Latest / DX Today | No-Hype Podcast & News About AI & DX / S&P Warns Meta as AI Spending Heads Past $225 Billion

Transcript

S&P Warns Meta as AI Spending Heads Past $225 Billion · DX Today | No-Hype Podcast & News About AI & DX

• 0:00
  SPEAKER_00: Welcome to the DX Today podcast, your weekly deep dive into the AI ecosystem.
• 0:04
  SPEAKER_00: I'm Rick Spare, and joining me as always is Laura.
• 0:07
  SPEAKER_01: Thanks, Rick.
• 0:08
  SPEAKER_01: And I have to say, this is one of those weeks where the money side of artificial intelligence suddenly became the whole story.
• 0:14
  SPEAKER_01: And one rating agency sort of said the quiet part out loud.
• 0:18
  SPEAKER_00: That's exactly where I want to go today because late on Friday, SP Global Ratings changed its outlook on Meta from stable to negative.
• 0:27
  SPEAKER_00: And the reason was the sheer scale of the AI spending.
• 0:31
  SPEAKER_01: Right.
• 0:31
  SPEAKER_01: And let's be precise about what that means.
• 0:33
  SPEAKER_01: Because people hear negative and think downgrade.
• 0:36
  SPEAKER_01: Meta still holds a double A minus rating, which is very strong.
• 0:40
  SPEAKER_01: A negative outlook is a formal warning, not a cut.
• 0:43
  SPEAKER_00: In other words, the agency is giving the company a clear signal.
• 0:47
  SPEAKER_00: It is watching closely now.
• 0:48
  SPEAKER_00: If the numbers keep moving this way, the rating could come down within two years.
• 0:53
  SPEAKER_01: Exactly.
• 0:54
  SPEAKER_01: And the line from S P that got quoted everywhere was this one we are less certain that it will curtail its spending as previously anticipated.
• 1:02
  SPEAKER_01: That sentence is really the entire story in a nutshell.
• 1:05
  SPEAKER_00: Let's put real numbers on it because I think listeners need to feel how big this is.
• 1:10
  SPEAKER_00: What does SP now expect Meta to spend on capital investment over the next few years?
• 1:15
  SPEAKER_01: The forecast is about$140 billion this year, then more than$225 billion next year, and roughly$252 billion the year after that, before it finally starts to flatten out.
• 1:27
  SPEAKER_00: And if I remember right, those later years were revised up dramatically because the agency previously expected something far more modest for both of those years.
• 1:38
  SPEAKER_00: How big was that jump compared with the earlier view?
• 1:41
  SPEAKER_01: It's a big jump.
• 1:42
  SPEAKER_01: The earlier forecasts were about$163 billion for next year and$178 billion the year after.
• 1:49
  SPEAKER_01: So the new estimates are roughly$50 billion and$70 billion higher, respectively.
• 1:54
  SPEAKER_00: That is an enormous revision for a company that size.
• 1:58
  SPEAKER_00: Now spending alone doesn't scare a rating agency if the cash keeps rolling in.
• 2:03
  SPEAKER_00: So the real question is what happens to free cash flow under this plan?
• 2:06
  SPEAKER_01: And that's where it gets uncomfortable.
• 2:08
  SPEAKER_01: SP projects free operating cash flow of negative$22 billion this year, negative$71 billion next year, and negative$39 billion the year after.
• 2:18
  SPEAKER_01: It used to expect positive cash flow by then.
• 2:34
  SPEAKER_01: And it's already showing up in the actual reported results, not just the forecasts.
• 2:39
  SPEAKER_01: Meta's third quarter free cash flow came in at negative$2.12 billion, which would have been unthinkable for this company a couple of years ago.
• 2:47
  SPEAKER_00: Here's the part I find most interesting, though.
• 2:49
  SPEAKER_00: The debt load isn't really what's driving the warning.
• 2:51
  SPEAKER_00: Meta's long-term debt is around$81 billion, which is manageable for a business this profitable.
• 2:58
  SPEAKER_00: So what's the real issue?
• 2:59
  SPEAKER_01: The real issue is leases.
• 3:01
  SPEAKER_01: Meta has been renting a huge amount of data center capacity.
• 3:04
  SPEAKER_01: Instead of building or borrowing for all of it directly, an SP treats those lease commitments as if they were debt because economically they basically are.
• 3:13
  SPEAKER_00: It's a subtle point that a lot of coverage glossed over.
• 3:16
  SPEAKER_00: A lease feels off the books to an investor scanning headlines, but to a credit analyst, a long commitment to pay for capacity is an obligation just like a bond.
• 3:26
  SPEAKER_01: And the size of those commitments is startling.
• 3:28
  SPEAKER_01: Meta's future lease obligations reached about$355.7 billion, up from roughly$279 billion just one quarter earlier.
• 3:38
  SPEAKER_01: That's about$77 billion added in three months.
• 3:41
  SPEAKER_00: So walk us through the leverage math because I think that's the trigger point for an actual downgrade, and it's the number that bond investors are going to be watching most closely from here on out.
