# GOOGL Full forecast · Q3 2026

> A newer forecast, cutoff 29 October 2026, replaces this one. This page stays as published.

Canonical: https://pricedin.markets/googl/fq3-2026/forecast/

cutoff 27 October 2026, 03:39 ET

A newer forecast, cutoff 29 October 2026, replaces this one. This page stays as published.

[Full forecast, Q4 2026 →](https://pricedin.markets/googl/fq4-2026/forecast/)

GOOGL · as of 2026-10-27 03:39:04 ET · 6 AI forecasters · 80% ranges, 1 in 10 to 9 in 10

## Sales and earnings outlook

Next quarter · Q3 2026

Sales $128B +25% on Q3 2025

Adjusted EPS $2.94 +34% on Q3 2025 ($2.19: $2.87 reported, less $0.68 for gain (loss) on equity securities, net, in the Q3 2025 release)

Quarter after · Q4 2026

Sales $143B +25% on Q4 2025

Adjusted EPS $3.22 +21% on Q4 2025 ($2.67: $2.82 reported, less $0.15 for gain (loss) on equity securities, net, in the Q4 2025 release)

This year · FY2026

Sales $500B +24% on FY2025

Adjusted EPS $11.76

Next year · FY2027

Sales $608B +22% on FY2026

Adjusted EPS $13.53 +15% on FY2026

Year +2 · FY2028

Sales $713B +17% on FY2027

Adjusted EPS $15.84 +17% on FY2027

##### Sales

PricedIn **$128B** $122B – $134B

$122B $125B $128B $131B $134B

##### adjusted EPS

PricedIn **$2.94** $2.51 – $3.35

$2.51 $2.74 $2.94 $3.14 $3.35

**Middle** the forecast **50% range** 1 in 4 chance the result comes in below this bar, 1 in 4 above it. Half the time it lands inside. **80% range** 1 in 10 chance below this line, 1 in 10 above it. These are the low and high printed by each number.

The models agree revenues grow about a quarter on last year and margins stay near 34%; they split on Cloud. Muse Spark Open is lowest at $124B, with Cloud adding only “the modest ~$1B sequential assumed” and having “kept ad-growth assumptions conservative”. Sol is highest at $131B, forecasting “cloud sales of about $31 billion in the third quarter”. Muse Spark Open's range is widest at the bottom: its $115B low has Search and YouTube slowing “to low single digits”.

The models expect third-quarter revenues of about $128B, as fast Cloud growth outweighs a tougher Search comparison and a slight currency drag.

Sales **$128B** +25% on Q3 2025 ($102B)

*50% range* $125B – $131B *80% range* $122B – $134B

Adjusted EPS **$2.94** +34% on Q3 2025 ($2.19: $2.87 reported, less $0.68 for gain (loss) on equity securities, net, in the Q3 2025 release)

*50% range* $2.74 – $3.14 *80% range* $2.51 – $3.35

#### Investment and cash generation

Base case

Capital spending **$56.2B**

Free cash flow **−$5.83B**

Free cash flow: operating cash flow less capital spending.

Alphabet's second quarter set a high bar. Revenues reached $119.8B, up 24% from a year earlier. Cloud revenues rose 82% to $24.8B, Cloud's backlog of signed contracts climbed to $514B, and Search and other advertising grew 17%.

The company flagged brakes on the third quarter. Search starts lapping last year's acceleration, currency turns from a help into a slight headwind, and the World Cup lift to YouTube fades. It is also renting outside computing capacity, which it says will squeeze margins for a while.

The quarter has ended, but its results were not out when the forecasts were made. The models' average of $128B is 25% above a year earlier and 6.6% above the second quarter. They put Cloud near $28.5B and Google Services near $98.8B, with an operating margin near 34%.

They differ mostly on Cloud. Sol puts it near $31B on strong customer wins, while Muse Spark Open allows only a small step up. The ranges run from $115B to $140B, but most models sit between $123B and $132B.

The models expect earnings per share of about $2.94, leaving out gains on equity securities, which added $6.26 to the second quarter's $9.11. A margin near 34% and more shares after June's stock sale keep it close to last year's $2.87.

#### Sales by quarter

Q1 2025 $90B +12%

Q2 2025 $96B +14%

Q3 2025 $102B +16%

Q4 2025 $114B +18%

Q1 2026 $110B +22%

Q2 2026 $120B +24%

Q3 2026 $128B +25%

Each percent is on the same quarter a year earlier. The last row is the models' forecast.

#### The quarter forecast, line by line

| Line | Q2 2026 reported | Q3 2025 reported | Q3 2026 forecast | Q4 2025 reported | Q4 2026 forecast |
| --- | --- | --- | --- | --- | --- |
| Revenues | $120B | $102B | $128B | $114B | $143B |
| Sales growth, year on year | +24% | +16% | +25% | +18% | +25% |
| Google Services | $95B | $87B | $99B | $96B | $109B |
| Google Cloud | $24.8B | $15.2B | $28.5B | $17.7B | $33.7B |
| Other Bets | $382M | $344M | $400M | $370M | $377M |
| Hedging gains (losses) | $106M | −$207M | $18M | −$67M | $15M |
| Cost of revenues | $45.9B | $41.4B | $49.9B | $45.8B | $56.4B |
| Research and development | $18.2B | $15.2B | $19.5B | $18.6B | $21.9B |
| Sales and marketing | $8.4B | $7.2B | $8.9B | $8.2B | $10.2B |
| General and administrative | $6.46B | $7.39B | $5.86B | $5.34B | $6.47B |
| Total costs and expenses | $79.0B | $71.1B | $84.1B | $77.9B | $94.9B |
| Income from operations | $40.8B | $31.2B | $43.5B | $35.9B | $47.8B |
| Operating margin | 34.0% | 30.5% | 34.1% | 31.6% | 33.5% |
| Other income (expense), net | $98.0B | $12.8B | $2.4B | $3.2B | $2.6B |
| Income before income taxes | $139B | $44B | $46B | $39B | $50B |
| Provision for income taxes | $26.6B | $9.0B | $8.2B | $4.7B | $8.9B |
| Tax rate | 19.1% | 20.5% | 17.8% | 11.9% | 17.7% |
| Net income | $112B | $35B | $38B | $34B | $41B |
| Net income available to common stockholders | $112B | — | $37B | — | $41B |
| Number of shares used in diluted net income per common share calculation | 12.31B | 12.20B | 12.39B | 12.23B | 12.44B |
| Adjusted EPS | $2.85 | $2.19 | $2.94 | $2.67 | $3.22 |

Q2 2026 adjusted EPS $2.85: $9.11 reported, less $6.26 for gain (loss) on equity securities, net, in the Q2 2026 release.

Q3 2025 adjusted EPS $2.19: $2.87 reported, less $0.68 for gain (loss) on equity securities, net, in the Q3 2025 release.

Q4 2025 adjusted EPS $2.67: $2.82 reported, less $0.15 for gain (loss) on equity securities, net, in the Q4 2025 release.

The lines as the company's statement shows them. The forecast columns are the average of the models, line by line. Sales growth is the change on the same quarter a year earlier. Earnings per share is the headline's: the company's own for the reported quarter, and for a forecast quarter worked out from the averaged lines with the company's adjustments.

#### What it rests on

**Second-quarter revenues**

$119.8B, up 24% from a year earlier “Consolidated revenues were $119.8 billion, up 24% with 23% in constant currency.” Second-quarter earnings call

**Cloud growth**

Cloud revenues rose 82% to $24.8B in the second quarter “•Google Cloud saw a meaningful acceleration in growth as revenues increased 82% to $24.8 billion, led by an increase in Google Cloud Platform (GCP) across enterprise AI Solutions and enterprise AI Infrastructure, as well as core GCP services.” Second-quarter earnings release

**Cloud backlog**

$514B at the end of June, up more than $50B in the quarter “Google Cloud's backlog increased by more than $50 billion sequentially, reaching $514 billion in the second quarter.” Second-quarter earnings call

**Tougher Search comparison**

Search begins lapping last year's third-quarter acceleration “Q3, we will begin lapping an acceleration in Search performance that began in the third” Second-quarter earnings call

**Currency**

A slight headwind in the third quarter, after a one-point tailwind in the second “at the current spot rates, we would expect a slight FX headwind to our consolidated revenue in Q3, compared to a one percentage point FX tailwind in Q2.” Second-quarter earnings call

**Rented capacity**

Outside capacity added in the third quarter as a bridge to its own data centers “capacity in Q3 as a bridging strategy while we build out more internal capacity. This” Second-quarter earnings call

