Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Meta is financing an unusually large AI infrastructure build-out with operating cash flow, existing liquidity, new debt and long-term contractual commitments. Total capital expenditures reached $72.22 billion in 2025, while management’s September 2026 outlook calls for $135 billion to $145 billion in 2026. Because Meta does not report an AI-only capex subtotal, these figures also include infrastructure for its advertising, social and messaging businesses.
The financial effect is already visible: Meta generated $115.80 billion of operating cash flow and $43.59 billion of free cash flow in 2025, but second-quarter 2026 free cash flow fell to just $784 million after $31.08 billion of capex. Long-term debt rose from $58.74 billion at the end of 2025 to $83.66 billion on June 30, 2026. Liquidity remains substantial, so the immediate issue is not insolvency; it is whether spending, leverage and fixed commitments grow faster than AI-driven returns and advertising cash generation.
As an Amazon Associate I earn from qualifying purchases.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →How much is Meta spending on AI infrastructure?
There is no disclosed line item that isolates AI spending. The most defensible public proxy is total capital expenditure, which covers data centers, servers, networking equipment and other infrastructure serving both AI workloads and Meta’s established products. That distinction matters: treating every dollar of capex as AI spending would overstate the amount devoted exclusively to model training or inference.
| Period | Reported figure | What it shows |
|---|---|---|
| Full year 2025 | $72.22 billion capex | Actual infrastructure investment reported in Meta’s 2025 Form 10-K |
| Full year 2025 | $115.80 billion operating cash flow | Cash produced before capital spending |
| Full year 2025 | $43.59 billion free cash flow | Operating cash flow less capex |
| 2026 outlook | $135 billion-$145 billion capex | Management guidance filed in September 2026; includes core and AI infrastructure |
| Second quarter 2026 | $31.08 billion capex | Quarterly cash investment in property and equipment |
The 2026 guidance implies a major acceleration from 2025’s full-year spending. It is guidance rather than a completed result, and the eventual amount can change with chip availability, construction schedules, supplier terms and demand. For readers asking “How much is Meta spending on AI?”, the accurate answer is therefore a range based on total infrastructure spending, not a separately audited AI budget.
#1 Best Overall
Is Meta’s AI spending hurting free cash flow?
Yes. The clearest evidence is the gap between operating cash flow and capital expenditure. In the second quarter of 2026, Meta produced $31.86 billion of operating cash flow and spent $31.08 billion on capex, leaving only $784 million of free cash flow.
Why the quarterly squeeze is so large
Free cash flow is the cash left after the company pays for property and equipment. A profitable advertising business can therefore show strong operating cash flow while producing little cash available for debt repayment, share repurchases, acquisitions or a larger liquidity reserve once data-center construction and equipment purchases are paid.
Meta’s 2025 numbers illustrate the same mechanics at an annual level. Operating cash flow of $115.80 billion exceeded $72.22 billion of capex by $43.59 billion. The second-quarter 2026 result is more compressed because spending is arriving at a much higher run rate. One quarter does not establish a permanent trend—capex can be lumpy—but it demonstrates how quickly AI-related infrastructure can absorb cash generated by the existing business.
What free-cash-flow pressure does and does not mean
- It does mean less internally generated cash is available for discretionary uses while the build-out is at peak intensity.
- It does not mean Meta is losing money or unable to pay current bills; operating cash flow remained strongly positive.
- It does not prove that all capex is unproductive. Capacity can support better ad ranking, recommendation systems, engagement, safety tools and new AI products over several years.
- It does increase sensitivity to delays, underused data centers, chip obsolescence, weaker advertising or higher financing costs.
How much debt is Meta taking on for data centers and AI?
