First-party TPU sales are expected to contribute approximately $192 billion in revenue in 2027–2028, while three AI catalysts could drive an Alphabet rerating

Overview

The report focuses on the potential for first-party sales of Alphabet's internally developed TPU systems and raises forecasts for TPU, Google Cloud, and overall company earnings based on new information about supply scale and pricing. Morgan Stanley believes TPU sales can directly increase earnings, but the key determinants of a further valuation uplift remain Gemini 4, efficient low-cost models, and the scaled deployment of AI products.

Core views

First, Morgan Stanley substantially raises its monetization assumptions for Alphabet's sales of TPU systems to first-party customers. Recent reports indicate that Alphabet has committed to supplying approximately 1 million TPUs, which the research team estimates represents approximately 1.3GW and a contract value of approximately $35 billion, implying revenue of approximately $27 billion per GW. This is above the prior assumption of approximately $20 billion per GW. The report also views Marvell's expanded custom-chip partnership as evidence supporting higher TPU pricing and average selling prices. At the same time, the report raises the associated gross margin assumption from 20% to 30% and expects Alphabet to sell 0.3GW, 3.2GW, and 4.2GW of TPU systems in the second half of 2026, 2027, and 2028, respectively. Driven jointly by volume and price, TPU-related Google Cloud revenue is expected to reach $84 billion in 2027 and $108 billion in 2028, 35% and 37% above previous forecasts, respectively. The two-year total is approximately $192 billion, close to the $200 billion cited in the title. Second, the higher TPU monetization assumptions significantly expand Google Cloud's revenue and profit scale. The report raises its 2027 cloud revenue growth forecast from 106% to 118% and its EBIT growth forecast from 105% to 137%. Overall Google Cloud revenue forecasts for 2027 and 2028 are raised by 6% and 7%, respectively, while EBIT forecasts are increased by 2% in both years. By 2028, cloud is expected to contribute 48% of Alphabet's companywide EBIT, or nearly half. However, because the incremental TPU sales are large and carry lower margins than the core cloud business, together with margin pressure from the SPCX compute agreement, the report modestly lowers reported cloud margin forecasts for 2027 and 2028 from 38% and 41% to 37% and 40%, respectively. Excluding the effects of TPU and SPCX, the core cloud business's incremental margin is still expected to remain at approximately 50% through 2027, while TPU's EBIT margin is estimated at approximately 30%. The report therefore does not conclude that cloud operating efficiency is deteriorating; rather, the rapid expansion of new, lower-margin revenue is changing the revenue mix. Third, the change in TPU assumptions flows through to Alphabet's consolidated financial statements, but the simultaneous increase in costs means the EPS upgrades are smaller than the revenue upgrades. Morgan Stanley raises its companywide revenue forecasts for 2027 and 2028 by 2% and 3%, respectively. To reflect higher cost of sales and align incremental margins for the non-TPU businesses, it raises its “Other CoR” forecasts for the same periods by 4% and 6%, ultimately increasing its 2027 and 2028 EPS forecasts by 1% and 2%, respectively. EPS in the detailed model is approximately $15 and $18.50, with the 2028 table value at $18.49; the risk-reward scenarios round these figures to approximately $15 and $18. The report also expects Alphabet's total revenue to rise from $402.836 billion in 2025 to $500.611 billion in 2026, $665.846 billion in 2027, and $813.676 billion in 2028, representing a 26% CAGR from 2025 to 2028. Over the same period, GAAP operating margin is expected to increase from 32.0% to 34.3%, 34.5%, and 35.0%. The research team will continue monitoring newly disclosed inventory data on the balance sheet to assess inventory accumulation and digestion after TPU sales commence. Fourth, the report believes TPU sales primarily provide EPS upside, while a valuation rerating still depends on three product catalysts. The most important is whether Gemini 4, expected to launch in late 2026 or early 2027, can return Google to or near the industry frontier, thereby restoring the valuation to levels reached after Gemini 3 launched in November 2025. The second catalyst is efficient, low-cost models, including Gemini Flash. As enterprises focus more closely on budgets when implementing AI, the performance-to-cost ratio of these models may affect Google's competitive position. The third is Google's full-scale rollout of generative AI products across Search, YouTube, and other user scenarios, demonstrating that AI innovation can deliver more durable Search and YouTube growth than the market currently expects rather than cannibalizing the core search business. Finally, Morgan Stanley maintains its $400 price target and Overweight rating. The price target is based on the average of the new 2027 and 2028 EPS estimates of approximately $15 and $18 to $18.50, applying a P/E multiple of approximately 24x. The report also states that this valuation represents approximately 1.6x PEG, a premium of approximately 35% to the peer median PEG. On a growth-adjusted basis, the $400 price target implies an approximately 20% valuation premium for Alphabet relative to peers, versus a current trading premium of approximately 5%, and offers approximately 15% upside from the $344.82 closing price. In the risk-reward analysis, the bull case is $460, based on a multiple of approximately 26x the average of 2027 and 2028 EPS of approximately $16 and $20, assuming AI innovation accelerates Search and Cloud growth and improves profitability. The base case is $400, based on a multiple of approximately 24x the average of EPS of approximately $15 and $18. The bear case is $225, based on a multiple of approximately 15x the average of EPS of approximately $14 and $16, assuming global advertising growth continues to slow, search advertising growth weakens, expense discipline fails to materialize, and AI products depress margins due to lower monetization rates and higher compute intensity.