• 3:52
  SPEAKER_01: In 2025, Meta's leverage was about 0.1 times, which is almost nothing.
• 3:57
  SPEAKER_01: SP now says it could go above one times as early as 2027, and that's the downgrade threshold for this rating.
• 4:04
  SPEAKER_01: It could reach 1.3 times once the leases begin.
• 4:08
  SPEAKER_00: So in roughly two years, the company could go from essentially no leverage to pass the line where the agency says it can no longer justify a AA minus.
• 4:19
  SPEAKER_00: That's a dramatic trajectory for one of the strongest balance sheets in technology.
• 4:23
  SPEAKER_01: And Meta isn't funding this purely with cash flow or debt.
• 4:27
  SPEAKER_01: It's also selling stock.
• 4:28
  SPEAKER_01: The company has already raised more than$37 billion in new equity, and it plans about$60 billion more through an at-the-market program starting in January.
• 4:38
  SPEAKER_00: Which tells you something on its own, because a profitable company selling close to$100 billion of its own stock is unusual.
• 4:47
  SPEAKER_00: Normally a business like Meta is buying back shares, not issuing them in this kind of volume.
• 4:53
  SPEAKER_01: That's true.
• 4:54
  SPEAKER_01: And here's a detail I love.
• 4:56
  SPEAKER_01: S P actually left that future equity out of its forecast.
• 5:00
  SPEAKER_01: So in theory, if Meta really does sell all that stock, the leverage picture could look better than the agency's base case suggests.
• 5:07
  SPEAKER_00: Fair point.
• 5:07
  SPEAKER_00: And I want to be fair to Meta here.
• 5:09
  SPEAKER_00: A spokesperson disagreed with the assessment and pointed to what they called significant product and business opportunities ahead.
• 5:17
  SPEAKER_00: So the company is clearly not accepting the framing that this is reckless.
• 5:22
  SPEAKER_01: And Mark Zuckerberg made a pretty bold argument back in the middle of October.
• 5:25
  SPEAKER_01: He said the profitable core business and strong credit let Meta fund these costs, and even if its own AI products fail, other companies would gladly pay for all that computing capacity.
• 5:35
  SPEAKER_00: That's a fascinating hedge, essentially saying the worst case is that we become a landlord for compute.
• 5:41
  SPEAKER_00: But I'd push back on that because it assumes demand for that capacity stays strong, even in a world where Meta's own bets don't land.
• 5:50
  SPEAKER_01: And that's exactly the gap SP is pointing at.
• 6:18
  SPEAKER_01: Alphabet, Amazon, and Microsoft all run big cloud businesses.
• 6:22
  SPEAKER_01: When they build a data center, they can rent that capacity to outside customers and earn revenue directly.
• 6:28
  SPEAKER_01: Meta doesn't have a public cloud, so its payoff has to come in directly, mostly through advertising.
• 6:33
  SPEAKER_00: So when Meta spends a dollar on AI infrastructure, the return has to show up as better ad targeting, better ranking of content in the feed, better creative tools for advertisers or new products like its assistant.
• 6:47
  SPEAKER_00: That's a much longer and less certain chain.
• 6:50
  SPEAKER_01: Exactly.
• 6:50
  SPEAKER_01: And some of those returns are real.
• 6:52
  SPEAKER_01: Meta has said AI-driven ranking improves engagement and ad performance.
• 6:57
  SPEAKER_01: The problem is that it's hard to measure, and it's hard to prove that it scales with$250 billion a year in spending.
• 7:04
  SPEAKER_00: Let me play devil's advocate against the agency for a moment.
• 7:07
  SPEAKER_00: Rating agencies are conservative by design.
• 7:11
  SPEAKER_00: They were laid on plenty of things in the past.
• 7:13
  SPEAKER_00: Is it possible they're simply uncomfortable with a new kind of investment cycle they don't fully understand?
• 7:19
  SPEAKER_01: That's a real argument, and some investors are making it.
• 7:22
  SPEAKER_01: David Wagner at Aptus Capital Advisors said the worries about Meta's debt and spending are overstated because the big tech firms have far stronger debt coverage than the broader market.
• 7:33
  SPEAKER_01: That's a fair counterweight.
• 7:34
  SPEAKER_00: And yet, Laura, this didn't happen in a vacuum the same week.
• 7:38
  SPEAKER_00: Alphabet had what was reported as its worst single trading day on record, falling about 12% after its own cash flow turned more negative in the third quarter.
• 7:48
  SPEAKER_01: Right.
• 7:49
  SPEAKER_01: And the financing activity across the sector that week was wild.
• 7:52
  SPEAKER_01: Microsoft raised about$25 billion in a bond sale, and SpaceX closed$40 billion in debt financing.
• 7:59
  SPEAKER_01: Everyone is reaching for capital at the same moment.
• 8:02
  SPEAKER_00: And earlier this year, SP downgraded Oracle to Triple B, which is just one notch above junk territory, partly because of its own AI build-out.
• 8:11
  SPEAKER_00: So there's a pattern of credit analysts getting more skeptical of this spending wave.
