Meta plans to issue $25 billion in debt to back AI: In 2026, capital expenditures are set to reach $145 billion

Meta Platforms completed a $25 billion investment-grade bond issuance on April 30, its second major debt offering within six months. The proceeds are mainly for funding its 2026 AI capital expenditure plan, which has been revised up to $125-145 billion. This issuance is divided into six tranches, with the longest-maturity long-dated notes due in 2066. Initial pricing was about 1.8 percentage points higher than the U.S. Treasury yields of the same period. After Meta released its Q1 earnings on 4/30, its stock price plunged 7%. Even so, the deal still attracted about $96 billion in subscription orders, but that is below the previous $125 billion level, suggesting growing hesitation in the market about Meta’s AI capital expenditure strategy.

$2.5 billion issuance in six tranches: longest due in 2066, priced about 1.8 percentage points above Treasury yields

Meta designed this bond offering with six tranches, spanning maturities from intermediate-term notes to the long-dated tranche due in 2066. The most closely watched is the 40-year tranche, priced initially at a spread about 1.8 percentage points (180 basis points) above the yield of the corresponding-maturity U.S. Treasury. While this spread remains within the investment-grade range, it is significantly wider than the spread from Meta’s prior $3 billion issuance in October 2025, indicating that bond investors have started to mark up the pricing of Meta’s credit risk.

In terms of overall subscription multiple, the peak subscription amount for this deal was $96 billion, and at final pricing most of the six tranches were priced at a “risk premium higher than the October issuance.” Compared with the $125 billion subscription amount from the prior $3 billion issuance, this offering is smaller in size but comes with higher “funding tightness.” With the same $2.5 billion asset size, Meta could only attract $96 billion in orders, reflecting that the bond market’s acceptance of “financing Meta’s AI capital expenditures at low interest rates” is starting to decline at the margin.

AI capital expenditure further revised upward: $125-145 billion, Zuckerberg admits “no precise product plan”

Meta’s 4/30 Q1 earnings release also revised its estimated full-year 2026 capital expenditures upward, increasing the range from the previous $115-135 billion to $125-145 billion, implying an additional $10 billion in AI investment. At the earnings call, Mark Zuckerberg made a rare admission: “there is currently no precise AI plan laid out product by product.” This candor instead triggered investor concerns—when the company cannot clearly explain which specific product lines the $140 billion spending will go into, the risk premium for whether the “burning cash can translate into returns” rises.

Meta’s stock fell 7% after the 4/30 earnings report (the largest single-day drop in six months). Meanwhile, Alphabet’s stock rose 6% on the same period due to positive earnings news. Both companies revised up AI capital expenditures substantially, yet the market delivered sharply different reactions. The key difference is that Alphabet’s Cloud business has already shown concrete monetization with quarterly revenue of $20 billion, while Meta’s advertising business has not demonstrated that AI investment can directly lift ad ROI.

What to watch next: AI ROI signals, bond issuance density, changes in subscription structure

The key metric for Meta in the next few quarters is whether “AI investment can translate into improved advertising ROI.” Reels recommendation algorithms, Threads push notifications, and ad matching systems are potential beneficiary areas for AI investment. However, the current Q1 earnings do not show any structural improvement in ad CPM or ROI. If no monetization signal is seen in Q2, the high spread on this bond issuance may be more than just short-term volatility—it could indicate that the market has begun to systematically doubt whether an “AI arms race” will overwhelm free cash flow.

For the bond market, after Meta, Alphabet, Amazon, and Microsoft all revised up AI capital expenditures, a wave of technology bond issuance totaling $400-500 billion is expected over the next 12 months. Meta’s Q1 earnings already revealed dual pressures: capital expenditures surging and user numbers falling for the first time. If the whole industry issues debt in sync, bond spreads could widen further and force the AI cash-burning race into the next phase of “subduing costs of debt.”

This article about Meta’s $2.5 billion bond issuance to support AI: 2026 capital expenditure to push toward $145 billion first appeared on Chain News ABMedia.