Big Tech AI Capex and Q3 2026 Earnings Results

Microsoft stood out among hyperscalers in late October 2026 earnings with accelerating Azure growth (48% YoY to $33.17B in its fiscal Q1 ended September 30), positive free cash flow of $18.6B (well above the ~$8.5B Visible Alpha estimate), and capex of $45.2B that came in below its own ~$50B projection due to third-party capacity delays—signaling that its AI infrastructure investments are already translating into revenue and cash generation while peers burn cash.[1][2]

• Microsoft Cloud revenue hit $63.9B (+30% YoY); commercial RPO surged 73% to $677B; Azure guidance for fiscal Q2 was 49-50% growth (above ~47.6% Street estimates).[3]
• Calendar 2026 capex guidance lowered slightly to ~$170B (from ~$175B earlier); Q2 capex expected >$50B with acceleration in H2 from GPUs/CPUs and delayed 3P deliveries.[4]
• Stock reaction: Shares rose ~4% in after-hours trading post-earnings on the cloud beat and FCF resilience, though the broader sector later pressured names.[5]
• Analysts viewed this as validation of Microsoft’s data moat and balance-sheet strength versus debt-reliant peers; some noted creative accounting on leases but highlighted sustained demand exceeding supply.[6]

This positions Microsoft as relatively insulated in the AI capex race—its productivity/cloud ecosystem converts infrastructure spend into high-margin recurring revenue faster than pure infrastructure plays, giving it an edge for competitors trying to match scale without similar cash-flow buffers.

Alphabet’s Google Cloud exploded to $30.7B (+103% YoY) with operating income tripling to $11.1B (36.2% margin), yet capex of $51.4B exceeded operating cash flow of $42.3B, producing negative FCF of ~$9.1B and triggering a sharp stock selloff—highlighting how even explosive cloud/AI demand can pressure cash generation when spending outpaces monetization timelines.[7][8]

• Q3 2026 consolidated revenue $127.6B (+25% YoY); TPU system sales contributed meaningfully and are expected to ramp further in 2027.[9]
• FY2026 capex guidance held at $195-205B; 2027 expected to rise significantly (details in Q4). Google Cloud backlog ~$561B (55%+ recognizable in 24 months).[10]
• Stock reaction: Shares fell ~10-12.6% (worst one-day drop on record per some reports), erasing hundreds of billions in market value amid earnings miss, negative FCF, and capex concerns.[11]
• Analysts and commentators focused on the cash-burn signal and near-term margin pressure from third-party compute ramp; some saw TPU progress as a long-term Nvidia alternative but questioned near-term ROI visibility.[12]

For entrants or competitors, Alphabet’s results underscore the risk of aggressive capex outrunning cash flow even with hyper-growth in cloud—success depends on converting infrastructure into diversified AI revenue (e.g., TPU sales, agents) before debt or dilution pressures mount.

Amazon’s AWS delivered 39% growth to $45.98B (fastest in nearly five years, beating estimates) with operating income up 55% to $17.8B, but Q3 property/equipment purchases hit $77.3B (well above ~$58.7B Street est.), driving TTM FCF to a $42.4B outflow and a stock decline despite the cloud strength.[13][14]

• Total Q3 revenue $201.9B (+12%); 2026 cash capex guidance held at ~$220B; 2027 expected “meaningfully higher”; AWS backlog $578B (more than double YoY); Jassy highlighted potential for two $1T revenue businesses (AWS + another) in 5-10 years and >40% AWS growth in 2027-2028.[15]
• Raised $48.1B in long-term debt in the quarter; can borrow up to $80B against AI hardware.[13]
• Stock reaction: Fell ~7% post-earnings on capex magnitude, retail guidance softness, and sector-wide AI spend fatigue.[16]
• Commentators noted AWS acceleration as a positive but flagged the scale of cash burn and questions around sustained returns; some highlighted customer prepayments or vendor financing as mitigators seen elsewhere (e.g., Oracle).[17]

Amazon illustrates how even dominant cloud positions require ever-larger absolute spend to maintain growth; competitors must either achieve similar scale efficiencies or differentiate on cost/niche AI workloads to avoid matching this burn rate.

