Summary
- Cloud Next as catalyst: Google Cloud Next 2026 (April 22-24) delivered the most comprehensive AI infrastructure announcement cycle in the company's history — TPU 8 silicon in two variants, Gemini 3.1 model family, a full agentic AI platform, and a new chip partnership with Marvell — establishing Google as the vertically integrated AI infrastructure leader ahead of Q1 earnings on April 29.
- $243 billion backlog: Google Cloud's remaining performance obligations have exploded to $243 billion, providing multi-year revenue visibility that dwarfs the segment's current $58 billion annualized run rate and signals enterprise AI adoption is accelerating, not plateauing.
- TPU 8 dual-chip strategy: The simultaneous launch of TPU 8t for training (scaling to 9,600-chip superpods with 2PB shared memory) and TPU 8i for inference (80% performance improvement over prior generation) creates a complete custom silicon stack that competes directly with NVIDIA's GPU dominance.
- Pre-earnings setup: With consensus expecting $107 billion in Q1 revenue (+19% YoY) and Cloud revenue growth of 57.5%, Alphabet trades at approximately 29 times forward earnings — a reasonable multiple for a company growing at this pace with multiple re-rating catalysts ahead.
Cloud Next 2026: The Agentic Era Begins
Google Cloud Next has evolved from a product conference into a strategic declaration. The 2026 edition, held April 22-24, was arguably the most consequential in the event's history — not because of any single announcement, but because of the coherence of the overall vision. Alphabet is no longer simply offering cloud infrastructure; it is building the complete operating system for enterprise AI.
The centerpiece of the enterprise strategy is the Agent Platform, a suite of tools designed to make AI agents a production-ready enterprise capability rather than a research curiosity. Agent Studio provides a visual development environment for building, testing, and deploying AI agents without requiring deep machine learning expertise. Agent Registry offers a centralized catalog for discovering and managing agents across an organization. Agent Gateway handles authentication, rate limiting, and observability for agent deployments.
But the most strategically significant component is Agent-to-Agent Orchestration, which Google brands as A2A. This protocol enables AI agents built on different platforms — whether Google's Gemini, Anthropic's Claude, or open-source models — to communicate and collaborate on complex tasks. The implications are profound. By positioning A2A as an open standard, Google is executing the same playbook that made Kubernetes the default container orchestration platform: define the standard, open-source it, and then monetize the managed service that runs it at scale.
The enterprise AI market is entering a phase where individual model quality matters less than the infrastructure that connects, orchestrates, and governs AI applications. Google's Agent Platform positions Cloud as that infrastructure layer. If A2A gains adoption — and early partnerships with Salesforce, SAP, and ServiceNow suggest it will — Google Cloud could capture platform-level economics similar to what Amazon Web Services achieved with compute infrastructure in the 2010s.
This is not speculative. Google Cloud's revenue grew 57% year-over-year in the most recently reported quarter, and the $243 billion backlog suggests that enterprises are not merely experimenting with Google's AI platform — they are signing multi-year commitments.
TPU 8: Silicon Supremacy
The launch of TPU 8 in two distinct variants — TPU 8t optimized for training and TPU 8i optimized for inference — represents Alphabet's most aggressive move yet in the custom silicon arms race.
TPU 8t is designed for the most demanding AI training workloads. A single superpod can scale to 9,600 TPU 8t chips interconnected with 2 petabytes of shared memory, creating a training cluster with enough capacity to train frontier models that would otherwise require tens of thousands of NVIDIA GPUs. The shared memory architecture is the key innovation. Traditional GPU clusters rely on network-based communication between nodes, creating bottlenecks as model sizes grow. TPU 8t's 2PB shared memory per superpod eliminates much of this overhead, enabling more efficient scaling for the massive models that define the current AI frontier.
TPU 8i, by contrast, targets the inference workloads that generate the majority of AI-related compute demand in production deployments. Google claims an 80% performance improvement over the previous generation, a leap that directly addresses the cost challenge facing every enterprise deploying AI at scale. Inference costs typically exceed training costs by a factor of three to ten over a model's lifetime, making inference efficiency the primary determinant of AI deployment economics.