Meta’s reported debt increased substantially during the first half of 2026. Long-term debt was $58.74 billion on December 31, 2025, compared with $83.66 billion on June 30, 2026. The June 2026 filing also reported $84.00 billion of notes outstanding and $4.40 billion of short-term future interest payments. Reuters reported that Meta completed a $25 billion investment-grade bond sale in April 2026.
| Date or period | Debt or financing measure | Qualification |
|---|---|---|
| December 31, 2025 | $58.74 billion long-term debt | Year-end balance reported by Meta |
| June 30, 2026 | $83.66 billion long-term debt | Six-month increase shown in Meta’s second-quarter release |
| June 2026 filing | $84.00 billion notes outstanding | Face amount of notes disclosed in the Form 10-Q |
| June 2026 filing | $4.40 billion future interest payments due within the short term | Contractual interest obligation disclosed by Meta |
| April 2026 | $25 billion bond sale | Investment-grade issuance reported by Reuters |
Debt is only one part of the financing picture. Large data-center programs can also involve operating and finance leases, equipment purchase commitments, take-or-pay arrangements, guarantees and other supplier contracts. Those obligations may not appear as conventional borrowing on the same timetable, but they still create fixed or semi-fixed cash requirements. Investors should read the commitments and contingencies sections of each filing alongside the debt footnotes.
Rank #2
Can Meta fund the AI build-out without borrowing more?
Meta can fund a substantial portion internally when advertising cash flow remains strong, but the 2026 spending plan makes additional financing more likely if management wants to preserve cash balances and other uses of capital. The answer depends on the relationship between three moving parts: operating cash flow, capex timing and the amount of liquidity management is willing to keep on hand.
Internal funding capacity
At year-end 2025, Meta’s $43.59 billion of free cash flow provided a meaningful self-funding cushion after capex. That cushion is not a fixed annual entitlement. If capex approaches the 2026 guidance range while operating cash flow grows more slowly, free cash flow could fall sharply or turn negative for periods. A company can still pay for projects from cash reserves, but doing so reduces its buffer for advertising volatility, litigation, acquisitions and maturities.
Why issuing debt can be rational
Borrowing allows Meta to spread the cost of long-lived infrastructure across the years in which that infrastructure is expected to generate value. It also avoids liquidating investments or cutting every other capital-allocation program at the exact moment AI capacity is strategically important. The trade-off is interest expense, refinancing exposure and a larger group of creditors with contractual claims ahead of equity holders.
What would make borrowing riskier
- Operating cash flow stagnates or declines while capex remains near the upper end of guidance.
- Advertising growth weakens during an economic slowdown, reducing the cash engine that supports infrastructure spending.
- Facilities are completed before demand, utilization or monetization catches up.
- Specialized chips become obsolete faster than expected, forcing replacement spending or impairment charges.
- Higher interest rates raise the cost of refinancing or new debt.
Does Meta have enough liquidity to cover its commitments?
Near-term liquidity is strong on the reported figures. Meta held $81.59 billion of cash, cash equivalents and marketable securities at December 31, 2025, rising to $90.26 billion on June 30, 2026. Its June 2026 filing says available funds, operating cash flow and financing should cover operational needs, infrastructure and AI commitments for the next 12 months and thereafter for the foreseeable future.
That statement is a liquidity assessment, not a promise that spending will be painless. Cash and marketable securities are a gross resource; they must be considered against debt maturities, interest, leases, purchase commitments, payroll, taxes and possible legal costs. A rising cash balance can coexist with rising debt if the company is deliberately financing an expansion while retaining a safety reserve.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Could AI spending threaten Meta’s credit rating?
AI spending raises credit risk through weaker cash-flow cushions and higher fixed commitments rather than through an immediate solvency crisis. S&P Global Ratings said financing for AI infrastructure is increasingly using debt, leases, guarantees and other structures, and projected negative free operating cash flow across six major hyperscalers in 2026-27. It also estimated more than $1.3 trillion of combined hyperscaler AI infrastructure investment by 2027.
For Meta, a credit analyst would focus on whether leverage and interest obligations rise faster than recurring operating cash flow, and whether AI investments produce measurable commercial benefits. A rating agency could tolerate temporarily lower free cash flow if it expects strong earnings, ample liquidity and disciplined capital spending. The risk becomes more acute if spending commitments are difficult to cancel and the advertising business weakens at the same time.
Indicators that would signal a deteriorating credit cushion
- Free cash flow after capex remains close to zero or negative for several consecutive quarters.
- Debt, lease liabilities, guarantees or take-or-pay commitments grow faster than operating cash flow.
- Interest expense rises materially without a corresponding increase in earnings before interest and taxes.
- Cash and marketable securities decline while infrastructure commitments continue to expand.
- Management repeatedly raises capex guidance but provides limited evidence of improved monetization or utilization.