Meta reported Q3 revenue of $64.6B (+26% YoY) with ad strength driving results, but capex of $34.9B (incl. finance leases) exceeded operating cash flow, yielding negative FCF of $2.1B; it narrowed 2026 capex to $135-145B (raising the floor) and introduced a non-voting Class C stock class to fund AI without diluting control.[18][19]

• Reality Labs losses guided to $18-19B in 2026 and ~$15B in 2027; AI investments already boosting ad engagement/monetization.[20]
• Stock reaction: Mixed/initially positive on revenue beat but pressured by FCF and spending narrative; broader sector weakness weighed later.[21]
• Analysts saw the new share class and narrowed capex as pragmatic funding moves but noted longer ROI timelines for frontier AI (e.g., Muse model) versus core ads.[22]

Meta’s approach—leveraging high-margin ad cash flow to subsidize AI while using equity tools to preserve founder control—offers a template for non-hyperscalers, but it requires proven ad uplift from AI to justify the spend trajectory.

Nvidia had no new earnings in the October window (next expected ~Nov 17); Oracle (Q1 FY2027 reported earlier in September) showed Cloud Infra +121% to $7.39B, RPO +$209B YoY to $664B (including >$30B new AI contracts), and capex of $28.5B largely offset by $11.36B customer prepayments, with FY2027 revenue guidance >$90B and plans to turn FCF positive in H2 FY2028 while limiting additional AI financing to $20B.[23][24]

Overall market read: Earnings revealed accelerating AI/cloud demand across the board (Azure 48%, Google Cloud 103%, AWS 39%, Oracle Cloud Infra 121%) but widespread capex escalation turning once-strong FCF generators into net cash users or near-breakeven, sparking sector-wide selloffs (hundreds of billions in market value erased) and questions about ROI timelines, debt reliance, and supply constraints.[25][16] Shares of neoclouds and related names also pressured. Analysts split between “spend now to win later” optimism (citing backlog/RPO surges and demand > supply) and caution on cash burn/depreciation ramp.[12]

Implication for competitors/entrants: The bar has risen—matching hyperscaler capex requires either massive existing cash flows (Microsoft/Meta model), customer/vendor financing/prepayments (Oracle/Amazon elements), or differentiation via efficiency/specialization (e.g., TPUs, niche clouds). Delays in capacity or slower monetization can trigger sharp valuation resets.

Key Metrics Comparison Table (Recent Reported Quarter vs. Prior-Year Quarter and Guidance Where Available)

Note: Periods differ slightly by fiscal calendars (Microsoft fiscal Q1 FY2027 ended Sep 30; Alphabet/Amazon Q3 calendar 2026 ended Sep 30; Meta Q3 2026 ended Sep 30; Oracle Q1 FY2027 ended Aug 31). Figures primarily from company releases and contemporaneous reporting.[1][7][13][18][24]

• Microsoft (Q1 FY2027): Revenue $92.3B (+19%); Azure $33.17B (+48%); Cloud $63.9B (+30%); Capex $45.2B (vs. $34.9B prior); FCF $18.6B (positive, beat est.); RPO $677B (+73%); Guidance: Q2 Azure 49-50%, CY2026 capex ~$170B.
• Alphabet (Q3 2026): Revenue $127.6B (+25%); Google Cloud $30.7B (+103%); Capex $51.4B (vs. $24.0B); FCF -$9.1B (negative); Cloud backlog ~$561B; Guidance: FY2026 capex $195-205B (unchanged), 2027 significantly higher.
• Amazon (Q3 2026): Revenue $201.9B (+12%); AWS $45.98B (+39%); Capex (PPE) $77.3B (vs. $35.1B); TTM FCF -$42.4B; AWS backlog $578B; Guidance: 2026 capex ~$220B, 2027 meaningfully higher, AWS >40% growth in 2027-28.
• Meta (Q3 2026): Revenue $64.6B (+26%); Capex $34.9B (incl. leases, vs. ~$19.4B); FCF -$2.1B; Guidance: FY2026 capex $135-145B (narrowed floor), Reality Labs losses narrowing.
• Oracle (Q1 FY2027): Revenue $19.3B (+30%); Cloud Infra $7.39B (+121%); Total Cloud $11.6B (+62%); RPO $664B (+$209B YoY); Capex $28.5B (offset by prepayments); Guidance: FY2027 revenue >$90B, capex $90-95B, FCF positive H2 FY2028.

Depreciation trends: Rising across the board with capex (e.g., Meta +44% YoY in one report; hyperscalers noting higher infra depreciation and energy costs pressuring near-term margins). Debt raised: Notable at Amazon ($48B+ in quarter) and Meta (new equity class + prior debt growth); Oracle limiting additional AI financing.[26]

These results collectively show AI infrastructure spend topping prior peaks with no near-term slowdown signaled, but investor tolerance for negative or declining FCF is waning—favoring those demonstrating fastest revenue conversion.

