ビジネス · AI売上高の会計処理OpenAIの売上高200億ドルの差は、会計上の違いであり、売上が消えたわけではない執筆 2026年10月9日 21:04 UTC視聴文字起こし0:09Core Weave shares fell 7.77% in a sell-off that hit artificial intelligence0:14suppliers.0:15OpenAI's September revenue pace was reportedly about $50 billion a year, below the0:20$68 to $70 billion headline investors had seen.0:23Different ways of counting partner sales explain the gap.0:26The next test is whether OpenAI can reach its year-end target on a consistent0:29accounting basis.0:30OpenAI reportedly gave investors the $50 billion figure for September's end.0:35An annualized run rate projects a recent sales pace over a full year.0:39That pace can change.0:40And the company still has operating costs to pay.0:43The figure doesn't tell us how much cash OpenAI collected during 2026 or how much0:48profit it made.0:49It also doesn't guarantee those sales are contractually recurring.0:52The correction concerns what investors thought the revenue headline measured.0:55The earlier $68 to $70 billion came from investors adjusting OpenAI's numbers to1:00match Anthropic's partner sales method.1:02OpenAI's own figure counted its retained share of certain partner sales.1:06That produces an $18 to $20 billion measurement gap without showing an equivalent1:10loss of sales.1:11A customer payment can support two different revenue headlines.1:14The underlying demand can stay exactly the same.1:17The $68 billion and near $70 billion estimates were variants of the same adjusted1:21figure, rather than separate streams of revenue.1:24Anthropic counts the full customer payment through cloud partners as revenue, then1:28treats the partner's cut as an expense.1:29OpenAI records only its own share of certain partner sales.1:33Each approach depends on who controls the customer relationship and delivery.1:36Both treatments have been described as complying with generally accepted accounting1:40principles.1:41That description doesn't amount to an audit of either company.1:44Earlier explanations suggested OpenAI's figure came after payments to Microsoft and1:48Amazon.1:49The more precise explanation concerns its share of partner sales.1:52Computing costs still have to be paid separately.1:55Anthropic's reported pace reached about $65 billion at July's end.2:00OpenAI's $50 billion observation came two months later.2:04Those different dates and accounting bases prevent these figures from telling us2:08which company had the larger comparable revenue business.2:11OpenAI's investor presentation also described rapid expansion during the July to2:16September quarter.2:17Its total revenue run rate grew 77%, while its enterprise run rate grew 107%.2:24Those percentages describe sequential growth within the third quarter.2:27Enterprise customers are businesses paying for AI services.2:30The 107% figure measures how fast that business grew.2:34It doesn't mean enterprise supplied that percentage of total sales.2:37OpenAI expects these customers to drive much of its next increase.2:40OpenAI began 2026 at a reported annualized pace of about $20 billion,2:46a figure of about $6 billion circulated in 2024.2:50Those older numbers give a sense of expansion, but changing measurement details make2:54them unsuitable for a precise, like-for-like growth calculation.2:57OpenAI reportedly still expects its annualized pace to reach at least $70 billion by3:02December 31st.3:03That is a separate $70 billion figure with a future deadline.3:07Enterprise growth is expected to supply much of the increase.3:10If September and December use the same measurement basis, Reaching that target3:14requires about $20 billion more in annualized pace.3:17That's roughly 40% growth during the final quarter.3:19OpenAI competes for business customers with Google Cloud and Meta, but their effect3:23on its pricing and growth hasn't been quantified.3:26NVIDIA closed down 2.94% on October 8th.3:29Oracle lost 5.48%, and Coreweave fell 7.77%.3:35The NASDAQ composite declined 1.25%.3:38The heavier losses in AI suppliers show where concerns about demand behind the3:42infrastructure buildout were concentrated.3:44These closing percentages refine the earlier rounded descriptions of 3, 6, 8, and3:491%.3:50The revenue clarification coincided with the sell-off, and concern about AI demand3:54may have contributed.3:55Rising treasury yields also pressured markets, so the revenue news alone can't3:59explain every percentage point