{ "generatedAt": "2026-10-22T16:04:11Z", "url": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/", "disclaimer": "Analysis and opinion, not investment advice.", "id": "aws-custom-silicon-margin-proof", "number": 60, "question": "Will Amazon disclose by 2027-12-31 that its own AI chips lead a named AWS AI workload, or that AWS operating margin exceeds 38%?", "dek": "AWS's Trainium/Inferentia price-performance advantage drives enough inference migration off merchant GPUs to show up as a disclosed workload share or an expanded AWS operating margin.", "family": "T02", "subthesis": "T02.04", "sector": "AI.SEMI", "subsectors": [ "AI.SEMI.ASIC", "AI.SEMI.GPU", "AI.CLOUD.HYPER", "AI.NET.SWITCH" ], "relatedParts": [], "approvedAt": "2026-10-22T05:42:55Z", "sectorName": "Semiconductors & Compute Silicon", "api": "https://deepstack.ltd/api/v1/theses/aws-custom-silicon-margin-proof.json", "sides": [ { "id": "yes", "label": "Custom silicon takes share from merchant GPUs", "summary": "Custom chips claimed at 30-40% better price-performance plus near-universal Graviton adoption create switching costs and lower unit compute cost for captive AWS inference. AWS reduces per-unit compute cost and expands operating margin while publicly touting internal silicon share, and buyers increasingly compete on lowest cost per token as inference workloads get hard-wired to custom accelerators.", "voices": [ { "name": "Invest in Assets", "entity": "https://deepstack.ltd/people/invest-in-assets/", "claim": "Custom silicon adoption as switching cost", "date": "2026-08-11", "url": "https://www.investinassets.net/p/why-im-bullish-on-amazon" }, { "name": "Michael Spencer", "entity": "https://deepstack.ltd/people/michael-spencer/", "claim": "Amazon's custom chips achieve 30-40% better price-performance than competing solutions like NVIDIA GPUs. This creates a flywheel where lower-cost inference attracts customers, increases AWS usage, and improves unit economics. At scale, Trainium…", "date": "2026-04-14", "url": "https://www.ai-supremacy.com/p/what-amazons-shareholder-letter-says-about-ai-future-andy-jassy-2026" }, { "name": "Market Sentiment", "entity": "https://deepstack.ltd/people/market-sentiment/", "claim": "As buyers stop paying for frontier models on every task, competition turns to lowest cost per token; inference is repetitive enough to hard-wire, and at hyperscaler volume the $500M+ design cost is amortized, so the roughly half of inference cost…", "date": "2026-07-18", "url": "https://www.marketsentiment.co/p/asic" } ] }, { "id": "no", "label": "Nvidia's platform lock-in keeps custom silicon from showing up in disclosures", "summary": "Nvidia's investments across component suppliers keep downstream AI infrastructure dependent on Nvidia rails regardless of customer preference for custom silicon, and procurement teams favor fungible, general-purpose capacity over architectural purity. It would show first if AWS continues to describe Nvidia GPUs as the primary platform for its AI capacity through 2027, with no disclosed workload-share claim or margin threshold crossed.", "voices": [ { "name": "Gaetano", "entity": "https://deepstack.ltd/people/gaetano/", "claim": "By investing in and integrating component suppliers like Marvell across custom silicon, scale-up networking, and photonic fabric, Nvidia ensures downstream AI infrastructure remains dependent on Nvidia rails regardless of customer preferences for…", "date": "2026-03-31", "url": "https://gaetanoresearch.substack.com/p/marvellnvdapoet" }, { "name": "Dylan Patel", "entity": "https://deepstack.ltd/people/dylan-patel/", "claim": "By combining purpose-built hardware accelerators with flexible software compilation and task scheduling frameworks, a vendor can achieve the performance benefits of custom solutions while preserving the programmability and portability advantages of…", "date": "2022-04-06", "url": "https://newsletter.semianalysis.com/p/how-nvidias-empire-could-be-eroded" }, { "name": "Bryan Shan", "entity": null, "claim": "Fungibility beats architectural purity in procurement", "date": "2026-08-10", "url": "https://newsletter.semianalysis.com/p/ultra-high-interactivity-on-nvidia" } ] } ], "tests": [ { "id": "th060-t01", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "November 18, 2026", "whenLabel": "Nov 18, 2026", "date": "2026-11-18", "precision": "day", "event": "NVIDIA, long-term debt growth from a year earlier, quarter ended October 25, 2026: whether it comes in at or above 230%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "open", "callExpects": null, "callExpectsBy": null, "callExpectsAt": null, "callProbability": null, "callProbabilityAt": null, "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[0].date", "url": null }, "resolver": { "kind": "earnings", "symbol": "NVDA", "rule": "results