The dual-chip strategy has competitive implications that extend beyond Google's own cloud. By offering purpose-built silicon for both training and inference, Google is creating a vertically integrated alternative to NVIDIA's general-purpose GPU approach. This matters because NVIDIA's dominance has driven GPU prices to levels that are becoming prohibitive for many enterprises. Google can offer its TPU-based cloud services at lower price points while maintaining margins because it controls the silicon design, the software stack, and the cloud infrastructure.
The competitive landscape is shifting rapidly. Amazon has its Trainium and Inferentia chips, and Microsoft recently deployed its Maia accelerator. But neither has achieved the scale or the performance claims that Google is making with TPU 8. If these claims hold up under independent benchmarking — and Google's track record with prior TPU generations suggests they will — the TPU 8 family could meaningfully shift enterprise AI workloads toward Google Cloud.
The $243 Billion Backlog
Perhaps the most underappreciated number in Alphabet's financial profile is Google Cloud's $243 billion backlog of remaining performance obligations. This figure, which represents contracted but not yet recognized revenue, has grown at an extraordinary pace — roughly doubling over the past 18 months.
To put this in perspective, Google Cloud generated approximately $14.5 billion in revenue in Q4 2025, implying an annualized run rate of roughly $58 billion. The $243 billion backlog therefore represents more than four years of revenue at current run rates. While not all of this backlog will convert to revenue in the near term — contract durations range from one to ten years — the magnitude provides a level of revenue visibility that is rare among hyperscale cloud providers.
The backlog comparison with competitors is instructive. Microsoft's Azure does not disclose a directly comparable figure, though Microsoft's total remaining performance obligations across all commercial cloud services were approximately $298 billion as of Q2 FY2026. Amazon Web Services reported a backlog of approximately $189 billion. Google Cloud's $243 billion figure, relative to its much smaller revenue base, suggests a higher growth trajectory than either competitor.
What is driving this backlog accumulation? Three factors stand out. First, enterprises are committing to multi-year Google Cloud contracts to secure access to TPU capacity and Gemini models, often before they have fully defined their AI deployment strategies. Second, Google's consumption-based pricing model for AI services creates a natural backlog accumulation as enterprises ramp usage. Third, the Agent Platform announcements at Cloud Next 2026 are likely to accelerate contract signings as enterprises seek to build on the A2A orchestration framework.
The investment implications are clear. Even if Google Cloud's growth rate moderates from the current 57% level — and the law of large numbers makes some deceleration inevitable — the backlog provides a revenue floor that significantly reduces downside risk for the Cloud segment. Our model assumes Cloud revenue growth decelerates to approximately 45% by Q4 2026, which still implies roughly $72 billion in annual Cloud revenue, representing nearly 15% of Alphabet's total.
Alphabet's AI strategy extends far beyond Cloud. The Gemini model family, now in its 3.1 generation, is reshaping the company's core Search business in ways that the market has not fully priced.
Cloud Next 2026 introduced three new models: Gemini 3.1 Pro, the flagship model optimized for complex reasoning and extended context windows; Gemini 3.1 Flash, a lighter-weight model designed for high-throughput inference at lower cost; and Veo 3.1 Lite for video generation and Lyria 3 Pro for music generation. The breadth of the model portfolio reflects Google's strategy of competing across the full spectrum of AI applications rather than concentrating on a single frontier model.
But the most commercially significant Gemini application is AI Overviews in Search. Google now deploys AI-generated overview responses on more than 40% of search queries, up from less than 10% in early 2025. This is not merely a user experience enhancement — it is a fundamental shift in Search monetization.
AI Overviews enable Google to monetize queries that previously generated little or no revenue. Consider a query like "best enterprise CRM for a 500-person company." In traditional Search, this query returns a mix of organic results and advertisements, with a click-through rate on ads of perhaps 3% to 5%. With AI Overviews, Google can embed contextual recommendations within the AI-generated response, creating a new advertising surface that is both more useful to the user and more valuable to the advertiser. Early data suggests that AI Overview ad engagement rates are 1.5 to 2 times higher than traditional search ad click-through rates.