Analysis framework

The report first derives sales revenue per GW using TPU supply quantities, estimated compute capacity, and contract value, and then develops TPU revenue forecasts by incorporating average rack selling prices, gross margins, and annual sales capacity. It subsequently incorporates TPU revenue and costs into the Google Cloud model to estimate changes in cloud revenue, EBIT, and margins, and then flows these effects through to Alphabet's consolidated revenue, costs, and EPS. Finally, it determines the price target using average 2027–2028 EPS, P/E, PEG, and a growth-adjusted premium to peers, and tests the valuation range under different product, growth, and margin assumptions through bull, base, and bear scenarios.

Methodology notes

• Industry/Sector Analysis FrameworkVolume-price decomposition

  Decomposition of TPU capacity, selling price per GW, and margin

  The report forecasts the capacity of TPU systems sold in each year and calculates revenue using an assumption of approximately $27 billion per GW, then adds a 30% gross margin assumption to assess the profit contribution. This decomposition shows that the forecast upgrades result from both volume expansion and improved unit monetization.

• Company Fundamentals and Financial FrameworkOperating/Financial Leverage Analysis

  Cloud incremental margin and revenue mix analysis

  The report distinguishes between the core cloud business's approximately 50% incremental margin, TPU's approximately 30% EBIT margin, and the impact of the SPCX agreement to explain why cloud revenue and EBIT grow rapidly while the reported cloud margin still declines modestly.

• Valuation MethodPE/PEG valuation

  P/E and PEG valuation using average 2027–2028 EPS

  The $400 price target is primarily derived by applying an approximately 24x P/E multiple to average 2027 and 2028 EPS, with valuation reasonableness checked using an approximately 1.6x PEG and a growth-adjusted premium relative to peers.

• Valuation Method

  Bull, base, and bear risk-reward scenarios

  The report sets different assumptions for product innovation, Search and Cloud growth, expense discipline, and AI compute costs, producing scenario values of $460, $400, and $225, respectively, to illustrate the impact of changes in key variables on earnings and valuation.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Alphabet Inc. (GOOGL.O, GOOGL.US)

  First-party TPU sales raise Google Cloud revenue and EBIT forecasts, while the scaling of Gemini and AI products is the report's principal driver of a valuation rerating.

  Strengths

  Higher pricing and sales capacity for internally developed TPUs; approximately 50% incremental margin in the core cloud business; Search, YouTube, and Cloud provide a multi-scenario foundation for AI productization; expense discipline can generate operating leverage.

  Weaknesses

  TPU sales carry lower margins than the core cloud business, while the SPCX compute agreement also pressures cloud margins, making the EPS uplift from incremental revenue relatively limited.

  Comparison

  The $400 price target implies an approximately 20% growth-adjusted premium to peers, versus Alphabet's current approximately 5%; the report gives a peer median 2028 P/E of 18x.

  Risks

  Slowing advertising demand, decelerating Search growth, weaker-than-expected expense discipline, and lower AI product monetization rates combined with higher compute intensity could all depress earnings and valuation.

