GOOG · Research
Alphabet Raised $138B From Outside Investors in Nine Months Because Its Own Cash No Longer Covers the Buildout
Google Cloud grew 103% and the backlog reached $560.8B. But capex passed operating cash, owner cash turned negative $28.1B, buybacks stopped, and Alphabet sold debt, preferred stock and new common shares to pay for compute.
- Google Cloud revenue growth
- 103%
- Owner cash
- -$28.1B
- Outside capital raised
- $138.3B
1. Summary: what this piece answers
Alphabet is the largest owner of consumer attention and the fastest-growing large cloud on earth. Google Cloud revenue more than doubled in the third quarter of 2026, rising 103% to $30.7B, and total revenue rose 25% to $127.6B [2].
This piece asks a narrower question: who is paying for the buildout? For two decades the answer was Alphabet's own cash flow.
In 2026 that stopped being true.
Here is what you will learn. In the first nine months of 2026 Alphabet spent $131,953M on property and equipment, more than its $127,149M of operating cash flow [2].
After stock pay of $23,338M, owner cash, meaning operating cash flow less all capex less stock pay, was negative $28,142M, our calculation [2]. A year earlier the same nine months produced positive owner cash of $30,833M [2][3].
To close the gap Alphabet raised $138.3B of outside capital in nine months: $76.3B of net new debt, $19.1B of mandatory convertible preferred stock and $42.9B of new common stock, including a $10.0B private placement with an affiliate of Berkshire Hathaway [1][2]. Buybacks, which took $39,810M in the first nine months of 2025, went to zero
2. Thesis and the toll-gate test
| Test | Result | Evidence |
|---|---|---|
| Owns the record | Pass | Search index, YouTube library, Gmail and Workspace documents, Android and Play accounts; the Gemini app reached one billion users [2] |
| Owns the connection | Partial | Alphabet still pays $16,366M a quarter in traffic acquisition costs to partners for distribution, 20.0% of advertising revenue, and court remedies limit some default deals [1][2] |
| Prices close to outcomes | Pass | Search advertising is priced per click or conversion; Cloud model APIs are priced per token, and token volume is nearly 32 billion a minute, more than four times a year ago [2] |
| R&D on the gate vs features | Pass | R&D of $55,891M in nine months, 15.6% of revenue, funds the Gemini models and in-house TPU chips that now sell as systems to outside customers [2][3] |
Three passes and one partial. The partial matters: Alphabet owns the destination, but it still rents part of the road that leads there.
3. What Alphabet sells
Google Services is the consumer franchise: Search and other properties, YouTube, Android, Chrome, Maps, Play, Gmail, hardware and subscriptions such as YouTube Premium and Google One [1]. It earns most of its money from advertising placed next to intent, either a query or a video.
Google Cloud sells infrastructure (compute, storage, networking and TPU systems), the Gemini models through APIs, data and security software including Wiz, and Workspace subscriptions to businesses [1][2]. Other Bets holds Waymo, Verily and smaller ventures [1].
New products ship fast. In the quarter Alphabet launched Gemini 4 Argon and Carbon, which management describes as leading at the frontier, and said TPU system sales to outside customers were a meaningful contributor to Cloud growth [2].
YouTube Music and Premium passed 150 million subscribers including trials [2]. In May 2026 Alphabet agreed with Blackstone to form a US joint venture selling data center capacity, networking, hardware and software as a single compute service; Blackstone committed up to $5.0B of capital and Alphabet committed to backstop data center and energy infrastructure [1].
4. How each line makes money, customers and pricing
Search is an auction. Advertisers bid on queries and pay when users click or convert, so revenue tracks commercial intent rather than seats.
Management says query growth in the quarter was the fastest in five years, including commercial queries [2]. YouTube sells video ads and subscriptions.
Network, the ads Alphabet places on other publishers' sites, is barely growing and is the weakest line. Cloud is a mix of consumption billing (compute and tokens) and committed contracts; most of the backlog is multi-year commitments [1].