---

Recent Findings Supplement (November 2026)

Microsoft delivered a standout October 28, 2026, earnings beat on fiscal Q1 2027 (ended September 30), with Azure growth accelerating to 48% and free cash flow far exceeding expectations despite heavy AI infrastructure outlays. The company attributed lower-than-guided capex to third-party GPU/data center delivery delays rather than softening demand, which remains supply-constrained, and trimmed its full-year 2026 capex outlook while signaling acceleration ahead.[1][2]

• Revenue reached $92.3 billion (+19% YoY), with Microsoft Cloud at $63.9 billion (+30%) and Azure specifically at $33.17 billion (+48%, beating estimates); overall operating income rose 21% to $45.9 billion.[1][3]
• Capex totaled $45.2 billion (vs. ~$50 billion prior projection and $34.9 billion YoY), with the shortfall tied to delayed third-party capacity (some GPU-as-a-Service contracts pushed to H2 FY2027); calendar 2026 guidance lowered to ~$170 billion from $175 billion.[2][4]
• Free cash flow hit $18.6 billion (well above ~$8.5 billion Visible Alpha estimate), supported by strong cloud collections; Q2 capex expected above $50 billion.[1][5]
• Commercial remaining performance obligation (backlog/RPO proxy) jumped 73% to $677 billion; commercial arrangements with OpenAI contributed ~$10 billion (~11% of quarterly revenue).[6][7]
• Stock rose ~3% in after-hours trading. Analysts highlighted the FCF resilience and demand-supply imbalance as evidence AI bets are paying off, distinguishing Microsoft from peers facing sharper cash flow squeezes.[1][8]

This positions Microsoft favorably for competitors: its scale, existing capacity efficiencies, and diversified revenue provide a buffer against pure infrastructure spend risks, raising the bar for pure-play or less-mature AI/cloud players to demonstrate similar monetization velocity.

Alphabet reported Q3 2026 results on October 28 showing Google Cloud more than doubling (+103% to $30.7 billion) on AI infrastructure, TPU system sales, and enterprise demand, but this came with record quarterly capex that drove negative free cash flow and a slight revenue miss. Guidance for 2026 capex held steady amid plans for significant 2027 increases and near-term margin pressure from third-party compute.[9][10]

• Consolidated revenue $127.6 billion (+25%); Google Cloud operating income $11.1 billion (margin 36.2%, up sharply YoY); total backlog ~$567.7 billion (~$561 billion Google Cloud).[11][12]
• Capex (purchases of property and equipment) $51.4 billion (record; vs. $24 billion YoY), exceeding operating cash flow of $42.3 billion and producing FCF of -$9.1 billion (TTM FCF down 73% to $19.7 billion).[9][13]
• 2026 capex guidance unchanged at $195–205 billion; expects “significant” increase in 2027; began utilizing third-party compute late Q3 (ramp in Q4) with expected near-term margin headwinds from higher depreciation and energy costs.[11][13]
• Shares fell sharply (reports of ~10% decline) on the FCF print, revenue miss, and spending trajectory despite the cloud strength.[10]

For market entrants or rivals, Alphabet’s results underscore the execution risks of rapid AI scaling—strong demand validation but immediate cash flow compression and reliance on external capacity—potentially favoring integrated players with better-matched supply/demand timing.

Amazon’s October 29 Q3 2026 report highlighted AWS accelerating to its fastest growth in nearly five years (+39.3% to $46 billion, $184 billion annualized run rate) with a $578 billion backlog, but record capex produced sharply negative trailing FCF and prompted debt issuance. The company held 2026 capex guidance while signaling meaningfully higher spending in 2027.[14][15]

• Net sales $201.9 billion (+12%); operating income $28.1 billion (record margin 13.9%); AWS operating income $17.8 billion (38.7% margin).[15]
• Capex (purchases of property and equipment) $77.3 billion (vs. $35.1 billion YoY and well above estimates); TTM FCF swung to -$42.4 billion outflow (vs. +$14.8 billion prior year), driven by AI-related property/equipment investments.[14][16]
• Raised $48.1 billion in long-term debt in the quarter; 2026 cash capex guidance held at ~$220 billion; expects “meaningfully more” in 2027.[14][17]
• Shares fell ~6.5% after hours amid cash burn concerns, despite AWS momentum and CEO commentary on capacity being “gobbled up” as fast as installed.[18]

This highlights a key competitive dynamic: AWS’s scale and backlog provide visibility, but the cash flow math intensifies pressure on all hyperscalers to convert AI infrastructure into higher-margin, faster-monetizing services or risk prolonged negative FCF periods that smaller or less-capitalized players cannot sustain.