Key data

• First-party TPU sales unit revenue assumptionapproximately $27bn/GWpreviously approximately $20bn/GW
• TPU sales gross margin assumption30%previously 20%
• TPU system sales capacity2H26: 0.3GW; 2027: 3.2GW; 2028: 4.2GWused to estimate annual TPU system sales revenue
• TPU-related Google Cloud revenue2027: $84bn; 2028: $108bn35% and 37% above previous forecasts, respectively, totaling $192bn over two years
• 2027 Google Cloud growth forecastRevenue 118%; EBIT 137%previously 106% and 105%, respectively
• Google Cloud forecast revisions2027/2028 revenue: +6%/+7%; EBIT: +2%/+2%primarily driven by higher TPU pricing and margin assumptions
• Reported Google Cloud margin2027: 37%; 2028: 40%previously 38% and 41%, respectively
• Core cloud incremental marginapproximately 50%excluding the effects of TPU and the SPCX agreement, expected to be maintained through 2027
• Cloud EBIT contribution in 202848%, approximately 50%share of Alphabet's companywide EBIT
• Alphabet revenue forecast revisions2027: +2%; 2028: +3%driven by higher TPU revenue assumptions
• Alphabet EPS forecast revisions2027: +1%; 2028: +2%net impact on earnings per share after the simultaneous increase in cost forecasts
• New EPS forecasts2027: approximately $15; 2028: approximately $18.50 (table value $18.49)rounded to approximately $15 and approximately $18 in the risk-reward scenarios
• Alphabet total revenue forecasts2025: $402.836bn; 2026E: $500.611bn; 2027E: $665.846bn; 2028E: $813.676bn26% CAGR from 2025 to 2028
• GAAP operating margin forecasts2025: 32.0%; 2026E: 34.3%; 2027E: 34.5%; 2028E: 35.0%the report expects operating margins to improve overall
• Price target and valuation$400; approximately 24x P/E; approximately 1.6x PEGbased on average 2027–2028 EPS, with approximately 15% upside
• Peer valuation premiumprice target implies approximately 20%; current approximately 5%growth-adjusted premium relative to peers
• Risk-reward scenariosBull $460; Base $400; Bear $225corresponding to approximately 26x, 24x, and 15x average 2027–2028 scenario EPS valuations

Impact & implications

The report believes that TPU has evolved from internal compute infrastructure into a substantial source of first-party system sales revenue, significantly expanding Google Cloud's revenue scale and contribution to companywide profit. However, because TPU margins are lower than those of the core cloud business, the revenue upgrades will not produce proportional increases in cloud margins or Alphabet EPS. Whether Alphabet's growth-adjusted premium to peers can rise from the current approximately 5% to the approximately 20% implied by the price target will depend primarily on whether Google can demonstrate more durable long-term growth through Gemini 4, efficient models, and cross-product AI commercialization.

Risks

• If global advertising growth slows further, search advertising revenue and margins could come under pressure.
• Alphabet has relatively high exposure to small and medium-sized businesses and travel advertisers, and a recession could depress advertising revenue.
• If expense discipline fails to materialize, margin expansion, adjusted EBITDA, EPS, and free cash flow could fall below expectations.
• If new AI products have lower monetization rates and higher-than-expected compute intensity, they could create greater margin pressure.
• Margins on TPU sales and the SPCX compute agreement are lower than those of the core cloud business, potentially pressuring reported cloud margins.
• If AI products cannibalize the core search business rather than generate incremental revenue, the growth and valuation expansion in the bull case may not materialize.

What to watch

• Monitor whether Gemini 4, expected to launch in late 2026 or early 2027, can return Google to or near the frontier-model level.
• Monitor the adoption and competitive performance of efficient, low-cost models such as Gemini Flash as enterprises place greater emphasis on AI budgets.
• Monitor the full-scale deployment of generative AI products across Search, YouTube, and other user scenarios, as well as their contribution to sustained growth and monetization.
• Monitor new balance-sheet inventory disclosures to assess inventory accumulation and digestion after TPU sales commence.
• Monitor whether new Google Cloud partnerships, backlog, and TPU sales capacity can support the 2027–2028 revenue forecasts.
• Monitor search advertising's share of global advertising budgets, mobile search's share of online advertising budgets, and YouTube's long-term monetization.
• Monitor whether expense growth can continue to slow and generate the expected operating leverage.