#### Low and high

Base the middle forecast

Key assumptions

- ****Cloud keeps climbing as backlog converts, to about $28.5B, with chip system sales mostly later.** Backlog reached $514B, and the company expects just over half of it as revenue within 24 months. *6 of 6***
  - **Opus 5.5** Cloud backlog turning into revenue
  - **GPT-6.1 Sol** Cloud capacity expansion and 2027 chip-system delivery ramp
  - **GLM 5.3** Google Cloud backlog of $514 billion converting to revenue, with just over half recognized over 24 months and TPU system sales ramping into 2027
  - **Kimi K3.1 Pro** Cloud backlog conversion and accelerator-system ramp
  - **Muse Spark Open** Cloud sustains strong growth with backlog conversion and TPU sales starting
  - **DeepSeek V4.1 Flash** Google Cloud revenue growth off the $514 billion backlog, more than half converting within 24 months
- ****Search growth eases to about 15% to 16%, from 17%, as it laps last year's acceleration.** The company warned of the tougher comparison and a slight currency headwind in the third quarter. *6 of 6***
  - **Opus 5.5** I slowed Search growth to about 16% in the third quarter (from 17%) because of the tougher comparison and a small currency headwind.
  - **DeepSeek V4.1 Flash** I take Search growth from about 17% in the second quarter to 15% in the third
  - **GLM 5.3** Search and YouTube advertising growing low-to-mid teens as AI features expand queries and monetization, decelerating as the 2025 acceleration is lapped
  - **Muse Spark Open** CFO said Q3 will see slight FX headwind vs Q2 tailwind and lapping of Search acceleration starting Q3 last year
  - **GPT-6.1 Sol** Search comparisons, the fading World Cup boost and capacity constraints are valid cautions
  - **Kimi K3.1 Pro** Mid-teens Google Services growth
- ****The operating margin holds near 34%, as rented capacity and depreciation offset growth.** The company said renting outside capacity will create modest margin pressure in the near term. *6 of 6***
  - **Opus 5.5** hold the third-quarter operating margin at about 34%
  - **GLM 5.3** the company guided to margin pressure in H2-2026 from third-party capacity bought as a bridge
  - **DeepSeek V4.1 Flash** Rented capacity costs more than capacity Alphabet owns, so it eats Cloud margin in the near term.
  - **Muse Spark Open** Use of third-party capacity as bridging strategy will create modest margin pressure in near-term
  - **Kimi K3.1 Pro** near-term third-party capacity pressure
  - **GPT-6.1 Sol** warns that rented capacity creates near-term margin pressure
- ****Capital spending near $56.2B in the quarter keeps free cash flow negative, near −$5.8B.** The company raised its 2026 capital spending range twice and expects free cash flow to stay under pressure. *6 of 6***
  - **Opus 5.5** I take 2026 capital spending at $200 billion, the middle of the new guidance
  - **GPT-6.1 Sol** I use about $200 billion of capital spending this year
  - **GLM 5.3** I use $200 billion, about $55 billion in Q3
  - **Kimi K3.1 Pro** $200 billion of 2026 capital spending
  - **Muse Spark Open** drove base capex 200B and 260B
  - **DeepSeek V4.1 Flash** I take $200 billion for the year, split about $57 billion in the third quarter
- ****No further gains on equity securities count in earnings per share.** The second quarter's $99.0B equity gain added $6.26 a share; such marks cannot be forecast. *5 of 6***
  - **Opus 5.5** I assume no equity gains from here on
  - **Kimi K3.1 Pro** forecast no future gain or loss
  - **DeepSeek V4.1 Flash** I assume no net gain or loss on equity securities in any future period
  - **GLM 5.3** my adjusted EPS removes them
  - **GPT-6.1 Sol** Their size does not improve adjusted earnings.

If all of this happens

**$128B** sales

**$2.94** adjusted EPS

**$56.2B** Capital spending

**−$5.83B** Free cash flow

Cash walk not supplied by every model.

Low about 1 chance in 10 of worse

What the low case assumes

- ****Search growth slows to 11% to 12%, against about 16% in the base.** The quarter laps last year's Search acceleration, and oil near $115 a barrel weighs on advertisers. *6 of 6***
  - **Opus 5.5** Search slows toward 11 to 12% against the tough comparison
  - **Muse Spark Open** Search and YouTube growth slows to low single digits on macro and lapping
  - **DeepSeek V4.1 Flash** Search and other revenue growth falls to about 13%
  - **GLM 5.3** Search decelerates to single digits
  - **Kimi K3.1 Pro** Slower Search and YouTube advertising growth
  - **GPT-6.1 Sol** Weaker advertiser spending and tougher search competition
- ****Cloud stalls near $27B rather than $28.5B as capacity and chip deliveries slip.** Management has said demand outruns supply and expected a tight demand-supply environment through 2026. *6 of 6***
  - **Opus 5.5** Cloud capacity shortfall and slower backlog conversion
  - **GPT-6.1 Sol** Delayed capacity delivery and chip-system shipments
  - **GLM 5.3** TPU shipments slip
  - **Kimi K3.1 Pro** Delayed Cloud capacity and accelerator-system deliveries
  - **Muse Spark Open** Cloud expansion slows and third-party capacity creates margin pressure
  - **DeepSeek V4.1 Flash** Google Cloud sequential growth stalling near $1.5-2 billion a quarter as TPU system deliveries slip into 2027
- ****Depreciation and rented capacity push the operating margin below 34%.** Depreciation rose 38% in 2025, and rented capacity costs more than owned data centers. *6 of 6***
  - **Opus 5.5** depreciation and rented capacity costs
  - **GPT-6.1 Sol** High depreciation, energy and external-capacity costs
  - **GLM 5.3** Operating margin compresses as depreciation, third-party capacity costs and supply-chain inflation outrun revenue leverage
  - **Kimi K3.1 Pro** Higher infrastructure and energy costs
  - **Muse Spark Open** Depreciation and AI talent costs rise faster than sales
  - **DeepSeek V4.1 Flash** Rented third-party capacity and a hardware-heavy mix cutting Cloud operating margin
- ****A stronger dollar trims revenues earned abroad more than the slight headwind expected.** About half of sales come from outside the United States. *3 of 6***
  - **Opus 5.5** stronger dollar
  - **Kimi K3.1 Pro** Foreign-exchange and regulatory pressure
  - **GLM 5.3** it trims my reported revenue middle slightly

If all of this happens

**$122B** sales · −5% vs base

**$2.51** EPS · −15% vs base

Revenues would have to land near $122B, about 4.7% under the base. Search would slow to 11% to 12% against the tough comparison, Cloud would stall near $27B as capacity stays short, and high oil prices, weak consumers and a stronger dollar would trim ad budgets. Rented capacity and depreciation would push the margin down, taking earnings per share toward $2.51, about 15% below the base. Muse Spark Open's low is furthest, at $115B, because it alone lets Search and YouTube growth fall to low single digits in a quarter that has already ended.

In the second quarter Search grew 17% and Cloud 82%, and the company is renting extra capacity precisely to keep serving demand.

The six lows run from $115B (Muse Spark Open) to $125B (GLM 5.3).

First sign The third-quarter report, the next one due: Cloud revenues below $27B or Search growth below 13%.

High about 1 chance in 10 of better

What the high case assumes

- ****Cloud tops $30B as capacity frees backlog, against about $28.5B in the base.** Customers typically spend more than 50% above their commitments, the Cloud chief said in September. *6 of 6***
  - **Opus 5.5** Rented capacity and new data centers release pent-up demand, so Cloud tops $30 billion in the third quarter
  - **GPT-6.1 Sol** Faster conversion of cloud commitments and additional large contracts
  - **GLM 5.3** TPU system shipments ramp faster than planned and new mega-deals pull revenue forward from 2027 into late 2026
  - **Kimi K3.1 Pro** Faster Cloud capacity and accelerator-system delivery
  - **Muse Spark Open** Cloud backlog 514 billion converts faster with new customers doubling
  - **DeepSeek V4.1 Flash** Faster conversion of the $514 billion Cloud backlog and earlier TPU system revenue
- ****Search holds near 17% growth, against about 16% in the base, as AI Mode advertising scales.** AI Mode passed one billion monthly users, and the company says it adds to Search queries overall. *6 of 6***
  - **GLM 5.3** Search sustains 17%+ growth as AI Mode monetization expands ad coverage to previously unmonetizable queries
  - **DeepSeek V4.1 Flash** Search and other revenue holds near 17% growth as AI Mode and AI Max widen monetisable queries
  - **Opus 5.5** AI Mode and agentic commerce advertising
  - **Muse Spark Open** Search queries hit new highs on AI Mode with over one billion users driving upside
  - **Kimi K3.1 Pro** Sustained strong Search and subscription growth
  - **GPT-6.1 Sol** Improved search advertising coverage and subscription pricing
- ****The operating margin holds near 35% as Cloud's scale offsets rented capacity.** Cloud's operating margin rose to 35.6% in the second quarter from 20.7% a year earlier. *6 of 6***
  - **Opus 5.5** operating leverage in Cloud
  - **GLM 5.3** Cloud operating margin keeps expanding beyond 35.6% as internal capacity replaces costly third-party bridging capacity
  - **Kimi K3.1 Pro** Better-than-expected Cloud operating leverage
  - **Muse Spark Open** Operating leverage from AI efficiencies and pricing power
  - **DeepSeek V4.1 Flash** Operating leverage outrunning depreciation and energy costs
  - **GPT-6.1 Sol** Efficient internal capacity replacing expensive rented capacity

If all of this happens

**$134B** sales · +5% vs base

**$3.35** EPS · +14% vs base

Revenues would have to reach about $134B, 4.9% above the base. Rented capacity and new data centers would release pent-up Cloud demand, lifting Cloud past $30B, and Search would hold near 17% as AI Mode advertising scales. The margin would stay near 35%, lifting earnings per share toward $3.35, about 14% above the base. Sol's high is furthest, at $140B, because it lets capacity arrive sooner and enterprise demand convert quickly into both usage and chip system sales, on top of a middle that already has Cloud near $31B.

In September the Cloud chief described a run rate of nearly $100B a year, which matches about $25B a quarter.

The six highs run from $130B (GLM 5.3) to $140B (GPT-6.1 Sol).

First sign The third-quarter report, the next one due: Cloud growth above 95% or backlog well above $550B.

#### Watch

The third-quarter report is the test. Cloud near $28.5B with backlog still rising fits the base; below $27B points lower, above $30B higher. Search growth below 13% points lower, near 17% higher. Watch also the operating margin against 34% and what management says about rented capacity costs.

##### Sales

PricedIn **$143B** $134B – $151B

$134B $139B $143B $147B $151B

##### adjusted EPS

PricedIn **$3.22** $2.73 – $3.72

$2.73 $2.98 $3.22 $3.45 $3.72

**Middle** the forecast **50% range** 1 in 4 chance the result comes in below this bar, 1 in 4 above it. Half the time it lands inside. **80% range** 1 in 10 chance below this line, 1 in 10 above it. These are the low and high printed by each number.