reported => read the print", "question": "", "source": null, "ambiguity": "", "specState": "pending" }, "url": "https://deepstack.ltd/docket/#test-th060-t01", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 33, "at": "2026-10-22T06:19:12Z" }, { "model": "deepseek-reasoner", "p": 96, "at": "2026-10-22T06:19:32Z" } ], "readers": null }, { "id": "th060-t02", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "December 1, 2026", "whenLabel": "Dec 1, 2026", "date": "2026-12-01", "precision": "day", "event": "Marvell, revenue growth from a year earlier, quarter ended October 31, 2026: whether it comes in at or above 44%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "pre", "callExpects": "yes", "callExpectsBy": "Desk Editor (level A)", "callExpectsAt": "2026-10-22T12:53:49Z", "callProbability": 76, "callProbabilityAt": "2026-10-22T12:53:49Z", "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[1].date", "url": null }, "resolver": { "kind": "manual", "rule": "Compare the revenue that Marvell reports for the quarter ended October 31, 2026 with the quarter ended October 31, 2025, in its 10-Q or 10-K filing on SEC EDGAR. The answer is yes if growth is at or above 44.0%; no if it is below 44.0%.", "question": "Did the revenue of Marvell for the quarter ended October 31, 2026 grow at or above 44.0% from the quarter ended October 31, 2025?", "source": { "label": "SEC EDGAR, XBRL company facts of Marvell (Forms 10-Q and 10-K)", "url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0001835632.json" }, "ambiguity": "If Marvell has not filed by December 17, 2026, or a filing restates a period the rule reads, the test is inconclusive, and the Result says what was read and what was missing.", "specState": "approved", "specVersion": 1, "specUpdatedOn": "2026-10-22", "specHistory": [ { "version": 1, "date": "2026-10-22", "reason": "First specification.", "by": "Desk Editor (level A)" } ], "callExpects": "yes", "format": "ladder", "group": "th060-mrvl-revenue-yoy", "threshold": 44, "direction": "above", "unit": "%", "think": { "baseRate": "Marvell's revenue growth from a year earlier was 36.5% for the quarter ended August 1, 2026, by its filing on SEC EDGAR.", "outsideView": "The year-earlier quarter is not on the same concept in the filings, so the rungs sit on the last print alone.", "wouldChange": "The results-calendar consensus for the quarter ended October 31, 2026 implies 52.0%, and the middle rung is centred there." } }, "url": "https://deepstack.ltd/docket/#test-th060-t02", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 60, "at": "2026-10-22T06:19:36Z" }, { "model": "deepseek-reasoner", "p": 60, "at": "2026-10-22T06:20:05Z" } ], "readers": null }, { "id": "th060-t03", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "December 1, 2026", "whenLabel": "Dec 1, 2026", "date": "2026-12-01", "precision": "day", "event": "Marvell, revenue growth from a year earlier, quarter ended October 31, 2026: whether it comes in at or above 52%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "pre", "callExpects": "either", "callExpectsBy": "Desk Editor (level A)", "callExpectsAt": "2026-10-22T12:53:49Z", "callProbability": 50, "callProbabilityAt": "2026-10-22T12:53:49Z", "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[2].date", "url": null }, "resolver": { "kind": "manual", "rule": "Compare the revenue that Marvell reports for the quarter ended October 31, 2026 with the quarter ended October 31, 2025, in its 10-Q or 10-K filing on SEC EDGAR. The answer is yes if growth is at or above 52.0%; no if it is below 52.0%.", "question": "Did the revenue of Marvell for the quarter ended October 31, 2026 grow at or above 52.0% from the quarter ended October 31, 2025?", "source": { "label": "SEC EDGAR, XBRL company facts of Marvell (Forms 10-Q and 10-K)", "url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0001835632.json" }, "ambiguity": "If Marvell has not filed by December 17, 2026, or a filing restates a period the rule reads, the test is inconclusive, and the Result says what was read and what was missing.", "specState": "approved", "specVersion": 1, "specUpdatedOn": "2026-10-22", "specHistory": [ { "version": 1, "date": "2026-10-22", "reason": "First specification.", "by": "Desk Editor (level A)" } ], "callExpects": "either", "format": "ladder", "group": "th060-mrvl-revenue-yoy", "threshold": 52, "direction": "above", "unit": "%", "think": { "baseRate": "Marvell's revenue growth from a year earlier was 36.5% for the quarter ended August 1, 2026, by its filing on SEC EDGAR.", "outsideView": "The year-earlier quarter is not on the same concept in the filings, so the rungs sit on the last print alone.", "wouldChange": "The results-calendar consensus for the quarter ended October 31, 2026 implies 52.0%, and the middle