The main cost of the ad business is traffic acquisition: payments to Apple, phone makers, browsers and Network partners for distribution. TAC rose 10.0% to $16,366M, slower than advertising revenue growth of 10.6%, so the toll Alphabet pays for its connection is not getting more expensive [2].
5. Growth
Cloud did the heavy lifting: it added $15,552M of the $25,231M increase in quarterly revenue, or 62% of growth, our calculation [2]. Services added $9,250M.
Alphabet gives no revenue guide, so the backlog is the best forward indicator.
6. Retention and the proxy we use
Alphabet does not disclose net revenue retention. For Cloud we use the backlog as the proxy: $560.8B of Cloud commitments not yet recognized, against trailing Cloud revenue that is a small fraction of that [1].
For consumers the measures are paying subscribers, 150 million for YouTube Music and Premium, and one billion Gemini app users [2]. Search query growth at a five-year high is the clearest sign that the core habit is not eroding as answer engines spread [2].
7. Operations: segments, costs, headcount and efficiency
The operating story is Cloud margin. A year ago Cloud earned $3,594M on $15,157M, a 23.7% margin; this quarter it earned $11,121M, a 36.2% margin, so Cloud profit tripled [2].
The cost of that profit sits partly elsewhere: Alphabet-level costs, mostly shared model research and development, rose 71% to $7,655M [2]. Group operating margin was 31.8% [2].
Depreciation rose 49% in the quarter to $8,369M, and property and equipment on the balance sheet rose from $246,597M to $372,827M in nine months, so depreciation will keep climbing [2].
Alphabet had 204,757 employees at Sep 30, 2026, up 7.7% from 190,167 a year earlier [2]. Trailing revenue per employee is about $2.30M, our calculation.
Stock pay was $8,630M in the quarter, 6.8% of revenue [2]. Cash paid for taxes on vested stock awards was $17,424M in nine months, and the ATM program is primarily intended to fund those taxes, according to the 10-Q [1]
8. Cash flow and owner cash
| Line | Nine months 2026 | Trailing twelve months | Fiscal 2025 |
|---|---|---|---|
| Operating cash flow | $127,149M | $179,551M | $164,713M |
| Less property and equipment | $131,953M | $159,804M | $91,447M |
| Less stock pay | $23,338M | $30,409M | $24,953M |
| Owner cash | -$28,142M | -$10,662M | $48,313M |
| Capex as share of operating cash | 103.8% | 89% | 55.5% |
Sources: [2][3], our calculation. Free cash flow before stock pay was negative in both the second quarter (-$5,855M) and third quarter (-$9,065M) of 2026 [2].
Interest income of $4,579M over nine months, earned on corporate cash, would make owner cash worse still if stripped out as our standard requires [1]. Third-quarter operating cash flow fell 12.6% to $42,290M even as operating income rose 30% [2]; tax payments and working capital explain part of that, and the 10-Q does not isolate the rest.
GAAP net income was $201,239M for nine months, but that figure includes $126,755M of gains on equity and debt securities, mostly marks on private stakes [2]. Those gains produce no cash.
Owners should read profit and cash as two different stories this year.
9. Balance sheet and capital allocation
The financing calendar of 2026 [1][2]: senior notes throughout the year, including $28.7B net in the third quarter; a public offering of Class A and Class C shares and a concurrent private placement to a Berkshire Hathaway affiliate on Jun 4, together $30.5B net; 6.25% mandatory convertible preferred on Jun 5 for $19.0B net, converting into common stock around May 15, 2029; and a $40.0B at-the-market program from Jun 1, of which $12.4B was sold in the third quarter and $27.6B remains. Alphabet also paid $1.0B for capped calls to limit dilution from the preferred.
Our rule treats debt as leverage, and new shares as dilution, not as cash the business made. Alphabet raised more outside money than it spent on buildings and chips in excess of operating cash, so cash and marketable securities more than doubled to $265,098M [2].
That gives a large cushion. It also means owners are now partly funding the buildout with their own ownership share.
Interest expense rose from $438M to $2,778M over nine months [1].