The models agree on growth near 25% from a year earlier but split on Cloud. Muse Spark Open is lowest at $138B; it keeps Cloud near “65 to 70 percent” growth and has “kept ad-growth assumptions conservative”. Sol and Kimi are highest at $147B: Sol puts Cloud near “$37 billion in the fourth”, and Kimi sees Services sustaining “mid-teens growth”. Sol's range is widest, from $132B with “Delayed capacity delivery and chip-system shipments” to $163B if “Capacity is delivered sooner”.

The models expect fourth-quarter revenues of about $143B, as Cloud's chip system sales begin to ramp at the end of the year.

Sales **$143B** +25% on Q4 2025 ($114B)

*50% range* $139B – $147B *80% range* $134B – $151B

Adjusted EPS **$3.22** +21% on Q4 2025 ($2.67: $2.82 reported, less $0.15 for gain (loss) on equity securities, net, in the Q4 2025 release)

*50% range* $2.98 – $3.45 *80% range* $2.73 – $3.72

#### Investment and cash generation

Capital spending: not supplied by every model

Free cash flow: not supplied by every model

Free cash flow: operating cash flow less capital spending.

The fourth quarter is the one the company would discuss at its third-quarter report. Alphabet prints no revenue outlook, but it said it expects a relatively small portion of its chip system sales revenue this year, ramping as it exits 2026, with the vast majority in 2027.

A year earlier, fourth-quarter revenues were $114B and earnings per share $2.82. The models' average of $143B is 25% higher and about 12% above their third-quarter figure. They put Cloud near $33.7B and Google Services near $109B, with an operating margin near 33%.

Costs step up too. Depreciation keeps rising with $200B of capital spending this year, rented capacity costs more than owned data centers, and the Wiz purchase weighs on Cloud's margin for the rest of 2026.

The models' middles run from Muse Spark Open's $138B to $147B for Sol and Kimi, mostly on how much chip system revenue lands before the year ends. Their full ranges span $130B to $163B.

The models expect about $3.22 a share, 14% above a year earlier, leaving out equity gains. The margin slips toward 33% as depreciation and lower-margin chip hardware weigh, and more shares and preferred dividends trim what reaches each common share.

#### The quarter forecast, line by line

| Line | Q2 2026 reported | Q3 2025 reported | Q3 2026 forecast | Q4 2025 reported | Q4 2026 forecast |
| --- | --- | --- | --- | --- | --- |
| Revenues | $120B | $102B | $128B | $114B | $143B |
| Sales growth, year on year | +24% | +16% | +25% | +18% | +25% |
| Google Services | $95B | $87B | $99B | $96B | $109B |
| Google Cloud | $24.8B | $15.2B | $28.5B | $17.7B | $33.7B |
| Other Bets | $382M | $344M | $400M | $370M | $377M |
| Hedging gains (losses) | $106M | −$207M | $18M | −$67M | $15M |
| Cost of revenues | $45.9B | $41.4B | $49.9B | $45.8B | $56.4B |
| Research and development | $18.2B | $15.2B | $19.5B | $18.6B | $21.9B |
| Sales and marketing | $8.4B | $7.2B | $8.9B | $8.2B | $10.2B |
| General and administrative | $6.46B | $7.39B | $5.86B | $5.34B | $6.47B |
| Total costs and expenses | $79.0B | $71.1B | $84.1B | $77.9B | $94.9B |
| Income from operations | $40.8B | $31.2B | $43.5B | $35.9B | $47.8B |
| Operating margin | 34.0% | 30.5% | 34.1% | 31.6% | 33.5% |
| Other income (expense), net | $98.0B | $12.8B | $2.4B | $3.2B | $2.6B |
| Income before income taxes | $139B | $44B | $46B | $39B | $50B |
| Provision for income taxes | $26.6B | $9.0B | $8.2B | $4.7B | $8.9B |
| Tax rate | 19.1% | 20.5% | 17.8% | 11.9% | 17.7% |
| Net income | $112B | $35B | $38B | $34B | $41B |
| Net income available to common stockholders | $112B | — | $37B | — | $41B |
| Number of shares used in diluted net income per common share calculation | 12.31B | 12.20B | 12.39B | 12.23B | 12.44B |
| Adjusted EPS | $2.85 | $2.19 | $2.94 | $2.67 | $3.22 |

Q2 2026 adjusted EPS $2.85: $9.11 reported, less $6.26 for gain (loss) on equity securities, net, in the Q2 2026 release.

Q3 2025 adjusted EPS $2.19: $2.87 reported, less $0.68 for gain (loss) on equity securities, net, in the Q3 2025 release.

Q4 2025 adjusted EPS $2.67: $2.82 reported, less $0.15 for gain (loss) on equity securities, net, in the Q4 2025 release.

The lines as the company's statement shows them. The forecast columns are the average of the models, line by line. Sales growth is the change on the same quarter a year earlier. Earnings per share is the headline's: the company's own for the reported quarter, and for a forecast quarter worked out from the averaged lines with the company's adjustments.

#### What it rests on

**Chip system timing**

A relatively small portion of chip system revenue this year, ramping as 2026 ends “We continue to expect to recognize a relatively small portion of the revenues from our existing TPU system sales agreements this year, ramping as we exit 2026.” Second-quarter earnings call

**Most chip revenue later**

The vast majority of chip system revenue lands in 2027 “We anticipate the vast majority of the revenues from these agreements will be realized in 2027.” Second-quarter earnings call

**Wiz weighs on Cloud**

A low single-digit point drag on Cloud's margin for the rest of 2026 “we expect a low single-digit percentage point headwind to cloud's operating margin for the remainder of 2026 related to the acquisition” First-quarter earnings call

**Capital spending range**

Raised to $195B to $205B for 2026, from $180B to $190B “Moving to investments, we are updating our full year 2026 CapEx guidance range to” Second-quarter earnings call

**Cloud backlog**

$514B at the end of June, up more than $50B in the quarter “Google Cloud's backlog increased by more than $50 billion sequentially, reaching $514 billion in the second quarter.” Second-quarter earnings call

#### Low and high

Base the middle forecast

Key assumptions

- ****Cloud steps up to about $33.7B as the first chip system sales ramp.** The company expects a relatively small portion of chip system revenue this year, ramping as 2026 ends. *6 of 6***
  - **Opus 5.5** I put Cloud at about $28.5 billion in the third quarter and $33 billion in the fourth
  - **GLM 5.3** about $28.6 billion in Q3 and $34.4 billion in Q4 as TPU shipments ramp exiting the year
  - **GPT-6.1 Sol** $37 billion in the fourth
  - **DeepSeek V4.1 Flash** I allow only a modest step-up in Cloud in the fourth quarter
  - **Kimi K3.1 Pro** Cloud backlog conversion and accelerator-system ramp
  - **Muse Spark Open** Cloud sustains strong growth with backlog conversion and TPU sales starting
- ****Google Services grows in the low-to-mid teens, to about $109B.** Search faces a harder comparison all quarter, and the World Cup lift to YouTube is gone. *6 of 6***
  - **GLM 5.3** keeping Google Services at about 13-14% growth
  - **DeepSeek V4.1 Flash** hold it near 15% in the fourth
  - **Kimi K3.1 Pro** Mid-teens Google Services growth
  - **Opus 5.5** Search growing about 15% with AI features
  - **GPT-6.1 Sol** Services grow at low-teens rates
  - **Muse Spark Open** Search and YouTube sustain double-digit growth on AI Overviews and AI Mode
- ****The operating margin slips to about 33%, from 34% in the third quarter.** Depreciation, rented capacity, the Wiz drag and lower-margin chip hardware all rise faster than revenue. *6 of 6***
  - **Opus 5.5** Fourth-quarter costs usually step up with marketing and holiday device sales.
  - **GLM 5.3** plus a low-single-digit headwind from the Wiz acquisition
  - **DeepSeek V4.1 Flash** I raise cost of revenues from 38.4% of revenue in the second quarter to about 39.5% in the third and 39.9% in the fourth
  - **Kimi K3.1 Pro** Heavy infrastructure spending and rising depreciation
  - **GPT-6.1 Sol** Efficiency gains partly offsetting depreciation and research costs
  - **Muse Spark Open** Capex 195 to 205 billion in 2026 rising significantly in 2027 pressures depreciation
- ****More shares and preferred dividends trim earnings per common share.** June's sale of common and preferred stock raised $49.6B, and no buybacks were made in the first half. *5 of 6***
  - **Opus 5.5** more shares after the June equity raise
  - **GLM 5.3** with preferred dividends of about $290 million a quarter reducing income available to common stockholders
  - **Kimi K3.1 Pro** Gradual diluted-share growth and preferred dividends
  - **DeepSeek V4.1 Flash** I also let the diluted share count creep from 12,309 million in the second quarter to 12,400 million in the third and 12,450 million in the fourth
  - **GPT-6.1 Sol** Common earnings deduct about $300 million of preferred dividends per forecast quarter.