rung is centred there." } }, "url": "https://deepstack.ltd/docket/#test-th060-t03", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 22, "at": "2026-10-22T06:20:10Z" }, { "model": "deepseek-reasoner", "p": 30, "at": "2026-10-22T06:20:40Z" } ], "readers": null }, { "id": "th060-t04", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "December 1, 2026", "whenLabel": "Dec 1, 2026", "date": "2026-12-01", "precision": "day", "event": "Marvell, revenue growth from a year earlier, quarter ended October 31, 2026: whether it comes in at or above 60%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "pre", "callExpects": "no", "callExpectsBy": "Desk Editor (level A)", "callExpectsAt": "2026-10-22T12:53:49Z", "callProbability": 25, "callProbabilityAt": "2026-10-22T12:53:49Z", "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[3].date", "url": null }, "resolver": { "kind": "manual", "rule": "Compare the revenue that Marvell reports for the quarter ended October 31, 2026 with the quarter ended October 31, 2025, in its 10-Q or 10-K filing on SEC EDGAR. The answer is yes if growth is at or above 60.0%; no if it is below 60.0%.", "question": "Did the revenue of Marvell for the quarter ended October 31, 2026 grow at or above 60.0% from the quarter ended October 31, 2025?", "source": { "label": "SEC EDGAR, XBRL company facts of Marvell (Forms 10-Q and 10-K)", "url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0001835632.json" }, "ambiguity": "If Marvell has not filed by December 17, 2026, or a filing restates a period the rule reads, the test is inconclusive, and the Result says what was read and what was missing.", "specState": "approved", "specVersion": 1, "specUpdatedOn": "2026-10-22", "specHistory": [ { "version": 1, "date": "2026-10-22", "reason": "First specification.", "by": "Desk Editor (level A)" } ], "callExpects": "no", "format": "ladder", "group": "th060-mrvl-revenue-yoy", "threshold": 60, "direction": "above", "unit": "%", "think": { "baseRate": "Marvell's revenue growth from a year earlier was 36.5% for the quarter ended August 1, 2026, by its filing on SEC EDGAR.", "outsideView": "The year-earlier quarter is not on the same concept in the filings, so the rungs sit on the last print alone.", "wouldChange": "The results-calendar consensus for the quarter ended October 31, 2026 implies 52.0%, and the middle rung is centred there." } }, "url": "https://deepstack.ltd/docket/#test-th060-t04", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 22, "at": "2026-10-22T06:20:45Z" }, { "model": "deepseek-reasoner", "p": 12, "at": "2026-10-22T06:21:00Z" } ], "readers": null }, { "id": "th060-t05", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "December 9, 2026", "whenLabel": "Dec 9, 2026", "date": "2026-12-09", "precision": "day", "event": "Broadcom, gross margin, quarter ended November 1, 2026: whether it comes in at or above 67.5%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "pre", "callExpects": "yes", "callExpectsBy": "Desk Editor (level A)", "callExpectsAt": "2026-10-22T08:14:29Z", "callProbability": 74, "callProbabilityAt": "2026-10-22T08:14:29Z", "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[4].date", "url": null }, "resolver": { "kind": "earnings", "symbol": "AVGO", "rule": "Divide the gross profit by the revenue that Broadcom reports for the quarter ended November 1, 2026 in its 10-Q or 10-K filing on SEC EDGAR. The answer is yes if the ratio is at or above 67.5%; no if it is below 67.5%.", "question": "Did the gross profit of Broadcom for the quarter ended November 1, 2026 come to at or above 67.5% of its revenue?", "source": { "label": "SEC EDGAR, XBRL company facts of Broadcom (Forms 10-Q and 10-K)", "url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0001730168.json" }, "ambiguity": "If Broadcom has not filed by January 7, 2027, or a filing restates a period the rule reads, the test is inconclusive, and the Result says what was read and what was missing.", "specState": "approved", "specVersion": 1, "specUpdatedOn": "2026-10-22", "specHistory": [ { "version": 1, "date": "2026-10-22", "reason": "First specification.", "by": "Desk Editor (level A)" } ], "callExpects": "yes", "format": "ladder", "group": "th060-avgo-gross-margin", "threshold": 67.5, "direction": "above", "unit": "%", "think": { "baseRate": "Broadcom's gross margin was 69.1% for the quarter ended August 2, 2026, by its filing on SEC EDGAR.", "outsideView": "A year earlier, for the quarter ended August 3, 2025, the same line read 67.1%, so the rungs sit on the last print.", "wouldChange": "A change in product mix, a one-time charge, or a reclassified line would move the reading." } }, "url": "https://deepstack.ltd/docket/#test-th060-t05", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 45, "at": "2026-10-22T06:21:04Z" }, { "model": "deepseek-reasoner", "p": 