2MAY

Alphabet books $35.7bn Q1 capex; backlog hits $460bn

4 min read

15:17UTC

Alphabet booked $35.7 billion of capex in Q1 2026 alone, with Google Cloud growing 63 per cent past $20 billion in a single quarter and total backlog nearly doubling quarter-on-quarter to $460 billion as Sundar Pichai called Google compute constrained.

← Back to Beijing court bans AI sackings as Big Tech burns cash Jump to analysis ↓

EconomicDeveloping

Key takeaway

Alphabet's $460 billion backlog measures unmet AI demand, and Pichai says compute capacity is what limits Google.

Alphabet reported Q1 2026 revenue of $109.9 billion, up 20 per cent year-on-year, with Google Cloud growing 63 per cent past $20 billion in a single quarter for the first time⁠¹. Quarter-one capex was $35.7 billion. Total backlog nearly doubled quarter-on-quarter to $460 billion, the disclosure that mattered most. CEO Sundar Pichai told the call that Google was "compute constrained in the near term".

The backlog figure is the cleanest signal in any of the Big Tech earnings this cycle. Backlog measures contracted future revenue that the company cannot currently fulfil because it does not yet have the compute capacity. A near doubling in a single quarter to $460 billion means demand is accelerating faster than Alphabet can build, and the $35.7 billion in Q1 capex is the spending rate that response requires. Annualised, Alphabet's capex run rate clears $140 billion, which puts it in the same range as Meta's revised guidance and within sight of Microsoft's $190 billion ceiling.

Pichai's compute constraint comment did the strategic work. When the CEO of a company spending $35 billion a quarter on data centres says he does not have enough capacity, the binding bottleneck on near-term AI revenue stops being model performance or product adoption and becomes physical infrastructure. Microsoft was running Azure at 40 per cent growth at the prior quarter and has since raised its own capex to $190 billion; Alphabet's backlog is the demand signal that justifies the $560 billion combined commitment across the four hyperscalers this year, and the matching pricing power that TSMC, SK Hynix and Samsung are converting into supplier margins. Some of the labour cost the hyperscalers are removing is being transferred into the supply chain rather than retained.

For savers exposed to the tracker funds that hold these four firms, the Pichai admission is reassuring on demand and unsettling on cost. Compute constraint at this scale means revenue is real; the capex it requires means free cash will not be returning to shareholders in the form most investors model.

Deep Analysis

In plain English

Alphabet is the company that owns Google. Most of its money comes from advertising, but it also runs Google Cloud, which businesses pay to use for computing power and AI services. In Q1 2026, Google Cloud grew 63%, from around $12 billion to over $20 billion in a single quarter, for the first time ever. More striking, Alphabet reported that customers have already signed contracts worth $460 billion in future cloud spending, roughly double what was committed three months earlier. Alphabet's CEO said that Google cannot yet build data centres fast enough to meet demand. Companies wanting to run AI workloads are signing contracts before the infrastructure exists, betting that Google will build fast enough to fulfil the commitment.

What could happen next?

Opportunity
Google Cloud's compute constraint creates an 18-24 month window for European cloud providers (OVHcloud, Deutsche Telekom) to capture enterprise AI workloads that cannot wait for Google capacity, particularly for EU-domiciled data.
Consequence
The $460 billion backlog commits Google Cloud to infrastructure delivery timelines that lock Alphabet into $35-plus billion quarterly capex through at least 2028, making headcount reductions the primary lever for operating margin management.

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: Oracle's 2000-2003 database infrastructure expansion saw enterprise backlog rise to $11 billion (equivalent to roughly $20 billion today) as companies committed to multi-year database licences before migrations were complete.

When the dot-com correction forced cost cuts across enterprise clients, 22% of the backlog was deferred or renegotiated. Alphabet's situation differs in that AI workload contracts are typically tied to specific compute allocations, not speculative licence agreements. But the backlog-conversion risk has the same structural shape.

Counter-parallel: Amazon Web Services grew from $5 billion in 2014 to $21 billion in 2017, a period during which its backlog expanded faster than revenue without a major renegotiation event. Google Cloud is tracking a faster growth curve. AWS's conversion rate held above 85% throughout that expansion phase, which Google Cloud bulls cite as the relevant precedent.