If all of this happens

**$143B** sales

**$3.22** adjusted EPS

Low about 1 chance in 10 of worse

What the low case assumes

- ****Chip system deliveries slip into 2027, holding Cloud below the $33.7B base.** Chip system revenue is booked on delivery and can vary from quarter to quarter. *6 of 6***
  - **GPT-6.1 Sol** Delayed capacity delivery and chip-system shipments
  - **DeepSeek V4.1 Flash** TPU system deliveries slip into 2027
  - **GLM 5.3** TPU shipments slip
  - **Kimi K3.1 Pro** Delayed Cloud capacity and accelerator-system deliveries
  - **Opus 5.5** Cloud capacity shortfall and slower backlog conversion
  - **Muse Spark Open** Cloud expansion slows and third-party capacity creates margin pressure
- ****High energy prices and weak consumers cut advertising budgets.** Oil neared $115 a barrel at the cutoff and consumer sentiment was 51.7, against 58.2 a year earlier. *6 of 6***
  - **GPT-6.1 Sol** an energy and financing shock weakens advertising
  - **DeepSeek V4.1 Flash** Advertising slowing as expensive energy squeezes consumers and a five per cent ten-year yield trims budgets
  - **GLM 5.3** A consumer downturn (sentiment 51.7, mortgage 7.3%) cuts advertising budgets across Search, YouTube and Network
  - **Kimi K3.1 Pro** Slower Search and YouTube advertising growth
  - **Opus 5.5** high oil prices and a stronger dollar trim advertising budgets
  - **Muse Spark Open** Search and YouTube growth slows to low single digits on macro and lapping
- ****Depreciation and rented capacity outgrow revenue, pushing the margin under 33%.** Depreciation rose 38% in 2025, and spending this year is about double 2025's $91.4B. *6 of 6***
  - **Opus 5.5** depreciation and rented capacity costs
  - **GPT-6.1 Sol** High depreciation, energy and external-capacity costs
  - **GLM 5.3** Operating margin compresses as depreciation, third-party capacity costs and supply-chain inflation outrun revenue leverage
  - **Kimi K3.1 Pro** Higher infrastructure and energy costs
  - **Muse Spark Open** Depreciation and AI talent costs rise faster than sales
  - **DeepSeek V4.1 Flash** Depreciation from about $200 billion of 2026 spending running ahead of revenue

If all of this happens

**$134B** sales · −6% vs base

**$2.73** EPS · −15% vs base

Revenues would have to come in near $134B, about 5.7% below the base. Chip system deliveries would slip into 2027, Cloud's quarterly gains would shrink, and an energy and financing shock would cut advertising budgets. Depreciation and rented capacity would grow faster than sales, pulling earnings per share toward $2.73, about 15% under the base. Sol's low is furthest, at $132B, because its middle carries the most Cloud, near $37B, so a delay in capacity delivery and chip system shipments takes the most away.

Fourth-quarter revenues a year earlier were $114B, so even the lowest case still shows strong growth.

The six lows run from $130B (Muse Spark Open) to $140B (GLM 5.3).

First sign At the third-quarter report: management pushing chip system revenue later, or Cloud growth below 70%.

High about 1 chance in 10 of better

What the high case assumes

- ****Chip systems ship sooner and new contracts convert fast, lifting Cloud above the $33.7B base.** Customers typically spend more than 50% above their commitments, and new customer wins doubled. *6 of 6***
  - **Opus 5.5** more capacity unlocking backlog
  - **GPT-6.1 Sol** Faster conversion of cloud commitments and additional large contracts
  - **GLM 5.3** TPU system shipments ramp faster than planned and new mega-deals pull revenue forward from 2027 into late 2026
  - **Kimi K3.1 Pro** Faster Cloud capacity and accelerator-system delivery
  - **Muse Spark Open** Cloud backlog 514 billion converts faster with new customers doubling
  - **DeepSeek V4.1 Flash** Faster conversion of the $514 billion Cloud backlog and earlier TPU system revenue
- ****Search holds near 17% growth as AI Mode advertising scales.** AI Mode passed one billion monthly users, and the company says it adds to Search queries overall. *6 of 6***
  - **GLM 5.3** Search sustains 17%+ growth as AI Mode monetization expands ad coverage to previously unmonetizable queries
  - **DeepSeek V4.1 Flash** Search monetisation of longer, previously unmonetisable AI queries
  - **Opus 5.5** AI Mode and agentic commerce advertising
  - **Muse Spark Open** Search queries hit new highs on AI Mode with over one billion users driving upside
  - **Kimi K3.1 Pro** Sustained strong Search and subscription growth
  - **GPT-6.1 Sol** Improved search advertising coverage and subscription pricing
- ****Margins hold near 34% as internal capacity replaces rented capacity.** Cloud's operating margin rose to 35.6% in the second quarter from 20.7% a year earlier. *6 of 6***
  - **Opus 5.5** operating leverage in Cloud
  - **GLM 5.3** Cloud operating margin keeps expanding beyond 35.6% as internal capacity replaces costly third-party bridging capacity
  - **Kimi K3.1 Pro** Better-than-expected Cloud operating leverage
  - **Muse Spark Open** Operating leverage from AI efficiencies and pricing power
  - **DeepSeek V4.1 Flash** Operating leverage outrunning depreciation and energy costs
  - **GPT-6.1 Sol** Efficient internal capacity replacing expensive rented capacity

If all of this happens

**$151B** sales · +6% vs base

**$3.72** EPS · +15% vs base

Revenues would have to reach about $151B, 6.1% above the base. Chip systems would ship faster than the company's small-portion guidance, new large contracts would add to backlog, and Search would hold near 17%. The margin would hold as internal capacity replaces rented capacity, lifting earnings per share toward $3.72, about 15% above the base. Sol's high is furthest, at $163B, because it lets capacity arrive sooner and demand convert rapidly into both usage and chip system sales, starting from a $37B Cloud middle.

Cloud backlog rose from $155B in September 2025 to $514B in June 2026.

The six highs run from $144B (DeepSeek V4.1 Flash) to $163B (GPT-6.1 Sol).

First sign At the third-quarter report: chip system revenue landing ahead of plan, or backlog additions above $50B.

#### Watch

The third-quarter report and its call: what management says about chip system deliveries, rented capacity costs and currency. More chip system revenue before the year ends points higher; delays, or Search growth falling below 13%, point lower.

##### Sales

PricedIn **$500B** $484B – $517B

$484B $492B $500B $508B $517B

##### adjusted EPS

PricedIn **$11.76** $10.75 – $12.80

$10.75 $11.28 $11.76 $12.27 $12.80

**Middle** the forecast **50% range** 1 in 4 chance the result comes in below this bar, 1 in 4 above it. Half the time it lands inside. **80% range** 1 in 10 chance below this line, 1 in 10 above it. These are the low and high printed by each number.

The models agree on growth of 22% to 26% and a margin of 34% to 35%. Muse Spark Open is lowest at $491B, with Cloud adding “the modest ~$1B sequential assumed” and having “kept ad-growth assumptions conservative”. Sol is highest at $508B, with “cloud sales of about $31 billion in the third quarter and $37 billion in the fourth”. Kimi's range is widest, from $480B if “Cloud capacity and accelerator deliveries slip” to $536B if “New capacity arrives faster than expected”.

The models expect 2026 revenues of about $500B, with Cloud's surge adding to steady advertising growth.

Sales **$500B** +24% on FY2025

*50% range* $492B – $508B *80% range* $484B – $517B

Adjusted EPS **$11.76** $5.61 reported so far (Q1 2026 $2.76: $5.11 reported, less $2.35 for gain (loss) on equity securities, net, in the Q1 2026 release; Q2 2026 $2.85: $9.11 reported, less $6.26 for gain (loss) on equity securities, net, in the Q2 2026 release)

*50% range* $11.28 – $12.27 *80% range* $10.75 – $12.80

Base case

Operating margin **34.4%** 32.0% in FY2025

#### Investment and cash generation

Base case

Capital spending **$200B**

Free cash flow **−$8.80B**

Free cash flow: operating cash flow less capital spending.

Alphabet booked $230B of revenues in the first half of 2026, after $403B in all of 2025. Growth quickened from 22% in the first quarter to 24% in the second, led by Cloud and Search.

The models put the full year near $500B, up 24%. They expect a bigger second half as Cloud's backlog converts and the first chip system sales arrive, while Search slows against last year's acceleration.

Spending is the other story. The company raised its 2026 capital spending range twice, to $195B to $205B, and all six models take $200B. Free cash flow, $73.3B in 2025, turns negative in their average, near −$8.8B, and the company raised money through debt and a June stock sale.

Reported earnings per share swing with gains on equity securities, which added $6.26 a share in the second quarter alone. The forecasts leave those gains out, so their full-year figure of about $11.76 sits well below the $14.22 already reported for the first half.

Leaving out equity gains, the models expect about $11.76 a share: roughly $2.76 and $2.85 in the first two quarters, then about $2.94 and $3.22. An operating margin near 34% carries it; more shares after June's stock sale hold it back.

#### Sales by quarter

Q1 2025 $90B +12%

Q2 2025 $96B +14%

Q3 2025 $102B +16%

Q4 2025 $114B +18%

Q1 2026 $110B +22%

Q2 2026 $120B +24%

Q3 2026 $128B +25%

Q4 2026 $143B +25%

Each percent is on the same quarter a year earlier. The forecast rows are the models' average.

#### What it rests on

**Capital spending range**

Raised to $195B to $205B for 2026, from $180B to $190B “Moving to investments, we are updating our full year 2026 CapEx guidance range to” Second-quarter earnings call

**Equity gain excluded**

A $99.0B gain added $6.26 to second-quarter earnings per share “For Q2 2026, the net effect of the gain on equity securities of $99.0 billion increased the provision for income tax, net income, and diluted net income per common share by $21.9 billion, $77.1 billion, and $6.26, respectively.” Second-quarter earnings release

**Free cash flow**

−$5.9B in the second quarter, driven by capital spending “We had negative free cash flow of $5.9 billion in the second quarter, driven by our investments in CapEx.” Second-quarter earnings call

**Backlog conversion**

Just over half of backlog expected as revenue within 24 months “for our broad mix of customers, and we expect to recognize just over 50% of the total” Second-quarter earnings call

**Cloud growth**

Cloud revenues rose 82% to $24.8B in the second quarter “•Google Cloud saw a meaningful acceleration in growth as revenues increased 82% to $24.8 billion, led by an increase in Google Cloud Platform (GCP) across enterprise AI Solutions and enterprise AI Infrastructure, as well as core GCP services.” Second-quarter earnings release

**No buybacks**

No share repurchases in the first half of 2026 “In the three and six months ended June 30, 2026, there were no repurchases of the company's Class A or Class C shares.” Second-quarter quarterly report