74, "at": "2026-10-22T06:21:23Z" } ], "readers": null }, { "id": "th060-t06", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "December 9, 2026", "whenLabel": "Dec 9, 2026", "date": "2026-12-09", "precision": "day", "event": "Broadcom, gross margin, quarter ended November 1, 2026: whether it comes in at or above 69%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "pre", "callExpects": "either", "callExpectsBy": "Desk Editor (level A)", "callExpectsAt": "2026-10-22T08:14:29Z", "callProbability": 46, "callProbabilityAt": "2026-10-22T08:14:29Z", "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[5].date", "url": null }, "resolver": { "kind": "earnings", "symbol": "AVGO", "rule": "Divide the gross profit by the revenue that Broadcom reports for the quarter ended November 1, 2026 in its 10-Q or 10-K filing on SEC EDGAR. The answer is yes if the ratio is at or above 69.0%; no if it is below 69.0%.", "question": "Did the gross profit of Broadcom for the quarter ended November 1, 2026 come to at or above 69.0% of its revenue?", "source": { "label": "SEC EDGAR, XBRL company facts of Broadcom (Forms 10-Q and 10-K)", "url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0001730168.json" }, "ambiguity": "If Broadcom has not filed by January 7, 2027, or a filing restates a period the rule reads, the test is inconclusive, and the Result says what was read and what was missing.", "specState": "approved", "specVersion": 1, "specUpdatedOn": "2026-10-22", "specHistory": [ { "version": 1, "date": "2026-10-22", "reason": "First specification.", "by": "Desk Editor (level A)" } ], "callExpects": "either", "format": "ladder", "group": "th060-avgo-gross-margin", "threshold": 69, "direction": "above", "unit": "%", "think": { "baseRate": "Broadcom's gross margin was 69.1% for the quarter ended August 2, 2026, by its filing on SEC EDGAR.", "outsideView": "A year earlier, for the quarter ended August 3, 2025, the same line read 67.1%, so the rungs sit on the last print.", "wouldChange": "A change in product mix, a one-time charge, or a reclassified line would move the reading." } }, "url": "https://deepstack.ltd/docket/#test-th060-t06", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 26, "at": "2026-10-22T06:21:27Z" }, { "model": "deepseek-reasoner", "p": 56, "at": "2026-10-22T06:22:12Z" } ], "readers": null }, { "id": "th060-t07", "part": null, "thesis": 60, "slug": "aws-custom-silicon-margin-proof", "when": "December 9, 2026", "whenLabel": "Dec 9, 2026", "date": "2026-12-09", "precision": "day", "event": "Broadcom, gross margin, quarter ended November 1, 2026: whether it comes in at or above 70.5%.", "status": "pending", "resultId": null, "sameAs": null, "stage": "pre", "callExpects": "no", "callExpectsBy": "Desk Editor (level A)", "callExpectsAt": "2026-10-22T08:14:29Z", "callProbability": 20, "callProbabilityAt": "2026-10-22T08:14:29Z", "callProbabilityHistory": [], "source": { "kind": "thesis", "ref": "tests[6].date", "url": null }, "resolver": { "kind": "earnings", "symbol": "AVGO", "rule": "Divide the gross profit by the revenue that Broadcom reports for the quarter ended November 1, 2026 in its 10-Q or 10-K filing on SEC EDGAR. The answer is yes if the ratio is at or above 70.5%; no if it is below 70.5%.", "question": "Did the gross profit of Broadcom for the quarter ended November 1, 2026 come to at or above 70.5% of its revenue?", "source": { "label": "SEC EDGAR, XBRL company facts of Broadcom (Forms 10-Q and 10-K)", "url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0001730168.json" }, "ambiguity": "If Broadcom has not filed by January 7, 2027, or a filing restates a period the rule reads, the test is inconclusive, and the Result says what was read and what was missing.", "specState": "approved", "specVersion": 1, "specUpdatedOn": "2026-10-22", "specHistory": [ { "version": 1, "date": "2026-10-22", "reason": "First specification.", "by": "Desk Editor (level A)" } ], "callExpects": "no", "format": "ladder", "group": "th060-avgo-gross-margin", "threshold": 70.5, "direction": "above", "unit": "%", "think": { "baseRate": "Broadcom's gross margin was 69.1% for the quarter ended August 2, 2026, by its filing on SEC EDGAR.", "outsideView": "A year earlier, for the quarter ended August 3, 2025, the same line read 67.1%, so the rungs sit on the last print.", "wouldChange": "A change in product mix, a one-time charge, or a reclassified line would move the reading." } }, "url": "https://deepstack.ltd/docket/#test-th060-t07", "storyUrl": "https://deepstack.ltd/theses/aws-custom-silicon-margin-proof/#tests", "resultUrl": null, "machines": [ { "model": "claude-opus-5", "p": 21, "at": "2026-10-22T06:22:17Z" }, { "model": "deepseek-reasoner", "p": 22, "at": "2026-10-22T06:22:59Z" } ], "readers": null } ] }