10. Acquisitions in the last 24 months
Wiz closed Mar 11, 2026 for $29.5B in cash after adjustments, excluding post-close compensation; it sits in Google Cloud and extends multicloud security [1]. Intersect, a renewable energy developer, closed Mar 10, 2026 for $5.9B and speeds data center power [1].
The filings do not disclose the revenue of either business. Both were cash deals, and together they equal about 26% of the outside capital raised this year, our calculation.
11. Superintelligence angle
Mixed, leaning tailwind. Alphabet is one of very few companies that owns every layer: chips (TPU), models (Gemini), distribution (Search, Android, Chrome, YouTube) and a cloud that rents all of it to others.
Token volume rising fourfold in a year and a $560.8B Cloud backlog say demand is real [1][2]. The threat is to the oldest gate: if answer engines replace the search results page, the per-click auction that funds everything could compress.
So far the data cut the other way, with query growth at a five-year high [2]. The cost is capital intensity: the company that once converted most of its operating cash into free cash now spends almost all of it on plant.
12. Leadership and control
Sundar Pichai is chief executive officer of Google and Alphabet [2]. Alphabet has three common classes; Class B carries ten votes.
The proxy states that Larry Page and Sergey Brin control 52% of total voting power while owning less than 11% of outstanding voting stock [4]. The Vanguard Group is listed as a holder above 5% [4].
Control rests with the founders, which is why the board could pivot from buybacks to equity issuance in one year without a shareholder vote on the strategy. We name the chief financial officer by role only, as the sources pulled for this piece do not name the officer.
13. Risks and catalysts
Risks: returns on new capacity fall short of the cost of capital if model prices fall faster than usage grows; more equity issuance under the $27.6B left on the ATM program [1]; conversion of the preferred in 2029 adds shares [1]; search and ad technology antitrust remedies in the US, with a final search judgment entered December 2025 and the ad tech case ongoing [1]; marks on private stakes can reverse, swinging GAAP profit [2]. Catalysts: the Q4 2026 results and 2027 capex outlook, expected early February 2027 (unconfirmed date); the preferred dividend payment on Nov 15, 2026 (confirmed) [2]; Blackstone joint venture capacity coming online (unconfirmed timing) [1].
14. Scorecard, questions and predictions
| Metric | Value |
|---|---|
| R&D share of revenue | 15.6% over nine months [2][3] |
| GAAP net margin | 20.5% in Q3, inflated in 2026 by securities gains [2] |
| API quality | 5 of 5: public APIs with self-serve keys, public token pricing, agent tool integrations, SDKs in several languages and a large marketplace, our scoring |
| Cloud operating margin | 36.2% in Q3 [2] |
| Owner cash margin | -2.3% trailing, our calculation |
| Toll-gate score | 3 of 4 pass, 1 partial |
Questions for the owner. First, what Cloud margin on new capacity justifies spending more than operating cash?
Second, should equity issuance be judged against the buybacks of 2022 to 2025, which retired shares at very different terms? Third, at what point does depreciation growth outrun Cloud profit growth?
Fourth, conviction on a 1 to 5 scale: our view is 3, quality is clear but owner cash is negative, so we watch with a trigger of two consecutive quarters of positive free cash flow.
Predictions to register: Q4 2026 revenue grows 20% or more (65%); Cloud revenue growth stays above 80% in Q4 (60%); free cash flow before stock pay is negative again in Q4 2026 (60%); Alphabet sells more stock under the ATM program in Q4 (70%); buybacks stay at zero through Q4 2026 (75%).
Stock Profiles & Comparable Companies
GOOG Stock Profile →More Published Research
View All →Tenable Maps Every Weakness Agents Will Hunt. It Bought Back $516M of Stock on $81M of Owner Cash, and Borrowed the Difference
Tenable owns the exposure record and owner cash is finally positive, but growth is slowing toward 7%, the share count shrank 11.5% on convertible debt, and stockholders' equity turned negative.
SentinelOne Guards the Device Where Agents Run. Owners Lost $291M Last Year, and the Gap Is Widening
Revenue grows near 20% and the products target agent security, but stock pay was 5.6 times operating cash flow, owner cash worsened to -$291.0M, and a new convertible could add 46.1M shares.