metrics, the rating agencies consider all of that as well. And so, I wouldn't say that these companies are necessarily getting a free lunch, it's more around flexibility, I think, from a capacity perspective.
Sean Kenney:
Yeah, interesting. Yeah.
Erica Zieba:
I'll just add one thing on circular financing too, because I think it gets a lot of negative press, and to a large extent it's a watch item for us. But I think there's a couple of things that make it a little bit more healthy than what we certainly experienced in the early 2000s. The first is that a lot of these equity investments are gated by milestones. So NVIDIA investing $100 billion for 10 gigs at OpenAI, I mean, that is going to go gig by gig. And so, we'll see how that progresses but that's not a committed hundred billion. So, it looks like a massive circular financing wheel but it actually is gated pretty well. I would say that the second thing is these deals that we're seeing from chip makers into the model producers and trainers, a lot of that is strategic.
So, if you're NVIDIA and you're sitting from a position of advantage on a GPU, you want to make sure that your roadmap and your product roadmap is aligned with OpenAI, Anthropic, Meta, Microsoft to make sure that you are the chip that is taking the same amount of incremental share in the future. So, I think there's the circular aspect of it that gets a lot of news, but there's also the strategic aspect, which is a viability question I think, on the chip side. So, I think those two things are actually pretty important differences.
Sean Kenney:
Yeah. I feel like what we're already touching on is just the idiosyncrasies company by company, just in terms of how you finance, in terms of how you're going to generate revenue from these investments that they're making. So, maybe if we move a little bit more to the equity side, talking about free cash flow, when you think about some of these players, they can fund a lot of this through existing free cash flow. They can tap the debt markets because they're highly rated. The cost of capital is pretty light, so they're going to use the debt markets as well. But I would imagine that understanding company by company, how they're financing and how they intend to generate the revenue to justify the financing is very much a company by company evaluation. So, I'd love for you to put our listeners into the research room. You all benefit from not only cross capital structure collaboration, but collaborating with your colleagues across the technology sector more broadly, other sectors and industries as well.
Put us in the room. What are the things as you're evaluating equities and businesses, what are you looking for? What are the flags and the key tent poles that you're looking for?
Erica Zieba:
Yeah. I mean, I would say the biggest question that we've certainly focused on and we should continue to push on is just the return from all the CapEx, because I think that goes to the question of sustainable CapEx in the future and what sustainable free cash flow and free cashflow margins looks like business by business when we look out to '27 and '28, and you're right, it differs massively between companies. So if you take a company like Microsoft, it's a rational CapEx spender. It is building and spending its CapEx to the incremental revenue dollar that it's going to convert in cloud. It has a pretty consistent historical track record of actually monetizing cloud. And so, when it's underwriting to commitments from its customers, it knows exactly how much it's spending and what the return is going to look like. I think that's a very different value proposition than someone like a Meta who doesn't have a cloud and who is building its own open source model to distribute into a potential consumer product in the future, and we don't know what that looks like.
So, I think the models are quite different, but we spend a ton of time thinking about what the actual return needs to be to justify the CapEx. And I think the hard thing is it's coming in a number of places. Today we're seeing the return in cloud, today we're seeing the return in digital advertising. We're seeing Google search and Meta social advertising dollars certainly scale. But outside of that and some pockets of coding, we're really not seeing the revenue dollars convert. So, 2026 really for us is a year where we need to start seeing the demand inflection on the enterprise side that actually justifies the spend. We spend a ton of time with our colleagues on the ground talking to privates and talking to the people who are actually building the applications to make sure that when we're modeling to revenue at each individual company, idiosyncratically we are not missing something on the CapEx spend.
And what we see today as well, overall returns for the hyperscalers are certainly going down, returns that we see are above the cost of capital. And so, if Microsoft or Google is actually getting revenue and getting returns above the cost of capital, I think this cycle's potentially too big to miss for them to start underspending. And so, I think that's probably the bulk of where we spend our time now thinking about is just the justified return profile, but you're right, it's different between every single company, which makes the job fun.
Sean Kenney:
Yeah, yeah. I would imagine that when you're collaborating across the platform, you mentioned getting access to the private markets and we're managing public equities and public fixed income for our clients, but how do you tap into the private markets and the research and the roadshows? Tell us a little bit about that.