The market's fear that AI would destroy Search economics is increasingly looking misplaced. Search revenue is expected to grow approximately 16.5% year-over-year in Q1 2026, an acceleration from prior quarters. If AI Overviews continue expanding to cover 60% or more of queries by year-end — a trajectory that management has indicated is likely — Search could sustain mid-teens growth rates through 2027, confounding the bears who predicted ChatGPT-driven disruption would erode Google's search monopoly.
Search accounted for 48.2% of Alphabet's fiscal year 2025 revenue of $410.2 billion. The resilience and acceleration of this segment is the foundation of the investment case because it funds the company's massive capital expenditure program while generating substantial free cash flow.
$175-185 Billion Capex: The Infrastructure Arms Race
CEO Sundar Pichai reaffirmed at Cloud Next 2026 that Alphabet will spend between $175 billion and $185 billion in capital expenditures during fiscal year 2026. To contextualize this figure: it exceeds the GDP of countries like Ukraine, Hungary, or Kuwait. It is roughly equivalent to what the entire U.S. federal government spent on education in 2025.
Approximately half of this capex — $87 to $93 billion — is directed toward Google Cloud infrastructure, including data center construction, TPU fabrication, and networking equipment. The remainder supports Google Services infrastructure (Search, YouTube, Android) and Other Bets (primarily Waymo's autonomous vehicle fleet).
The Marvell partnership announced at Cloud Next adds a new dimension to Alphabet's infrastructure strategy. Google and Marvell are co-developing a custom media processing unit (MPU) and an inference-optimized TPU variant, making Marvell the third chip partner after Broadcom and MediaTek. This diversified chip supply strategy reduces Google's dependence on any single silicon partner while enabling specialized chip designs for different workload types.
The capex debate among investors centers on return on investment. Bears argue that $175 to $185 billion in annual spending represents a leap of faith — that Alphabet is building infrastructure capacity for AI demand that may not materialize at the projected scale. Bulls counter that the $243 billion Cloud backlog already validates the demand thesis and that underinvesting in AI infrastructure would be the greater strategic risk.
Our analysis favors the bull interpretation. The hyperscale cloud market is exhibiting characteristics of a natural oligopoly, where scale advantages create durable competitive moats. Alphabet, Amazon, and Microsoft are the only companies with the financial resources, technical expertise, and customer relationships to compete at this level. Each dollar of capex invested today creates infrastructure that will generate revenue for 10 to 15 years. The question is not whether the capex will generate returns — it is whether the returns will justify the magnitude of the investment. The Cloud backlog and growth trajectory suggest they will.
Valuation: Fairly Priced or Still Cheap?
At $535.82 per share, Alphabet trades at approximately 29.2 times forward earnings estimates, a modest premium to its five-year average of roughly 25 times but a discount to the growth-adjusted multiples commanded by peers like Microsoft (33 times forward) and Amazon (42 times forward). For a company growing revenue at 19% annually with improving margins and a $243 billion Cloud backlog, we believe this valuation underprices the optionality embedded in Alphabet's business.
Our three-scenario valuation model produces the following outcomes.
Bull Case ($720, 25% probability): Google Cloud revenue growth accelerates past 60% as A2A adoption drives enterprise platform lock-in. AI Overviews expand to 70% of search queries by year-end, driving Search revenue growth above 20%. Waymo begins contributing meaningful revenue as robotaxi services expand beyond current markets. YouTube advertising benefits from connected TV penetration. Applying 35 times forward earnings on estimated calendar year 2027 EPS of approximately $20.50 yields $720 per share, implying a market capitalization of roughly $5.5 trillion.
Base Case ($620, 45% probability): Revenue grows 19% in fiscal year 2026 and 16% in fiscal year 2027. Cloud maintains 50% or better growth through 2026, decelerating to 40% in 2027. Search grows at mid-teens rates as AI Overviews expand. Margins expand modestly as Cloud reaches operating profitability at scale. Applying 31 times forward earnings on estimated calendar year 2027 EPS of approximately $20 yields $620, implying a market capitalization of roughly $4.7 trillion. This is our primary price target.