#### Low and high

Base the middle forecast

Key assumptions

- ****Cloud carries the growth as backlog converts, with chip system sales still small this year.** Backlog reached $514B; management expects only a small portion of chip system revenue in 2026. *6 of 6***
  - **Opus 5.5** Cloud backlog turning into revenue
  - **GPT-6.1 Sol** Cloud capacity expansion and 2027 chip-system delivery ramp
  - **GLM 5.3** Google Cloud backlog of $514 billion converting to revenue, with just over half recognized over 24 months and TPU system sales ramping into 2027
  - **Kimi K3.1 Pro** Cloud backlog conversion and accelerator-system ramp
  - **Muse Spark Open** Cloud sustains strong growth with backlog conversion and TPU sales starting
  - **DeepSeek V4.1 Flash** Timing of TPU system sales, small in 2026 and mostly in 2027
- ****Advertising keeps growing in the low-to-mid teens as AI features lift Search.** Search grew 17% in the second quarter, though it now laps last year's acceleration. *6 of 6***
  - **Opus 5.5** Search growing about 15% with AI features
  - **GPT-6.1 Sol** Healthy search advertising and paid subscription growth
  - **GLM 5.3** Search and YouTube advertising growing low-to-mid teens as AI features expand queries and monetization, decelerating as the 2025 acceleration is lapped
  - **Kimi K3.1 Pro** Mid-teens Google Services growth
  - **Muse Spark Open** Search and YouTube sustain double-digit growth on AI Overviews and AI Mode
  - **DeepSeek V4.1 Flash** Search and other revenue holding mid-teens growth as it laps the 2025 acceleration, with a slight third-quarter currency headwind
- ****The operating margin holds near 34% for the year, above 2025's 32%.** First-half margins were 36% and 34%; depreciation and rented capacity weigh on the second half. *6 of 6***
  - **Opus 5.5** The operating margin holds near 34%
  - **GLM 5.3** Operating margin holds around 33-34% in 2026
  - **DeepSeek V4.1 Flash** Operating margin holds near 34-35% because advertising leverage and cost discipline absorb a step-up in depreciation
  - **GPT-6.1 Sol** Consolidated margins stay around 35%
  - **Muse Spark Open** margins 34 to 35 percent supported by Cloud margin expansion
  - **Kimi K3.1 Pro** scale largely offsets depreciation and near-term capacity costs
- ****Capital spending of $200B leaves free cash flow near −$8.8B on average.** The company sets its range at $195B to $205B and expects free cash flow to stay under pressure. *6 of 6***
  - **Opus 5.5** I take 2026 capital spending at $200 billion, the middle of the new guidance
  - **GPT-6.1 Sol** I use about $200 billion of capital spending this year
  - **GLM 5.3** I use $200 billion
  - **Kimi K3.1 Pro** $200 billion of 2026 capital spending
  - **Muse Spark Open** drove base capex 200B and 260B
  - **DeepSeek V4.1 Flash** I take $200 billion for the year
- ****Equity gains are left out, and more shares follow June's stock sale.** First-half equity gains added $2.35 and $6.26 a share; no buybacks were made in the first half. *5 of 6***
  - **Opus 5.5** I assume no equity gains from here on
  - **Kimi K3.1 Pro** forecast no future gain or loss
  - **DeepSeek V4.1 Flash** I assume no net gain or loss on equity securities in any future period
  - **GLM 5.3** my adjusted EPS removes them
  - **GPT-6.1 Sol** Their size does not improve adjusted earnings.

If all of this happens

**$500B** sales

**$11.76** adjusted EPS

**$200B** Capital spending

**−$8.80B** Free cash flow

Cash walk not supplied by every model.

Low about 1 chance in 10 of worse

What the low case assumes

- ****Advertising slows in the second half as energy prices and weak consumers cut budgets.** Oil neared $115 a barrel at the cutoff and consumer sentiment was 51.7, against 58.2 a year earlier. *6 of 6***
  - **Opus 5.5** high oil prices and a stronger dollar trim advertising budgets
  - **GPT-6.1 Sol** Weaker advertiser spending and tougher search competition
  - **GLM 5.3** A consumer downturn (sentiment 51.7, mortgage 7.3%) cuts advertising budgets across Search, YouTube and Network
  - **Kimi K3.1 Pro** Slower Search and YouTube advertising growth
  - **Muse Spark Open** Search and YouTube growth slows to low single digits on macro and lapping
  - **DeepSeek V4.1 Flash** Advertising slowing as expensive energy squeezes consumers and a five per cent ten-year yield trims budgets
- ****Cloud capacity and chip deliveries slip, holding second-half Cloud below the base.** Management says demand outruns supply, and its own risk factors cite limits on power, water and land. *6 of 6***
  - **Opus 5.5** Cloud capacity shortfall and slower backlog conversion
  - **GPT-6.1 Sol** Delayed capacity delivery and chip-system shipments
  - **GLM 5.3** TPU shipments slip
  - **Kimi K3.1 Pro** Delayed Cloud capacity and accelerator-system deliveries
  - **Muse Spark Open** Cloud expansion slows and third-party capacity creates margin pressure
  - **DeepSeek V4.1 Flash** Google Cloud sequential growth stalling near $1.5-2 billion a quarter as TPU system deliveries slip into 2027
- ****Depreciation and rented capacity pull the margin toward 30%, against 34% in the base.** Capital spending of $200B this year is more than double 2025's $91.4B. *6 of 6***
  - **Opus 5.5** Depreciation and rented capacity push the operating margin toward 30%
  - **GPT-6.1 Sol** High depreciation, energy and external-capacity costs
  - **GLM 5.3** Operating margin compresses as depreciation, third-party capacity costs and supply-chain inflation outrun revenue leverage
  - **Kimi K3.1 Pro** Higher infrastructure and energy costs
  - **Muse Spark Open** Depreciation and AI talent costs rise faster than sales
  - **DeepSeek V4.1 Flash** Depreciation from about $200 billion of 2026 spending running ahead of revenue
- ****Legal remedies or new charges add pressure.** A December 2025 judgment requires sharing search data with competitors; both sides are appealing. *3 of 6***
  - **GPT-6.1 Sol** Adverse legal remedies or additional charges
  - **GLM 5.3** DOJ search remedies
  - **Kimi K3.1 Pro** Foreign-exchange and regulatory pressure

If all of this happens

**$484B** sales · −3% vs base

**$10.75** EPS · −9% vs base

Revenues would have to finish near $484B, about 3.1% under the base, with earnings per share near $10.75, about 8.6% lower. Cloud capacity and chip deliveries would slip, Search would slow against the tough comparison, and high oil prices and a stronger dollar would trim ad budgets in the second half. Depreciation and rented capacity would pull the margin toward 30%. Muse Spark Open's low is furthest, at $470B, because it lets Search and YouTube growth slow to low single digits, the deepest advertising slowdown any model allows.

Half the year is already booked at $230B, so the range is far narrower than for later years.

The six lows run from $470B (Muse Spark Open) to $493B (GLM 5.3).

First sign The third-quarter report, the next one due: Search growth below 13% or Cloud revenues below $27B.

High about 1 chance in 10 of better

What the high case assumes

- ****Capacity arrives sooner and chip systems ship faster, lifting second-half Cloud above the base.** Customers typically spend more than 50% above their commitments, the Cloud chief said in September. *6 of 6***
  - **Kimi K3.1 Pro** Faster Cloud capacity and accelerator-system delivery
  - **Opus 5.5** more capacity unlocking backlog
  - **GPT-6.1 Sol** Faster conversion of cloud commitments and additional large contracts
  - **GLM 5.3** TPU system shipments ramp faster than planned and new mega-deals pull revenue forward from 2027 into late 2026
  - **Muse Spark Open** Cloud backlog 514 billion converts faster with new customers doubling
  - **DeepSeek V4.1 Flash** Faster conversion of the $514 billion Cloud backlog and earlier TPU system revenue
- ****Search holds near 17% growth as AI Mode advertising scales.** AI Mode passed one billion monthly users, and the company says it adds to Search queries overall. *6 of 6***
  - **Kimi K3.1 Pro** Sustained strong Search and subscription growth
  - **GLM 5.3** Search sustains 17%+ growth as AI Mode monetization expands ad coverage to previously unmonetizable queries
  - **DeepSeek V4.1 Flash** Search monetisation of longer, previously unmonetisable AI queries
  - **Opus 5.5** AI Mode and agentic commerce advertising
  - **Muse Spark Open** Search queries hit new highs on AI Mode with over one billion users driving upside
  - **GPT-6.1 Sol** Improved search advertising coverage and subscription pricing
- ****Scale offsets infrastructure costs sooner, holding the margin near 35%.** Cloud's operating margin rose to 35.6% in the second quarter from 20.7% a year earlier. *6 of 6***
  - **Kimi K3.1 Pro** Better-than-expected Cloud operating leverage
  - **Opus 5.5** operating leverage in Cloud
  - **GLM 5.3** Cloud operating margin keeps expanding beyond 35.6% as internal capacity replaces costly third-party bridging capacity
  - **Muse Spark Open** Operating leverage from AI efficiencies and pricing power
  - **DeepSeek V4.1 Flash** Operating leverage outrunning depreciation and energy costs
  - **GPT-6.1 Sol** Efficient internal capacity replacing expensive rented capacity

If all of this happens

**$517B** sales · +3% vs base

**$12.80** EPS · +9% vs base

Revenues would have to reach about $517B, 3.5% above the base, with earnings per share near $12.80, about 8.8% higher. Capacity would arrive sooner, letting Cloud ship chip systems and convert backlog faster, while Search held near 17% on AI Mode. Kimi's high is furthest, at $536B, because it combines faster capacity and chip shipments with Search monetization that “remains strong” on a mid-teens Services base, and scale efficiencies that “offset infrastructure costs sooner”.

First-half growth quickened each quarter, and Cloud grew 82% in the second quarter.

The six highs run from $506B (DeepSeek V4.1 Flash) to $536B (Kimi K3.1 Pro).

First sign The third-quarter report, the next one due: Cloud revenues above $30B, or Search growing faster than 17%.