Erica Zieba:
Yep. So, I mean, the great thing about being a large platform is that while we don't invest on the private side we have relationships with CFOs, and IRs, and CEOs who used to either be part of public companies or who are in our network that we can actually tap into to say, "We'd love to speak to someone at OpenAI." Can we find someone on our team to actually get us in to see how they're thinking about scaling, what their cost advantages, what the value proposition and how their financial goals are stacking up and how that builds to what we think about from a Microsoft or an Oracle perspective.
We attend a ton of conferences, we spend a ton of time in San Francisco just going to sell side conferences that facilitate VCs, private equity and privates who are building actual applications and businesses in AI. We spend a ton of time talking to them about the businesses that they're trying to build and how the tooling suites are evolving. Those are probably the two biggest ways, but I would say the other great thing is podcasts like this have democratized a lot of information. And so, while we're getting at what feels like at times we're getting new served out of a fire hose, that's also a great way that we keep our pulse on the market.
Sean Kenney:
Yeah. John, anything on your end there?
John Haddad:
No, I thought that was great.
Sean Kenney:
Yeah. Well, I mean, I feel like one of the things we've been talking about is uncertainty. I mean, this is uncharted territories, it's unprecedented levels of investment going into a capability that there's conviction in but is yet to be truly monetized by these businesses. What gives you confidence when you see a company or when you see a revenue stream, what gives you confidence that that will sustain? What are you looking for?
Erica Zieba:
That's a hard question. So I think the other thing, going back to your last question on returns, I think it's also important to think about not just the revenue conversion but the actual cost and the labor impact to a lot of these investments. And so, when we look at businesses like Google or Microsoft, we're looking at the top line and how the CapEx is actually translating to the viability and the profitability of the business. We're also looking at what the compute is going towards internally to actually create the tooling and the product suite, how they're getting their own internal efficiencies on engineering, how they're deploying GPU to actually scale product and go to market. And so, when we think about overall return, I think the numbers get somewhat muddied. And so, I think there's a lot of news flow out there in the press that returns are not where they need to be, and I would just caution that the return can be on the revenue line and the OpEx line.
And so, when we think about sustainable levels of investment, I think we need to think about both buckets. And so, I mean, even yesterday we had a large bank that we're invested in, and we were talking to them and they gave us a pretty large number that they think that they're hoping to target on OpEx saves for next year because they're getting savings from coding, they're getting savings from back office. We're seeing it in IT services numbers and financials as companies look at ways to utilize AI to take out costs. And so, when we think about sustainability, hearing those anecdotes and seeing those in numbers, I think is just as important as seeing the revenue conversion hyperscale.
Sean Kenney:
Yeah, it's a great point. I mean, so much of the industry is focused on the revenue side and there is value to be realized on the other side as well, so that's a good perspective. So, let's talk about another area of uncertainty, which is geopolitics. And we published in our 2026 outlook, one of the key themes we talked about in that as well was just really understanding and navigating geopolitics and the geopolitical risks that are out there. And I would imagine with all of this funding, given that it's not just U.S. data centers that are being built, these are being built around the world and this funding is going around the world, that geopolitics creates some risk to a business's investment decisions and their CapEx spend. Maybe spend a minute talking about that. How do you think about that? How do we as a team wrestle with that?
Erica Zieba:
Do you want to take it?
John Haddad:
Sure. So, I maybe have a little bit of a unique lens in that I actually cover aerospace and defense on the fixed income side as well. So, we spend a lot of time actually as part of that thinking about the budget and the administration's priorities as it relates to national security. And I think there's a pretty clear view among administration officials here in the U.S. as well as in other countries that AI has the potential to be a real game changer, and that leadership in that is going to be of strategic importance to us from just the national security perspective going forward. And so, as you think about the constraints and the question marks around our leadership, there's two big pieces to it. It's energy security and making sure that we have the energy available to power all these data centers, the labor in order to build the data centers, and reshoring semiconductors because we continue to remain very reliant on TSMC and Taiwan for the vast majority of these AI accelerators.
So, all of that is going into the mix and how they think about our needs going forward.