Bear Case ($450, 30% probability): Cloud growth decelerates to 35% as enterprise AI adoption plateaus and competition from Azure intensifies. The $175 to $185 billion capex program burns free cash flow, raising investor concerns about capital allocation discipline. DOJ antitrust remedies force structural changes to Chrome distribution or Search defaults, reducing Search revenue by 5% to 8%. Applying 24 times forward earnings on estimated calendar year 2027 EPS of approximately $18.75 yields $450, implying a market capitalization of roughly $3.4 trillion.
The probability-weighted target across all scenarios is approximately $600 per share, or 12% above the current price. We set our price target at $620, above the weighted figure, reflecting what we believe is underpriced AI infrastructure optionality and the unrecognized value of Waymo's autonomous mobility platform, which represents a long-term total addressable market measured in trillions.
Key Risks
Three risks require particular attention from investors.
First, the DOJ antitrust case represents the most significant regulatory overhang on Alphabet's stock. The Department of Justice has proposed remedies that could include forcing Google to divest Chrome, modifying Search default agreements with Apple and other device manufacturers, or imposing structural separation requirements on the advertising technology business. While the most draconian outcomes remain unlikely — courts have historically favored behavioral remedies over structural breakups — even modest changes to Search distribution could reduce revenue by $10 billion to $20 billion annually. The case is expected to reach a resolution phase in late 2026 or early 2027, creating an extended period of uncertainty.
Second, the return on $175 to $185 billion in annual capex is uncertain. Alphabet is betting that enterprise AI demand will sustain Cloud growth at 40% or better for multiple years. If adoption curves flatten — because enterprises struggle to realize ROI from AI deployments, because open-source models reduce demand for cloud-hosted AI services, or because a recession curtails IT budgets — the capex program could prove excessive, pressuring free cash flow and forcing Alphabet to choose between maintaining investments and returning capital to shareholders.
Third, Cloud competition from Microsoft Azure and Amazon Web Services remains intense. Azure's integration with OpenAI models and Microsoft's enterprise software ecosystem gives it a structural advantage in enterprise accounts. AWS's first-mover advantage and massive installed base create switching costs that make it difficult for Google to win incumbent workloads. Google Cloud's 14.6% share of Alphabet's total revenue, while growing rapidly, still represents the third-place position in the hyperscale cloud market. Sustaining 50%-plus growth rates requires winning a disproportionate share of new workloads — a task that becomes progressively harder as the Cloud segment scales.
Conclusion
We rate Alphabet a Buy with a $620 price target, representing approximately 16% upside from the current price of $535.82. Cloud Next 2026 validates the AI infrastructure thesis that underpins our investment case: Alphabet is the only company in the world that controls frontier AI models (Gemini), custom training and inference silicon (TPU 8), a hyperscale cloud platform (Google Cloud), the dominant search engine (Search + AI Overviews), and an autonomous mobility platform (Waymo). This vertical integration is unique and, we believe, undervalued at 29 times forward earnings.
The April 29 earnings report is the next catalyst. Consensus expectations for $107 billion in revenue and 57.5% Cloud growth are achievable based on the backlog trajectory, but the more important signal will be management's commentary on Cloud bookings trends post-Cloud Next and the early revenue contribution from TPU 8 deployments. A beat-and-raise scenario — which our 360-degree analysis assigns a combined 55% probability — could push the stock toward $580 to $600 in the near term.
For readers tracking the AI infrastructure supply chain, Marvell's AI custom silicon story directly intersects with Alphabet's chip diversification strategy, while IBM's enterprise AI pivot illustrates how legacy technology companies are attempting to compete for the same enterprise AI budgets that Google Cloud is capturing. See also our coverage of AMD's AI chip strategy for the competitive silicon landscape.
Frequently Asked Questions
What were the biggest announcements at Google Cloud Next 2026?
Google Cloud Next 2026 (April 22-24) featured five major announcement categories. First, TPU 8 in two variants: TPU 8t for training, capable of scaling to 9,600-chip superpods with 2 petabytes of shared memory, and TPU 8i for inference with 80% better performance than the prior generation. Second, the Gemini 3.1 model family including Gemini 3.1 Pro and Flash, plus Veo 3.1 Lite for video and Lyria 3 Pro for music. Third, the Agent Platform encompassing Agent Studio, Agent-to-Agent Orchestration (A2A), Agent Registry, and Agent Gateway for enterprise AI agent deployment. Fourth, a chip partnership with Marvell Technology to co-develop a custom media processing unit and inference-optimized TPU variant, making Marvell the third chip partner after Broadcom and MediaTek. Fifth, CEO Sundar Pichai reaffirmed the $175 to $185 billion capex plan for 2026, with approximately half directed toward Cloud infrastructure.