#### Watch

Two reports remain, for the third and fourth quarters. Cloud revenue and backlog, Search growth against last year's acceleration, and the cost of rented capacity decide where 2026 lands. A higher capital spending range or more stock sales would weigh on free cash flow and earnings per share.

##### Sales

PricedIn **$608B** $543B – $680B

$543B $574B $608B $641B $680B

##### adjusted EPS

PricedIn **$13.53** $10.93 – $16.44

$10.93 $12.20 $13.53 $14.91 $16.44

**Middle** the forecast **50% range** 1 in 4 chance the result comes in below this bar, 1 in 4 above it. Half the time it lands inside. **80% range** 1 in 10 chance below this line, 1 in 10 above it. These are the low and high printed by each number.

The models agree 2027 grows about 19% to 24% and margins dip near 33%. Muse Spark Open is lowest at $590B; it “kept ad-growth assumptions conservative”. Kimi is highest at $631B, expecting Services to sustain “mid-teens growth”. For earnings per share, GLM is lowest at $12.80, its margin dipping “to about 32.5% in 2027 under the depreciation ramp”; Muse Spark Open is highest at $14.20. Sol's range is widest, $535B to $735B: an “energy and financing shock” at one end, capacity “delivered sooner” at the other.

The models expect 2027 revenues of about $608B, as most chip system sales and much of Cloud's backlog turn into revenue.

Sales **$608B** +22% on FY2026

*50% range* $574B – $641B *80% range* $543B – $680B

Adjusted EPS **$13.53** +15% on FY2026

*50% range* $12.20 – $14.91 *80% range* $10.93 – $16.44

Base case

Operating margin **33.4%** 34.4% in FY2026

#### Investment and cash generation

Base case

Capital spending **$266B**

Free cash flow **−$28.3B**

Free cash flow: operating cash flow less capital spending.

2027 is when the company says most of its chip system revenue arrives. Cloud's backlog stood at $514B in June, and management expects just over half of it to become revenue within 24 months, before any new contracts or usage above commitments.

The models' average is $608B, up 22% on their 2026 figure. Their Cloud forecasts run from about $154B to $180B for the year, and they expect Google Services to keep growing in the low-to-mid teens.

Costs climb with the build. The company says 2027 capital spending will rise significantly; the models put it near $266B on average, from $255B to $290B. Depreciation follows, so the operating margin eases to about 33%, and free cash flow stays negative, near −$28.3B.

Earnings per share average about $13.53. The range is wide, from $10.93 to $16.44, because a delay or speed-up in chip deliveries and capacity could move 2027 sales by $40B or more, in Opus's estimate.

The models expect about $13.53 a share. Revenues grow 22% but earnings per share only 15%, because depreciation lowers the margin toward 33%, interest costs rise with new debt, and the share count keeps creeping up after the June stock sale.

#### Sales by year

FY2023 $307B

FY2024 $350B +14%

FY2025 $403B +15%

FY2026 $500B +24%

FY2027 $608B +22%

FY2028 $713B +17%

The forecast years are the models' average.

#### Google Cloud backlog at each quarter's end

September 2025 $155B

December 2025 $240B

March 2026 $460B

June 2026 $514B

**Where each number comes from**

**September 2025**

“Google Cloud's backlog increased 46% sequentially and 82% year-over-year, reaching $155 billion at the end of the third quarter.”

**December 2025**

“Backlog grew by 55% quarter over quarter to $240 billion representing a wide breadth of customers driven by demand for AI products.”

**March 2026**

“Cloud accelerated again this quarter due to strong demand for our AI products and infrastructure. Revenue grew 63%, exceeding $20 billion for the first time and our backlog nearly doubled quarter-on-quarter to over $460 billion.”

**June 2026**

“Google Cloud's backlog increased by more than $50 billion sequentially, reaching $514 billion in the second quarter.”

Signed Cloud contracts not yet booked as revenue. The models' 2027 and 2028 Cloud forecasts rest on how fast this converts.

#### What it rests on

**Most chip revenue in 2027**

The vast majority of chip system revenue is expected in 2027 “We anticipate the vast majority of the revenues from these agreements will be realized in 2027.” Second-quarter earnings call

**2027 capital spending**

Expected to increase significantly from 2026, details later “we continue to expect our CapEx to increase significantly in 2027, and we'll provide more details at a later date.” Second-quarter earnings call

**Backlog conversion**

Just over half of backlog expected as revenue within 24 months “for our broad mix of customers, and we expect to recognize just over 50% of the total” Second-quarter earnings call

**Spending above commitments**

Customers typically spend more than 50% above what they commit “when a customer gives us a commitment, say, for $100, typically they spend more than 50% more than that.” Cloud chief, September conference

**Depreciation climbing**

$21.1B in 2025, up 38% from $15.3B in 2024 “depreciation increased by nearly billion dollars or 38% from $15.3 billion in 2024 to $21.1 billion in 2025” Fourth-quarter 2025 earnings call

**Server payback**

Under two years for AI servers in aggregate “Second, our payback period on AI servers in aggregate is less than two years, and on” Cloud chief, September conference

#### Low and high

Base the middle forecast

Key assumptions

- ****Cloud reaches about $154B to $180B as backlog converts and chip system sales mostly land.** Management expects most chip system revenue in 2027 and just over half of backlog within 24 months. *6 of 6***
  - **Opus 5.5** chip system sales mostly land in 2027
  - **GLM 5.3** TPU system sales ramping into 2027
  - **GPT-6.1 Sol** Cloud capacity expansion and 2027 chip-system delivery ramp
  - **Kimi K3.1 Pro** Cloud backlog conversion and accelerator-system ramp
  - **DeepSeek V4.1 Flash** Timing of TPU system sales, small in 2026 and mostly in 2027
  - **Muse Spark Open** Cloud backlog 514B with >50% to be recognized over 24 months and TPU system sales ramping in 2027
- ****Google Services grows in the low-to-mid teens.** AI features expand Search queries while subscriptions grow faster than advertising. *6 of 6***
  - **DeepSeek V4.1 Flash** Google Services grows about 13% in 2026 and 12% in 2027
  - **GPT-6.1 Sol** Services grow at low-teens rates
  - **Kimi K3.1 Pro** Mid-teens Google Services growth
  - **GLM 5.3** Search and YouTube advertising growing low-to-mid teens
  - **Opus 5.5** Search growing about 15% with AI features
  - **Muse Spark Open** Search and YouTube sustain double-digit growth on AI Overviews and AI Mode
- ****Capital spending rises to about $266B, keeping free cash flow negative.** The company said 2027 spending will increase significantly from 2026's $200B. *6 of 6***
  - **Opus 5.5** $290 billion in 2027
  - **GPT-6.1 Sol** I forecast capital spending of about $275 billion in 2027
  - **GLM 5.3** I assume about $260 billion
  - **Kimi K3.1 Pro** a significant 2027 spending increase
  - **Muse Spark Open** drove base capex 200B and 260B
  - **DeepSeek V4.1 Flash** so about $258 billion in 2027
- ****Depreciation eases the operating margin to about 33%, from 34% in 2026.** Depreciation follows spending with a lag, and it rose 38% in 2025. *5 of 6***
  - **GLM 5.3** dips to about 32.5% in 2027 under the depreciation ramp
  - **Opus 5.5** I let margins slip to 33% in 2027
  - **DeepSeek V4.1 Flash** my operating margin slips from 34.8% in 2026 to 34.0% in 2027
  - **GPT-6.1 Sol** 34.5% next year
  - **Kimi K3.1 Pro** to taper Cloud growth through 2028 and to temper consolidated margins
- ****The share count keeps rising, with no buybacks.** June's stock sale raised $49.6B, and an at-the-market program can sell more shares. *4 of 6***
  - **GLM 5.3** roughly 12.7-13.0 billion across 2027-2028
  - **GPT-6.1 Sol** 12.8 billion in 2027
  - **Opus 5.5** about 12.47 billion by 2028
  - **Kimi K3.1 Pro** support a modestly rising diluted share count

If all of this happens

**$608B** sales

**$13.53** adjusted EPS

**$266B** Capital spending

**−$28.3B** Free cash flow

Cash walk not supplied by every model.

Low about 1 chance in 10 of worse

What the low case assumes

- ****Cloud growth falls toward 35% to 45%, against about 60% in the base, as chip shipments slip.** Chip system revenue is booked on delivery, and deliveries depend on power, land and supply. *6 of 6***
  - **GLM 5.3** Cloud growth falls back toward 35-45% as TPU shipments slip and enterprise AI spending pauses to digest
  - **GPT-6.1 Sol** Delayed capacity delivery and chip-system shipments
  - **Opus 5.5** Cloud capacity shortfall and slower backlog conversion
  - **Kimi K3.1 Pro** Delayed Cloud capacity and accelerator-system deliveries
  - **Muse Spark Open** Cloud expansion slows and third-party capacity creates margin pressure
  - **DeepSeek V4.1 Flash** Google Cloud sequential growth stalling near $1.5-2 billion a quarter as TPU system deliveries slip into 2027
- ****A consumer downturn cuts advertising budgets across Search and YouTube.** More than 70% of 2025 revenues came from advertising, and oil neared $115 a barrel at the cutoff. *6 of 6***
  - **GLM 5.3** A consumer downturn (sentiment 51.7, mortgage 7.3%) cuts advertising budgets across Search, YouTube and Network
  - **DeepSeek V4.1 Flash** Advertising slowing as expensive energy squeezes consumers and a five per cent ten-year yield trims budgets
  - **GPT-6.1 Sol** an energy and financing shock weakens advertising
  - **Opus 5.5** high oil prices and a stronger dollar trim advertising budgets
  - **Kimi K3.1 Pro** Slower Search and YouTube advertising growth
  - **Muse Spark Open** Search and YouTube growth slows to low single digits on macro and lapping
- ****Depreciation and rented capacity outrun revenue, pushing the margin below 33%.** Spending of $200B in 2026 lands on the income statement in 2027 whether or not revenue does. *6 of 6***
  - **DeepSeek V4.1 Flash** Depreciation from about $200 billion of 2026 spending running ahead of revenue
  - **Opus 5.5** depreciation and rented capacity costs
  - **GPT-6.1 Sol** High depreciation, energy and external-capacity costs
  - **GLM 5.3** Operating margin compresses as depreciation, third-party capacity costs and supply-chain inflation outrun revenue leverage
  - **Kimi K3.1 Pro** Higher infrastructure and energy costs
  - **Muse Spark Open** Depreciation and AI talent costs rise faster than sales
- ****Search remedies and other legal charges begin to bite.** A December 2025 judgment requires sharing search data and syndication with competitors; both sides are appealing. *3 of 6***
  - **GPT-6.1 Sol** Adverse legal remedies or additional charges
  - **GLM 5.3** DOJ search remedies
  - **Kimi K3.1 Pro** Foreign-exchange and regulatory pressure