Erica Zieba:
Yeah. I mean, I would say the last four to five years we've seen a number of regulations from the U.S. on China in terms of what in the semiconductor industry is actually allowed to pass from a technology perspective. So, selling GPUs into that market and what chips we can actually sell into that market just to protect our overall national interest in technology leadership. What's interesting is that the actual vast majority of chip production from NVIDIA's side is coming from Taiwan. And so, what we think about and how we think about the multiples we want to pay for something in semis that does have that geopolitical profile is, we spent the last three years restricting China, is there a point where China comes back and restricts access for us? And so, we've seen a lot of semiconductor build out and capacity build out in the U.S. and other parts of the world that start to globalize us a little bit, but the reality is in three years I think only 30% of TSMC production is going to sit outside of Taiwan.
Which is not nearly enough if you look at the CapEx dollars these companies are spending to give us the chip capacity that we would need. So, when we think about what to pay for businesses like this, that's certainly top of mind. We need to make sure that that access to capacity is actually there. Sean Kenney:
Yeah. I mean, this is uncertain times, it's a global problem and it's multi capital structure issues that we're dealing with here. So I'm curious, maybe as our last question, how do you all on the investment team support each other through this? You're navigating a lot of complexity. You've got a global team, you've got a cross capital structure team. Any examples or thoughts on how you support each other through this research process?
Erica Zieba:
Yeah, tons. I think John might be sick of me after the last couple of years.
John Haddad:
And you of me.
Erica Zieba:
I just think that, I mean, the velocity with which this is moving, this is moving faster than I think the last probably three technology or platform shifts have moved combined, and it's happening in equity and it's happening in fixed income. And it's not just a technology question, it's a market question. So, the concentration of the names that we follow is massive and has implications across the cap structure, but also globally and across every sector. And so, what we've tried to do is make sure that when we're building a CapEx model and we're building a free cash flow model, that we understand you can't just go out and buy 100,000 H200 chips and hope that the energy is going to be there. So, we are doing a lot more cross sector meetings and cross cap structure meetings on things like energy.
So, if John and I are modeling CapEx and we're getting to over a trillion dollars in global CapEx spend on AI in three years, what kind of energy and what kind of power do we need to do that? And can we make that make sense in the U.S.? And that's a question that our energy team has to step in. So, we find ourselves doing a lot more cross sector work. Another example would be Oracle. So, in the fall they signed a $300 billion RPO deal with OpenAI. And the question there with the balance sheet that was already three times levered is how are they going to make this work if there's a lag between when they need funding to actually capacity the data centers versus seeing revenue come online, how is it going to work?
And so John and I, mostly John, spent a lot of time helping me think about what the flows need to look like, what the high yield market is actually comfortable underwriting, what the investment grade market is comfortable underwriting and thinking about other special creative finance vehicles that might make sense outside of my market but might occur more on the fixed income side. And so we looked at it, I couldn't quite get there on balance sheet but that is a question that John had to help talk me through.
John Haddad:
Yeah. I mean, you'll remember in the fall there was a lot in the press around Oracle's credit worthiness, and its CDS spreads were blowing out and its bond spreads were blowing out. Through working with Erica on the fixed income side, it was immensely helpful in her ability to help me sensitize the CapEx numbers. So, last fall when Oracle brought a large deal, to be able to compare our models and to say, "Wow, I think the market's still really underestimating CapEx. I think there's another deal coming, I think this feels a little tight from a spread perspective and it feels like this could get a little bit worse." And so, that collaboration really does result in good outcomes across the platform and good investment decisions I think, that ultimately benefit the client.
Sean Kenney:
Yeah. Well, it's really helpful to get your perspective and for bringing some of this to life. As I like to say, it's like putting us in the room to hear the conversations and the things that you guys are wrestling with as a team, so thank you for that. And thank you for listening to All Angles. If you enjoyed this episode, subscribe so you don't miss any future episodes. And until next time, be sure to consider your investment decisions from all angles. The views expressed are to those of the speaker and are subject to change at any time. These views are for informational purposes only and should not be relied upon as a recommendation to purchase any security or as an offer of securities or investment advice. No forecast can be guaranteed. Past performance is no guarantee of future results.