• 8:15
  SPEAKER_01: That's what makes this moment feel different to me.
• 8:18
  SPEAKER_01: For two years, the market rewarded anyone who announced a bigger data center budget.
• 8:23
  SPEAKER_01: Now, the people whose job is to ask whether you can pay for it are starting to ask out loud.
• 8:28
  SPEAKER_00: So let's bring it down to the practical level for our listeners, many of whom run technology teams or advise executives.
• 8:36
  SPEAKER_00: What should a business leader actually take away from a credit warning on one of the biggest companies in the world?
• 8:43
  SPEAKER_01: First, the cost of AI infrastructure is real and someone ultimately pays it.
• 8:47
  SPEAKER_01: If the hyperscalers are under pressure to show returns, expect pricing on AI services, models, and compute to reflect that pressure over time, rather than staying artificially cheap forever.
• 8:59
  SPEAKER_00: I'd add that leaders should apply the same discipline internally that SP is applying to Meta.
• 9:06
  SPEAKER_00: Every AI initiative in your organization should have a clear line from the spend to the revenue or the savings, not just a vague promise of transformation.
• 9:17
  SPEAKER_01: Yes, and the lease lesson matters too.
• 9:19
  SPEAKER_01: Plenty of companies sign multi-year commitments for cloud capacity or AI platform seats.
• 9:24
  SPEAKER_01: Those are obligations and they should be modeled like obligations, especially if the usage you are promised doesn't materialize.
• 9:31
  SPEAKER_00: There's also a vendor risk angle here.
• 9:35
  SPEAKER_00: If you're building heavily on any single platform, it's worth understanding the financial health of that provider and asking how their spending pressures might change their pricing, their priorities, or their product roadmap down the line.
• 9:49
  SPEAKER_01: And for investors, the big signal is the divergence between companies that can sell compute to others and companies that can only use it themselves.
• 9:57
  SPEAKER_01: That distinction is going to matter more and more as these budgets grow toward a quarter of a trillion dollars.
• 10:03
  SPEAKER_00: Let's talk about what happens next because a negative outlook usually comes with a timeline.
• 10:08
  SPEAKER_00: What specific things will SP and the market be watching to decide whether this turns into an actual downgrade?
• 10:17
  SPEAKER_01: The biggest one is Meta's spending guidance for 2027, which it hasn't released yet.
• 10:22
  SPEAKER_01: If that number comes in near the$225 billion forecast or higher, the pressure grows.
• 10:28
  SPEAKER_01: If Meta shows restraint, the outlook could stabilize again.
• 10:32
  SPEAKER_00: And the second thing is execution on that stock program starting in January.
• 10:37
  SPEAKER_00: If Meta actually raises that$60 billion in equity, it directly offsets some of the leverage concern, which is something the agency deliberately left out of its base case.
• 10:47
  SPEAKER_01: The third thing is evidence of payoff.
• 10:49
  SPEAKER_01: If Meta can show AI driving real growth and ad revenue, or if its assistance starts generating meaningful income, the argument that monetization lags its peers gets much weaker very quickly.
• 11:00
  SPEAKER_01: That's the number that truly matters.
• 11:02
  SPEAKER_00: I keep coming back to the bigger picture, Laura.
• 11:05
  SPEAKER_00: This is a company with an incredibly profitable advertising business, and it's choosing to bet a huge share of that profit on a technology whose returns are still uncertain.
• 11:18
  SPEAKER_00: That's either visionary or dangerous.
• 11:20
  SPEAKER_01: Probably both, honestly.
• 11:22
  SPEAKER_01: And that's what makes it such a compelling story.
• 11:25
  SPEAKER_01: History is full of infrastructure booms like railroads and fiber optic networks, where the technology changed the world, but many of the companies that funded the build-out never earned their money back.
• 11:35
  SPEAKER_00: That comparison is sobering and it is fair.
• 11:37
  SPEAKER_00: The technology can succeed while the people who paid for it do not.
• 11:41
  SPEAKER_00: SP is asking which side Meta will end up on.
• 11:44
  SPEAKER_01: And the honest answer is nobody knows yet, including Meta.
• 11:48
  SPEAKER_01: What we do know is that the era of AI spending without scrutiny is ending, and the people holding the bonds are now asking the same questions as the people holding the stock.
• 11:57
  SPEAKER_00: That's a great place to wrap up our deep dive.
• 12:00
  SPEAKER_00: The headline is a credit outlook, but the real story is accountability arriving for the biggest technology investment cycle in a generation.
• 12:09
  SPEAKER_00: Laura, thanks for bringing all those numbers today.
• 12:11
  SPEAKER_01: Always a pleasure, Rick.
• 12:12
  SPEAKER_01: And to everyone listening, keep an eye out on Meta's spending guidance and that January stock program, because those two things will tell us a lot about where this entire AI spending wave is heading next.
• 12:23
  SPEAKER_00: That's all for today's episode of the DX Today podcast.
• 12:26
  SPEAKER_00: Thanks for listening, and we'll see you next time.