If all of this happens

**$543B** sales · −11% vs base

**$10.93** EPS · −19% vs base

Revenues would have to come in near $543B, about 11% under the base, with earnings per share near $10.93, about 19% lower. Chip system deliveries would slip, enterprises would pause AI spending, and a consumer downturn would cut ad budgets while depreciation and rented capacity outran revenue. Muse Spark Open's low is furthest, at $520B, starting from the lowest 2026 base and the most conservative advertising view. Sol's earnings low of $10.00 is the deepest, as it lets an energy and financing shock, legal restrictions and costs that do not fall with sales hit together.

Free cash flow fell from $73.3B in 2025 to −$5.9B in the second quarter of 2026, as quarterly capital spending reached $44.9B.

The six lows run from $520B (Muse Spark Open) to $562B (GLM 5.3).

First sign The fourth-quarter report: a 2027 capital spending range far above $270B, or chip system deliveries pushed later.

High about 1 chance in 10 of better

What the high case assumes

- ****Capacity arrives sooner and new large contracts convert fast, lifting Cloud above the base.** New customer wins and deals over $100M doubled from a year earlier, the Cloud chief said. *6 of 6***
  - **GPT-6.1 Sol** Faster conversion of cloud commitments and additional large contracts
  - **Opus 5.5** more capacity unlocking backlog
  - **GLM 5.3** TPU system shipments ramp faster than planned and new mega-deals pull revenue forward from 2027 into late 2026
  - **Kimi K3.1 Pro** Faster Cloud capacity and accelerator-system delivery
  - **Muse Spark Open** Cloud backlog 514 billion converts faster with new customers doubling
  - **DeepSeek V4.1 Flash** Faster conversion of the $514 billion Cloud backlog and earlier TPU system revenue
- ****Search sustains about 17% growth as AI Mode advertising scales.** AI Mode passed one billion monthly users, and the company says it adds to queries overall. *6 of 6***
  - **GLM 5.3** Search sustains 17%+ growth as AI Mode monetization expands ad coverage to previously unmonetizable queries
  - **DeepSeek V4.1 Flash** Search monetisation of longer, previously unmonetisable AI queries
  - **Opus 5.5** AI Mode and agentic commerce advertising
  - **Muse Spark Open** Search queries hit new highs on AI Mode with over one billion users driving upside
  - **Kimi K3.1 Pro** Sustained strong Search and subscription growth
  - **GPT-6.1 Sol** Improved search advertising coverage and subscription pricing
- ****Internal capacity replaces rented capacity, holding the margin near 34% to 35%.** Cloud's operating margin rose to 35.6% in the second quarter from 20.7% a year earlier. *6 of 6***
  - **GPT-6.1 Sol** Efficient internal capacity replacing expensive rented capacity
  - **GLM 5.3** Cloud operating margin keeps expanding beyond 35.6% as internal capacity replaces costly third-party bridging capacity
  - **Opus 5.5** operating leverage in Cloud
  - **Kimi K3.1 Pro** Better-than-expected Cloud operating leverage
  - **Muse Spark Open** Operating leverage from AI efficiencies and pricing power
  - **DeepSeek V4.1 Flash** Operating leverage outrunning depreciation and energy costs
- ****Free cash flow recovers sooner, with 2027 spending guided below $240B.** Chip hardware placed in customers' data centers brings strong cash flow, the Cloud chief said. *2 of 6***
  - **DeepSeek V4.1 Flash** Capital spending guided below $240 billion for 2027, easing free cash flow
  - **Kimi K3.1 Pro** Earlier free-cash-flow recovery

If all of this happens

**$680B** sales · +12% vs base

**$16.44** EPS · +21% vs base

Revenues would have to reach about $680B, 12% above the base, with earnings per share near $16.44, about 21% higher. New contracts would add to the June backlog, rented and new capacity would free pent-up demand, chip systems would ship faster, and Search would hold near 17% as AI Mode advertising scales. Sol's high is furthest, at $735B, because it lets demand convert rapidly into both usage and chip system sales on a Cloud middle already near $180B, the largest of any model.

Customers typically spend more than 50% above their commitments, and backlog grew more than $50B in the second quarter alone.

The six highs run from $634B (DeepSeek V4.1 Flash) to $735B (GPT-6.1 Sol).

First sign The fourth-quarter report: Cloud growth above 90% with backlog additions above $50B, and a 2027 spending range below $240B.

#### Watch

The 2027 capital spending range at the fourth-quarter report, quarterly Cloud revenue and backlog, and the pace of chip system deliveries. A range far above $270B or slipping deliveries point lower; backlog well above $550B and Search near 17% point higher.

##### Sales

PricedIn **$713B** $602B – $835B

$602B $655B $713B $771B $835B

##### adjusted EPS

PricedIn **$15.84** $11.87 – $20.27

$11.87 $13.79 $15.84 $17.92 $20.27

**Middle** the forecast **50% range** 1 in 4 chance the result comes in below this bar, 1 in 4 above it. Half the time it lands inside. **80% range** 1 in 10 chance below this line, 1 in 10 above it. These are the low and high printed by each number.

The models agree growth slows to about 15% to 20%, with margins of 32% to 35%. Muse Spark Open is lowest at $680B, having “kept ad-growth assumptions conservative”. Kimi is highest at $759B, holding Services to “mid-teens growth” while Cloud “then normalizes”. For earnings per share GLM is lowest at $14.60 and Sol highest at $16.79. Sol's range is far wider, $570B to $970B: delayed capacity and an “energy and financing shock” at the bottom, operating leverage “lifting earnings substantially” at the top.

The models expect 2028 revenues of about $713B, as Cloud growth cools from its 2027 peak and spending growth eases.

Sales **$713B** +17% on FY2027

*50% range* $655B – $771B *80% range* $602B – $835B

Adjusted EPS **$15.84** +17% on FY2027

*50% range* $13.79 – $17.92 *80% range* $11.87 – $20.27

Base case

Operating margin **33.8%** 33.4% in FY2027

#### Investment and cash generation

Base case

Capital spending **$300B**

Free cash flow **−$5.83B**

Free cash flow: operating cash flow less capital spending.

By 2028 the question shifts from whether demand exists to whether the build-out pays. The models expect Cloud's growth to cool after the 2027 chip system surge, with Cloud between about $205B and $253B for the year.

The average is $713B, up 17% on their 2027 figure, with an operating margin near 34%. Capital spending keeps rising, to about $300B on average, so free cash flow averages near −$5.8B, though three models see it turning positive.

The company has given no figures this far out. It said only that 2027 spending will rise significantly, and its Cloud chief said AI servers pay back in under two years, which the models read as support for margins holding.

The spread is widest here. Earnings per share average $15.84 but range from $11.87 to $20.27, because delays or speed in capacity, chip shipments and advertising compound over three years.

The models expect about $15.84 a share, up 17%. Revenue growth and a steady margin near 34% do the work, while depreciation from earlier spending, interest costs and a share count near 13 billion hold it back.

#### Each model's growth in FY2028

Opus 5.5 +16%

GPT-6.1 Sol +20%

GLM 5.3 +15%

Kimi K3.1 Pro +20%

Muse Spark Open +15%

DeepSeek V4.1 Flash +17%

Growth of each model's middle on its own forecast for the year before.

#### What it rests on

**Server payback**

Under two years for AI servers in aggregate “Second, our payback period on AI servers in aggregate is less than two years, and on” Cloud chief, September conference

**2027 capital spending**

Expected to increase significantly from 2026, details later “we continue to expect our CapEx to increase significantly in 2027, and we'll provide more details at a later date.” Second-quarter earnings call

**Spending above commitments**

Customers typically spend more than 50% above what they commit “when a customer gives us a commitment, say, for $100, typically they spend more than 50% more than that.” Cloud chief, September conference

**Search remedies**

A December 2025 judgment restricts distribution and requires sharing search data with competitors “In August 2024, the US District Court for the District of Columbia ruled against Google, and in December 2025, entered a final judgment requiring remedies, which, among other things, imposes restrictions on how we distribute our services and requires us to share certain search data with and offer syndication services to certain competitors.” 2025 annual report

**Supply limits**

Scaling is increasingly limited by power, water and land “Our ability to scale our technical infrastructure is increasingly constrained by the availability of power, water, and land.” 2025 annual report risk factors

**Advertising share**

More than 70% of 2025 revenues came from online advertising “We generated more than 70% of total revenues from online advertising in 2025.” 2025 annual report risk factors