Distributed by: U.S. – MFS Institutional Advisors, Inc. ("MFSI"), MFS Investment Management and MFS Fund Distributors, Inc., Member SIPC; Latin America – MFS International Ltd.; Canada – MFS Investment Management Canada Limited.; Note to UK and Switzerland readers: Issued in the UK and Switzerland by MFS International (U.K.) Limited ("MIL UK"), a private limited company registered in England and Wales with the company number 03062718, and authorised and regulated in the conduct of investment business by the UK Financial Conduct Authority. MIL UK, an indirect subsidiary of MFS®, has its registered office at One Carter Lane, London, EC4V 5ER.; Note to Europe (ex UK and Switzerland) readers: Issued in Europe by MFS Investment Management (Lux) S.à r.l. (MFS Lux) – authorized under Luxembourg law as a management company for Funds domiciled in Luxembourg and which both provide products and investment services to institutional investors and is registered office is at S.a r.l. 4 Rue Albert Borschette, Luxembourg L-1246. Tel: 352 2826 12800. This material shall not be circulated or distributed to any person other than to professional investors (as permitted by local regulations) and should not be relied upon or distributed to persons where such reliance or distribution would be contrary to local regulation; Singapore – MFS International Singapore Pte. Ltd. (CRN 201228809M); Australia/New Zealand - MFS International Australia Pty Ltd ("MFS Australia") (ABN 68 607 579 537) holds an Australian financial services licence number 485343. MFS Australia is regulated by the Australian Securities and Investments Commission.; Hong Kong - MFS International (Hong Kong) Limited ("MIL HK"), a private limited company licensed and regulated by the Hong Kong Securities and Futures Commission (the "SFC"). MIL HK is approved to engage in dealing in securities and asset management regulated activities and may provide certain investment services to "professional investors" as defined in the Securities and Futures Ordinance ("SFO").; For Professional Investors in China – MFS Financial Management Consulting (Shanghai) Co., Ltd. 2801-12, 28th Floor, 100 Century Avenue, Shanghai World Financial Center, Shanghai Pilot Free Trade Zone, 200120, China, a Chinese limited liability company registered to provide financial management consulting services.; Japan - MFS Investment Management K.K., is registered as a Financial Instruments Business Operator, Kanto Local Finance Bureau (FIBO) No.312, a member of the Investment Trust Association, Japan and the Japan Investment Advisers Association. As fees to be borne by investors vary depending upon circumstances such as products, services, investment period and market conditions, the total amount nor the calculation methods cannot be disclosed in advance. All investments involve risks, including market fluctuation and investors may lose the principal amount invested. Investors should obtain and read the prospectus and/or document set forth in Article 37-3 of Financial Instruments and Exchange Act carefully before making the investments. For readers in Saudi Arabia, Kuwait, Oman, and UAE (excluding the DIFC and ADGM). In Qatar strictly for sophisticated investors and high net worth individuals only. In Bahrain, for sophisticated institutions only: The information contained in this document is intended strictly for professional investors. The information contained in this document, does not constitute and should not be construed as an offer of, invitation or proposal to make an offer for, recommendation to apply for or an opinion or guidance on a financial product, service and/or strategy. Whilst great care has been taken to ensure that the information contained in this document is accurate, no responsibility can be accepted for any errors, mistakes or omissions or for any action taken in reliance thereon. You may only reproduce, circulate and use this document (or any part of it) with the consent of MFS international U.K. Ltd ("MIL UK"). The information contained in this document is for information purposes only. It is not intended for and should not be distributed to, or relied upon by, members of the public. The information contained in this document, may contain statements that are not purely historical in nature but are “forward-looking statements”. These include, amongst other things, projections, forecasts or estimates of income. These forward-looking statements are based upon certain assumptions, some of which are described in other relevant documents or materials. If you do not understand the contents of this document, you should consult an authorised financial adviser. Please note that any materials sent by the issuer (MIL UK) have been sent electronically from offshore. South Africa - This document, and the information contained is not intended and does not constitute, a public offer of securities in South Africa and accordingly should not be construed as such. This document is not for general circulation to the public in South Africa. This document has not been approved by the Financial Sector Conduct Authority and neither MFS International (U.K.) Limited nor its funds are registered for public sale in South Africa.
67556.1