Credit Update: META offers 5+% to 6+% yields (USD) while it ramps up its AI monetisation.

Key Points

Data as of 31 March 2026
Source: Company Data, iFAST Compilations. 

Data as of 31 March 2026
Source: Company Data, iFAST Compilations.

Capex Pain Today, Monetisation Gain Tomorrow 

The central bear case on Meta’s credit profile is the scale of the AI capex ramp – 2026 capex is expected to come in around US$125-$145b (as shown in chart 3 below), representing a double, compared to 2025’s US$72.2b, alongside persistent ~US$19b of annual RL operating losses. Near-term, this pressure is real: in Q126, total expenses rose 35% YoY, outpacing revenue growth of +33%, due to rising data centre and cloud infrastructure costs. Management has been explicit that the spend trajectory points higher still. While Susan Li declined to guide 2027 capex, she noted that Meta has “continued to underestimate” its compute needs, and Mark Zuckerberg has publicly committed to roughly US$600b in US data centre and AI infrastructure investment by 2028.

 
However, we think that this heavy investment spending, while it will generate short-term pain in terms of FCF, should translate into structural monetisation gains down the road. This is already playing out in Q126 metrics, where price per advertisement rose 12% YoY, with impressions following suit at +19% YoY. These increases are increasingly driven by specific AI systems that compound rather than simply showing users more advertisements: Meta’s rebuilt advertising engine, Andromeda, now sifts through roughly 10,000x more candidate advertisements per impression than before, while newer ranking models (GEM, Adaptive Ranking) lifted advertisement conversion rates by ~6% on certain formats in Q126 alone. The same AI ranking improvements are also driving engagement gains in the content feed, with 10% increase in time spent on Instagram Reels following content recommendation enhancements, which in turn expands the inventory available for monetisation. We highlight this as a crux: Meta is making each advertisement more likely to convert, not just expanding inventory, which makes advertisers willing to pay more for the same space. Once the AI infrastructure is built, its cost is largely fixed, so further conversion gains flow disproportionately to profit. Hence, today’s capex pain converts into tomorrow’s margin engine; a trend that is clearly seen in chart 4 below.

 
Looking forward, management has reiterated that 2026 operating income will exceed 2025’s US$83.3b despite the capex ramp, implying earnings power continues to compound even through this phase of heavy investment. We think the real test for Meta’s credit profile is not the absolute capex figure but whether AI-driven pricing keeps outpacing the depreciation that capex creates; today, it comfortably does. 

Chart 3: Rising capex trajectory

Data as of 31 March 2026
Source: Company Data, iFAST Compilations. 

Chart 4: Operating margin has been on an increasing trend since FY22

Strong credit profile, but off-balance sheet arrangements are a key watch

Historically, Meta has operated with one of the strongest balance sheets among technology issuers: as seen in chart 5 below, over the last five years, OCF has compounded from US$57.7b (FY21) to US$115.8b (FY25) at a 19% compound annual growth rate (CAGR), and free cash flow has grown from US$38.1b to US$43.6b over the same period despite capex more than tripling (see chart 5 below). That said, we remain mindful that Meta could print negative free cash flow for FY26 and even FY27 due to the group’s large capex spending.

As of 31 March 2026, Meta held US$81.2b in cash and short-term marketable securities against total debt of US$86.8b (US$28b in operating leases and US$58.7b in long-term debt). On a pro-forma basis incorporating US$25b of bonds issued on 4 May 2026, total debt rises to US$111.8b and long-term debt to US$83.7b. As seen in Table 2 below, although leverage has picked up, net debt/equity is still minimal at 2.3% (31 March 2026). Interest coverage (EBITDA / gross interest cost) paints a similar picture, with the TTM ratio remaining robust at 73.6x. As seen in chart 6 below, META’s debt maturity profile is comfortably staggered across the next 40 years, which presents low refinancing needs.

More importantly, Meta’s off-balance-sheet (OBS) exposure is the key credit watch point for the group’s credit profile. As seen in Table 2 below, OBS commitments have exploded– from US$289b as of 31 December 2025 to over US$500b pro-forma accounting for April 2026 additions. This stack consists of (1) the Hyperion data centre joint venture (JV) with Blue Owl Capital and PIMCO (US$46.0b of maximum exposure to loss, including a US$28.0b residual value guarantee), (2) US$182.9b of leases not yet commenced (mostly data centres commencing 2026-2036), (3) US$237.7b of non-cancellable contractual commitments relating to cloud and network infrastructure, and (4) a new US$5.0b restricted cash escrow related to a multi-year purchase agreement.

We highlight three relevant credit observations regarding Meta’s OBS exposure. Firstly, all these commitments are recourse to Meta’s cash on a deferred basis. Secondly, the Hyperion deal was flagged by Meta’s auditor as a critical audit matter, and Meta’s US$1.5b of held-for-sale data centre confirms a second JV is imminent, implying the OBS stack will keep growing. Finally, capitalising these OBS lease-equivalent commitments (US$182.9b) and the residual value guarantee of US$28.0b, would lift adjusted leverage (net debt/equity) materially to roughly 89%, a fundamentally different credit profile than the reported GAAP figure (2.3%) suggests.   

Chart 5: Cash flows

Data as of 31 March 2026
Source: Company Data, iFAST Compilations. 

Chart 6: Debt Maturity Profile 

Data as of 31 March 2026
Source: Company Data, iFAST Compilations. 

Table 2: Key credit metrics & off-balance-sheet exposure

Overall, we think META possess a comfortable investment-grade credit profile, supported by strong operating cash flows (US$124.0b on a TTM basis) and a solid total available liquidity (US$81.2b). However, with management explicitly biased toward continued capex escalation with FCF expected to be negative for 2027-2028, alongside its OBS commitments (which are expected to pick up as well), we expect the group’s credit profile to soften moving forward. That said, we remain comfortable with Meta’s credit profile, anchored by its world-class suite of social media products and the increasing positive impact from its AI initiatives.

In Table 4 above, we highlight a couple of META outstanding bonds for consideration: these issues offer yields-to-worst ranging from 5.19% to 6.19%, with tenors ranging from 9.46 years to 29.47 years. Against comparable US treasuries, we find a compelling yield spread of 75bps to 123bps. Comparing against its US technology peers (Amazon, Google, Microsoft), these bonds present a decently attractive yield pickup ranging from 40+bps to 70+bps. Although we note most of these issues have a higher credit rating than Meta’s bonds; conversely, Oracle’s bonds provide higher yield pickups compared to Meta’s issues due to their softer credit rating.

Investors looking for solid income from a quality name can consider these outstanding Meta bonds. That said, we emphasise the duration risk embedded in its longer tenor bonds (19+y to 29y) which could face steep price declines if interest rates rise further.

 

Declaration: For specific disclosure, at the time of publication of this report, IFPL (via its connected and associated entities) holds positions in META 5.500% 15Nov2045 Corp (USD), AMZN 5.450% 20Nov2055 Corp (USD), ORCL 5.875% 26Sep2045 Corp (USD), and GOOGL 4.500% 15May2035 Corp (USD). The analyst who produced this report hold NIL positions in the abovementioned securities. This research report was prepared with the assistance of artificial intelligence (AI) tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation; the content of this report – including all investment theses, ratings, price targets and conclusions – has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.