Meta’s October 21 Q3 2026 results showed revenue strength (+26% to $64.6 billion) alongside AI-driven capex that pushed quarterly FCF negative for the first time in recent periods, prompting a narrowed-but-raised 2026 capex range and a new non-voting share class to fund growth without diluting control.[19][20]

• Capex (including finance lease principal payments) $34.91 billion (vs. ~$19.4 billion YoY); FCF -$2.12 billion (vs. +$10.6 billion YoY).[21]
• 2026 capex guidance narrowed/raised to $135–145 billion (higher floor); Reality Labs losses expected to narrow further in 2027.[20]
• Announced Class C non-voting shares (effective 2027) explicitly to support AI acquisitions, talent, and infrastructure financing while preserving voting control.[20][22]
• Shares rose >4% after hours on the revenue beat and Q4 outlook.[20]

Meta’s moves illustrate how AI capex is reshaping capital structures even for high-margin ad businesses, creating opportunities (and dilution risks) around novel equity instruments for funding-intensive AI plays.

Oracle provided fresh October 28, 2026, Investor Day updates on its AI/cloud infrastructure buildout (following its earlier Q1 FY2027 report), including long-term targets and a commitment to limit additional capital raises through FY2030 while expecting FCF positivity in H2 FY2028.[23][24]

• FY2027 reaffirmed at ≥$90 billion revenue and $8.10 non-GAAP EPS; FY2030 targets >$225 billion revenue (>35% CAGR from FY2026), >40% non-GAAP operating margin, and >$21 non-GAAP EPS.[23]
• Capex expected to peak in FY2027–2028 then decline; additional funding limited to $20 billion total through FY2030 ($10B equity + $10B gross debt, excluding refinancings); self-funded growth from FY2029 onward with net debt/EBITDA <3.5x by FY2029.[24]
• Earlier Q1 results (Sep 10) showed cloud infrastructure +121% to $7.4 billion and RPO +$209 billion YoY to $664 billion.[25]

These commitments reinforce the industry-wide theme of front-loaded AI infrastructure investment with multi-year payback expectations.

Aggregate 2026 capex guidance across the group (Microsoft ~$170B, Alphabet $195–205B, Amazon ~$220B, Meta $135–145B) points to roughly $590 billion in combined spending, underscoring the scale of the AI buildout.[26] Market reactions were bifurcated—positive or resilient for Microsoft and Meta on execution/F CF signals or funding innovations, more negative for Alphabet and Amazon on cash flow metrics—while analysts emphasized demand exceeding supply as validation but flagged depreciation, energy costs, and return timelines as watchpoints.[6][10]

Key comparison table (recent quarter vs. year-ago; all figures approximate from reports):

• Microsoft (FQ1’27): Rev +19% ($92.3B); Azure +48% ($33.2B); Capex $45.2B (vs. $34.9B); FCF +$18.6B (strong beat); 2026 guide $170B (trimmed).
• Alphabet (Q3’26): Rev +25% ($127.6B); Cloud +103% ($30.7B); Capex $51.4B (vs. $24B); FCF -$9.1B; 2026 guide $195–205B (unchanged).
• Amazon (Q3’26): Rev +12% ($201.9B); AWS +39% ($46B); Capex $77.3B (vs. $35.1B); TTM FCF -$42.4B (vs. +$14.8B); 2026 guide ~$220B (unchanged); raised $48.1B debt.
• Meta (Q3’26): Rev +26% ($64.6B); Capex $34.9B (vs. ~$19.4B); FCF -$2.1B (vs. +$10.6B); 2026 guide $135–145B (raised floor).

Nvidia’s latest reported quarter (Q2 FY2027, Aug 2026) showed continued Data Center strength but fell outside the October window; Oracle’s October Investor Day added long-term framing without new quarterly results.[27]

For competitors or new entrants, these results signal that AI infrastructure spending is now table stakes with differentiated outcomes based on existing cloud scale, capacity timing, and ability to monetize via services/backlog conversion. Cash flow discipline and innovative financing (e.g., Meta’s share class, Oracle’s funding caps) are emerging as competitive advantages amid rising scrutiny on returns. Additional detail on depreciation schedules or exact RPO conversion rates would further clarify payback dynamics.