How does Alphabet's $243 billion Cloud backlog compare to competitors?
Google Cloud's $243 billion backlog of remaining performance obligations is remarkable relative to the segment's size. With an annualized revenue run rate of roughly $58 billion, the backlog represents more than four years of revenue at current rates, suggesting a significantly higher growth trajectory than peers. For comparison, Amazon Web Services reported approximately $189 billion in backlog, while Microsoft's total commercial cloud remaining performance obligations were roughly $298 billion — but Microsoft's figure covers a much larger revenue base including Office 365, Dynamics, and Azure combined. On a backlog-to-revenue ratio basis, Google Cloud's metric is the highest among the three major hyperscalers, indicating the strongest relative growth pipeline.
What is the risk from the DOJ antitrust case against Google?
The Department of Justice antitrust case is the most significant regulatory risk facing Alphabet. Proposed remedies include potentially forcing divestiture of the Chrome browser, modifying or eliminating Search default agreements with Apple and other device manufacturers, and imposing structural separation on Google's advertising technology business. If the most aggressive remedies are implemented, they could reduce Search revenue by an estimated $10 billion to $20 billion annually. However, courts have historically favored behavioral remedies over structural breakups in technology antitrust cases, and Alphabet has significant legal resources to contest unfavorable outcomes. The resolution phase is expected in late 2026 or early 2027. We assign a roughly 20% probability to outcomes that would materially impair Alphabet's Search economics.
Is Alphabet's $175-185 billion capex sustainable?
Alphabet's planned capital expenditure of $175 to $185 billion for fiscal year 2026 is the largest in the company's history and among the largest ever committed by a single corporation. The sustainability of this spending depends on whether Google Cloud revenue growth justifies the investment. With Cloud revenue growing at 57% year-over-year and a $243 billion backlog providing multi-year visibility, current demand trends support the investment thesis. Alphabet generated approximately $95 billion in free cash flow in fiscal year 2025, and even with the elevated capex, is expected to generate $55 to $65 billion in free cash flow in fiscal year 2026. The company also maintains approximately $100 billion in cash and marketable securities. The risk is not near-term financial viability but rather long-term return on investment — if AI demand curves flatten before the capex generates adequate returns, Alphabet could face pressure to curtail spending, which would signal a deceleration narrative that could compress valuation multiples.
Should I buy GOOGL stock before earnings on April 29?
Our analysis rates Alphabet a Buy with a $620 price target, representing 16% upside from the current price of $535.82. The pre-earnings setup is constructive: Cloud Next 2026 provided multiple positive catalysts, consensus estimates appear achievable based on backlog trends, and the stock trades at a reasonable 29 times forward earnings for its growth profile. Our 360-degree analysis assigns a combined 55% probability to beat or significant beat scenarios for Q1 earnings. However, investors should be aware of risks including potential disappointment on Cloud operating margins due to elevated infrastructure spending, any negative signals from the DOJ antitrust resolution timeline, and the possibility that the market's earnings expectations have risen following Cloud Next announcements. A prudent approach would be to establish a partial position before earnings with a plan to add on any post-earnings weakness, targeting a full position size at or below the current level.
Disclaimer
This article is for informational purposes only and does not constitute investment advice, a recommendation, or a solicitation to buy or sell any securities. The analysis, opinions, and price targets expressed herein are those of the author and Edgen.tech and do not represent the views of Alphabet Inc., Google LLC, or any affiliated entity. All financial data is sourced from public filings and proprietary Edgen 360° Reports and is believed to be accurate as of the publication date but is not guaranteed. Past performance is not indicative of future results. Investors should conduct their own due diligence and consult with a qualified financial advisor before making investment decisions. Edgen.tech and its contributors may hold positions in the securities discussed.