#### Low and high

Base the middle forecast

Key assumptions

- ****Google Services keeps growing in the low-to-mid teens.** Subscriptions grow faster than advertising, and AI features keep expanding Search queries. *6 of 6***
  - **DeepSeek V4.1 Flash** Google Services grows about 13% in 2026 and 12% in 2027
  - **GPT-6.1 Sol** Services grow at low-teens rates
  - **Kimi K3.1 Pro** Mid-teens Google Services growth
  - **GLM 5.3** Search and YouTube advertising growing low-to-mid teens
  - **Opus 5.5** Search growing about 15% with AI features
  - **Muse Spark Open** Search and YouTube sustain double-digit growth on AI Overviews and AI Mode
- ****Capital spending climbs to about $300B, so free cash flow hovers near zero.** Spending growth slows after 2027's jump, and chip hardware sales bring cash on delivery. *4 of 6***
  - **GLM 5.3** turning slightly positive in 2028 as TPU hardware sales bring cash on delivery
  - **GPT-6.1 Sol** $315 billion in 2028
  - **DeepSeek V4.1 Flash** $285 billion in 2028
  - **Kimi K3.1 Pro** turning positive in 2028
- ****Cloud growth cools to about 30%, with Cloud near $205B to $253B.** Backlog conversion keeps adding revenue, but growth slows on a much larger base. *4 of 6***
  - **GLM 5.3** about 30% in 2028 as the backlog converts
  - **Kimi K3.1 Pro** Cloud grows rapidly through 2027 from backlog and accelerator systems, then normalizes
  - **Opus 5.5** $205 billion in 2028
  - **GPT-6.1 Sol** $253 billion in 2028
- ****The operating margin holds near 34% as revenue catches up with depreciation.** AI servers pay back in under two years, the Cloud chief said in September. *4 of 6***
  - **GLM 5.3** recovers to about 33% in 2028
  - **GPT-6.1 Sol** 35.3% in 2028
  - **DeepSeek V4.1 Flash** I do not model a Cloud margin collapse even as spending surges
  - **Kimi K3.1 Pro** scale largely offsets depreciation and near-term capacity costs

If all of this happens

**$713B** sales

**$15.84** adjusted EPS

**$300B** Capital spending

**−$5.83B** Free cash flow

Cash walk not supplied by every model.

Low about 1 chance in 10 of worse

What the low case assumes

- ****Cloud growth falls back as chip shipments slip and enterprises pause AI spending.** Capacity depends on power, water and land, and chip revenue is booked only on delivery. *6 of 6***
  - **GLM 5.3** Cloud growth falls back toward 35-45% as TPU shipments slip and enterprise AI spending pauses to digest
  - **GPT-6.1 Sol** Delayed capacity delivery and chip-system shipments
  - **Opus 5.5** Cloud capacity shortfall and slower backlog conversion
  - **Kimi K3.1 Pro** Delayed Cloud capacity and accelerator-system deliveries
  - **Muse Spark Open** Cloud expansion slows and third-party capacity creates margin pressure
  - **DeepSeek V4.1 Flash** Google Cloud sequential growth stalling near $1.5-2 billion a quarter as TPU system deliveries slip into 2027
- ****A consumer downturn cuts advertising budgets.** Oil neared $115 a barrel at the cutoff and consumer sentiment was 51.7, against 58.2 a year earlier. *6 of 6***
  - **GLM 5.3** A consumer downturn (sentiment 51.7, mortgage 7.3%) cuts advertising budgets across Search, YouTube and Network
  - **GPT-6.1 Sol** an energy and financing shock weakens advertising
  - **DeepSeek V4.1 Flash** Advertising slowing as expensive energy squeezes consumers and a five per cent ten-year yield trims budgets
  - **Opus 5.5** high oil prices and a stronger dollar trim advertising budgets
  - **Kimi K3.1 Pro** Slower Search and YouTube advertising growth
  - **Muse Spark Open** Search and YouTube growth slows to low single digits on macro and lapping
- ****Depreciation and energy costs outrun revenue, pushing the margin well below 34%.** Spending near $300B a year keeps adding depreciation whether or not revenue follows. *6 of 6***
  - **GPT-6.1 Sol** Infrastructure depreciation, rented capacity and research spending do not fall as fast as sales.
  - **Opus 5.5** depreciation and rented capacity costs
  - **GLM 5.3** Operating margin compresses as depreciation, third-party capacity costs and supply-chain inflation outrun revenue leverage
  - **Kimi K3.1 Pro** Higher infrastructure and energy costs
  - **Muse Spark Open** Depreciation and AI talent costs rise faster than sales
  - **DeepSeek V4.1 Flash** Depreciation from about $200 billion of 2026 spending running ahead of revenue
- ****Search slows to single digits as AI assistants and search remedies take hold.** The search judgment requires data sharing and syndication to competitors; appeals play out over 2027 and 2028. *4 of 6***
  - **GLM 5.3** Search decelerates to single digits under AI substitution and DOJ search remedies (data sharing, syndication, distribution restrictions)
  - **GPT-6.1 Sol** Adverse legal remedies or additional charges
  - **Kimi K3.1 Pro** Foreign-exchange and regulatory pressure
  - **DeepSeek V4.1 Flash** it is one of the reasons the low case exists rather than a reason to cut the middle

If all of this happens

**$602B** sales · −16% vs base

**$11.87** EPS · −25% vs base

Revenues would have to land near $602B, about 16% under the base, with earnings per share near $11.87, about 25% lower. The AI build would have to digest: Cloud growth falls back as enterprises pause, Search slows to single digits under AI substitution and search remedies, and a consumer downturn cuts ad budgets while depreciation keeps climbing. Sol's low is furthest, at $570B and $9.50 a share, because it stacks an energy and financing shock, legal restrictions and costs that do not fall as fast as sales on top of delayed capacity.

More than 70% of 2025 revenues came from online advertising, so an ad downturn hits most of the company at once.

The six lows run from $570B (GPT-6.1 Sol) to $625B (DeepSeek V4.1 Flash).

First sign Quarterly Search growth falling toward single digits, or Cloud backlog failing to grow from one quarter to the next.

High about 1 chance in 10 of better

What the high case assumes

- ****Faster capacity and new large contracts keep Cloud growing well above the base's 30%.** New customer wins and large deals doubled from a year earlier, the Cloud chief said. *6 of 6***
  - **GPT-6.1 Sol** Faster conversion of cloud commitments and additional large contracts
  - **Opus 5.5** more capacity unlocking backlog
  - **GLM 5.3** TPU system shipments ramp faster than planned and new mega-deals pull revenue forward from 2027 into late 2026
  - **Kimi K3.1 Pro** Faster Cloud capacity and accelerator-system delivery
  - **Muse Spark Open** Cloud backlog 514 billion converts faster with new customers doubling
  - **DeepSeek V4.1 Flash** Faster conversion of the $514 billion Cloud backlog and earlier TPU system revenue
- ****Search monetization strengthens as AI Mode widens the queries that carry ads.** AI Mode passed one billion monthly users, and the company says it adds to Search queries overall. *6 of 6***
  - **GPT-6.1 Sol** Improved search advertising coverage and subscription pricing
  - **GLM 5.3** Search sustains 17%+ growth as AI Mode monetization expands ad coverage to previously unmonetizable queries
  - **DeepSeek V4.1 Flash** Search monetisation of longer, previously unmonetisable AI queries
  - **Opus 5.5** AI Mode and agentic commerce advertising
  - **Muse Spark Open** Search queries hit new highs on AI Mode with over one billion users driving upside
  - **Kimi K3.1 Pro** Sustained strong Search and subscription growth
- ****Operating leverage lifts the margin above the base's 34%.** Cloud's operating margin rose to 35.6% in the second quarter from 20.7% a year earlier. *6 of 6***
  - **GPT-6.1 Sol** Efficient internal capacity replacing expensive rented capacity
  - **GLM 5.3** Cloud operating margin keeps expanding beyond 35.6% as internal capacity replaces costly third-party bridging capacity
  - **Opus 5.5** operating leverage in Cloud
  - **Kimi K3.1 Pro** Better-than-expected Cloud operating leverage
  - **Muse Spark Open** Operating leverage from AI efficiencies and pricing power
  - **DeepSeek V4.1 Flash** Operating leverage outrunning depreciation and energy costs
- ****Enterprise AI products and subscriptions grow faster than expected.** The Cloud chief cited more than 17 product lines each above a billion dollars, growing quickly. *3 of 6***
  - **GPT-6.1 Sol** Strong adoption of enterprise agents, data and security products
  - **GLM 5.3** Subscriptions (Google One AI plans, YouTube Premium) compound above 20% growth
  - **Kimi K3.1 Pro** Sustained strong Search and subscription growth

If all of this happens

**$835B** sales · +17% vs base

**$20.27** EPS · +28% vs base

Revenues would have to reach about $835B, 17% above the base, with earnings per share near $20.27, about 28% higher. Capacity would arrive sooner, new large contracts would keep converting, Search monetization and subscription pricing would hold, and cheaper in-house chips would replace rented capacity. Sol's high is furthest, at $970B and $26.00 a share, because it lets these compound from the largest Cloud base, near $253B in its middle, with operating leverage lifting earnings substantially.

The Cloud chief said in September that AI servers pay back in under two years and that customers often spend well above their commitments.

The six highs run from $758B (DeepSeek V4.1 Flash) to $970B (GPT-6.1 Sol).

First sign Cloud growth staying above 80% with stable margins, and operating cash flow catching up with capital spending sooner.

#### Watch

The 2027 capital spending range and quarterly backlog first, then rulings in the search and ad technology cases. Backlog that keeps growing and spending growth that slows point higher; Search slowing toward single digits or depreciation outrunning revenue point lower.

The years at a glance · sales

FY2023 **$307B**

FY2024 **$350B**

FY2025 **$403B**

FY2026 **$500B** $484B – $517B

FY2027 **$608B** $543B – $680B

FY2028 **$713B** $602B – $835B

$400B $600B $800B

50% range 80% range

The years at a glance · adjusted EPS

FY2023 **$5.80**

FY2024 **$8.04**

FY2025 **$10.81**

FY2026 **$11.76** $10.75 – $12.80

FY2027 **$13.53** $10.93 – $16.44

FY2028 **$15.84** $11.87 – $20.27

$5.00 $10.00 $15.00 $20.00

50% range 80% range