Current issuer ratings from Moody’s and Fitch are not stated here because a verified, current rating is required for a reliable comparison. Ratings can change independently of one quarter’s cash flow and should be checked in the agencies’ latest releases.
How can investors judge whether the spending is paying off?
Financial statements rarely label a single dollar of capex as “AI revenue.” A practical assessment combines cash-flow data with operating evidence. Look for improvements in ad ranking and conversion, recommendation quality, user engagement, ad pricing, creator activity, business messaging and adoption of new AI features. These indicators do not prove a specific return on a data center, but together they show whether infrastructure is supporting the products that generate cash.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches| Question | Metric to monitor | Why it matters |
|---|---|---|
| Is spending affordable? | Operating cash flow minus capex | Shows how much cash remains after infrastructure investment |
| Is leverage rising? | Debt and interest obligations relative to operating cash flow | Measures the burden of fixed claims on future cash generation |
| Is liquidity adequate? | Cash and marketable securities less near-term maturities and commitments | Tests the available buffer during a downturn |
| Is AI creating value? | Ad performance, engagement, monetization and adoption of AI features | Provides operating evidence behind the investment case |
| How exposed is the plan? | Capex guidance changes, supplier commitments and utilization | Highlights execution, obsolescence and underused-capacity risk |
What are the main downside scenarios?
Advertising slowdown
Meta’s advertising business supplies most of the operating cash that funds infrastructure. A cyclical decline in ad demand, pricing or conversion would reduce cash generation just as data-center and equipment commitments continue.
Underutilized capacity
If model demand, user adoption or product monetization develops more slowly than expected, Meta could carry expensive facilities and accelerators that produce less revenue than planned. Fixed depreciation, leases and financing costs would remain.
Technology turnover
AI hardware evolves quickly. A faster replacement cycle can force additional purchases before existing equipment has delivered its expected economic return, increasing capex and potential impairment risk.
Complex financing structures
Debt is visible, but leases, guarantees and take-or-pay contracts can make obligations harder to assess. S&P’s warning about increasingly complex financing across hyperscalers is relevant because headline debt may understate the total set of claims on future cash.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallExecution and regulatory costs
Data-center construction, power availability, supply-chain constraints, privacy requirements and legal costs can delay benefits or increase the cost of deploying AI. Delays matter financially when interest and contractual payments begin before revenue benefits arrive.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What should readers watch in Meta’s next filings?
- Compare quarterly operating cash flow with capex and calculate the resulting free cash flow.
- Track total debt, notes, interest payments and the maturity schedule rather than looking only at one debt line.
- Read commitments, leases, guarantees and purchase obligations for fixed payments not captured by headline borrowing.
- Compare actual capex with the $135 billion-$145 billion 2026 guidance and note any revisions.
- Measure cash, cash equivalents and marketable securities against debt and near-term commitments.
- Look for operating evidence that AI improves advertising performance, engagement or monetization.
- Check whether management explains utilization, expected returns and the conditions under which spending would slow.
These checks distinguish a temporary investment trough from a structural weakening of credit quality. The central question is not whether Meta can spend billions on AI—it clearly can—but whether recurring cash generation and commercial returns keep pace with the obligations created by that spending.
Best Value
Frequently Asked Questions
Does Meta disclose a separate AI capex number?
No. Its filings combine AI infrastructure with capacity for advertising, social, messaging and other products, so total capex is only a proxy. Any precise AI-only dollar figure would require information Meta does not currently report.
Why can free cash flow be low when Meta is profitable?
Free cash flow subtracts property and equipment purchases from operating cash flow. A profitable company can therefore report strong earnings and operating cash while a data-center construction cycle consumes nearly all of that cash.
Are leases part of Meta’s AI debt risk?
They can be. Data-center leases, equipment contracts, guarantees and take-or-pay arrangements create future cash obligations even when they are not presented as conventional long-term debt, so the commitments footnotes matter.
What would make Meta’s AI spending safer for creditors?
Credit risk would improve if operating cash flow grew faster than capex, debt stabilized, liquidity stayed ample and AI-driven advertising or product revenue became more visible. Evidence of flexible spending plans would also reduce the danger of being locked into expensive capacity during a downturn.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