Alphabet's $35.7 billion Q1 capex is concentrated in two spending streams. Custom TPU (Tensor Processing Unit) silicon and networking infrastructure, estimated at 35-40% of the total per Evercore ISI analysis, and physical data centre construction across the US, Singapore, and Europe (Pichai confirmed European expansion on the call). EU data sovereignty requirements favour Google Cloud for European enterprise customers over AWS and Azure, both of which face EU Commission DMA scrutiny.

The compute constraint Pichai cited translates into a specific competitive dynamic: enterprises that cannot get immediate capacity from Google are signing longer-term contracts as a queue commitment. This explains the backlog doubling without a proportionate revenue jump. The revenue will follow, but the timing depends on construction lead times now running at 24-30 months for hyperscale data centre completion.

Consensus view: Wolfe Research analyst Alex Zukin and UBS analyst Karl Keirstead both highlighted the $460 billion backlog as the most important figure in Alphabet's Q1 report, arguing it represents durable, contracted future revenue that de-risks the capex programme relative to Amazon and Microsoft, where backlog figures are smaller. Keirstead estimated Google Cloud would overtake AWS in revenue growth rate for at least four consecutive quarters at current trajectories.

Counter-view: New Street Research (analyst Dan Ritzer) noted that large enterprise cloud backlogs have historically included a 15-25% 'renegotiation buffer' where customers exercise price-down or delay provisions as contract commencement approaches. At $460 billion, a 15% renegotiation would wipe $69 billion off the backlog. Alphabet does not disclose backlog conversion rates, making the 15-25% renegotiation risk unverifiable from public filings.

Key tension: Whether Pichai's compute constraint is a short-term supply bottleneck that resolves by Q4 2026 as new data centres come online, or a structural capacity gap that competitors (Microsoft, Amazon) fill while Google is building.

Sources:Alphabet / SEC

Mentions:Alphabet →Google →Sundar Pichai →Meta →Oracle →Microsoft →SK Hynix →Azure →TSMC →Google Cloud →Amazon →AWS →Samsung Electronics →

First Reported In

Update #8 · Beijing court bans AI sackings as Big Tech burns cash

Alphabet / SEC· 2 May 2026

Read original →

Different Perspectives

Salesforce, Synopsys and TD Bank Group

Salesforce, Synopsys and TD Bank Group each filed quarterly disclosures in late August booking restructuring charges, or none at all, without naming AI as a cause. Their silence matters because Challenger's tracker shows AI as a stated reason fell to fourth place in August even as the year-to-date AI-cut total still leads at 116,175.

Singapore, South Korea, Taiwan and Indonesia

Singapore launched its Skills and Workforce Development Agency on 16 September, giving citizens six months of free premium AI tools, while South Korea ring-fenced its AI tax windfall in a new Future Response Fund. Taiwan kept funding its AI build past NT$190bn and Indonesia rewired vocational training around AI literacy, betting state-built skills beat a market-led adjustment.

ver.di, CGT Fonction Publique and CCOO

Germany's ver.di banked a 3.3% pay rise on 1 September and opened talks on a Tarifvertrag Transformation covering dismissal bans and reskilling, while France's CGT rejected Paris's AI negotiating timetable the same week. Spain's CCOO went further on 21 September, proposing to tax companies by the jobs they generate rather than wait for the next bargaining round.

BIS General Manager and Federal Reserve governors

The BIS's General Manager said on 10 September that AI displacement remains limited, even as the BIS's own survey found nearly 80% of firms plan to automate roles. Two Federal Reserve governors made the same point in July, arguing the labour-market data does not yet show a mass-firing event.

Bank of Canada, ONS and ECB

The Bank of Canada found the job-finding gap between AI-exposed and unexposed occupations widened from 2.2 to 13.9 percentage points since 2015-19, while separations barely moved. That framing, a hiring freeze rather than a firing wave, is echoed by the ECB's finding that euro-area AI use hit 52% of workers in 2026, concentrated among the university-educated.

Office for National Statistics

Deferred its Transformed Labour Force Survey beyond November 2027 and disclosed a May 2026 telephone-collection failure. The ONS carries no AI-attribution layer at all, so Britain sits outside this month's cohort of measuring states by its own admission.