lost.4:00NVIDIA finished at $230.48, Oracle at $135.69,4:07and Coreweave at $81.58.4:09The NASDAQ closed at $27,193.34.4:14losing 345.35 points, OpenAI is privately held, so the immediate losses reduce the4:20value of investors' holdings in other companies.4:22OpenAI was also asking prospective investors to consider another enormous financing4:26round.4:27Discussions disclosed in late September sought at least $30 billion at a valuation4:32of about $1.4 trillion before adding the new funding.4:35March's round brought $122 billion in commitments at an $852 billion4:41valuation that included the new money.4:43September's proposed valuation excludes new funding.4:46Comparing the two headlines directly would mix different valuation bases.4:50The new discussions were early stage and terms could change.4:53No closing had been announced through October 9th.4:56The proposed round would provide bridge financing while a public offering was4:59deferred beyond 2026.5:01Investors were weighing that trillion dollar valuation as the meaning of the revenue5:05headline shifted.5:06Amazon committed up to $50 billion in the earlier round.5:10NVIDIA and SoftBank committed 30 billion each alongside other investors.5:14Some funding depended on conditions or scheduled payments, so the commitment total5:18doesn't show that all the cash had arrived.5:21The larger investment question is whether paying demand can support the computing5:24build out.5:25More AI processing doesn't automatically produce proportionally more revenue.5:30Token consumption measures model processing, while prices and margins determine how5:34much that activity can earn.5:35Revenue growth has to cover computing, staffing, and other costs before it can5:39produce profit.5:40Open AI's disclosures don't give investors a complete picture of profitability or5:44cash generation.5:45A faster sales pace alone can't settle the question of returns.5:53Goldman Sachs forecasts an even larger industry buildout.5:57Its estimate covers the five biggest American hyperscalers, the companies operating6:01enormous cloud computing systems.6:03It puts their AI infrastructure spending at about $800 billion in 2026.6:07Goldman expects about $1.2 trillion in 2027 against a Wall Street consensus of about6:13$1.1 trillion.6:14These figures cover several cloud companies' infrastructure plans.6:17They describe expected spending across that group, rather than OpenAI's own budget.6:21The underlying analysis projects about 54% growth in 2027 after nearly 100%6:28in 2026.6:29The rounded 800 billion and 1.2 trillion endpoints imply 50%.6:34Rounding explains why those displayed totals don't reproduce the underlying growth6:37rate exactly.6:38A broader estimate put next year's spending by hyperscalers and newer cloud6:42providers at $1.2 to $1.3 trillion.6:45The boundaries of that group weren't clear.6:47Comparing several companies' spending with one lab's revenue can't measure the6:51sector's overall return.6:52The accounting explanation accounts for the smaller OpenAI headline without6:56demonstrating a sudden collapse in demand, but OpenAI hasn't publicly supplied the6:59detailed reconciliation or calculation behind its September pace.7:03Investors still need a consistent basis for judging whether its growth can support7:06the valuation.7:07OpenAI's next measurable test arrives on December 31st.7:11It expects at least a $70 billion annualized pace, with enterprise customers driving7:16much of the increase.7:17Reaching that level on a clearly stated, consistent basis would show how much paying7:21demand has actually grown.記事を読むストーリーの時系列HeyDexショート動画 ↗誰が何を報じたか複数のメディアが報じた論点OpenAI’s annualized revenue at September’s end2媒体EU efforts to limit Chinese plug-in hybrid exports2媒体1社だけが報じた論点22Center for Strategic & International Studiesのみが報道Chinese automakers’ share of European auto sales10月5日 15:34 UTCConnected-vehicle data security risk10月5日 15:34 UTCBloomberg Televisionのみが報道EU consideration of limits on Chinese hybrid imports10月8日 6:57 UTCOpenAI’s expected year-end annualized revenue10月9日 5:41 UTCPBoC rejection of claims that the renminbi is undervalued10月9日 5:41 UTCDeadline for progress in EU-China negotiations10月8日 10:13 UTCPossible Chinese retaliation against EU trade measures10月8日 6:57 UTCSchedule for EU-China trade talks10月8日 11:16 UTCEU-China talks alongside PBoC currency-policy defense10月9日 5:41 UTCEU trade chief’s visit to China10月8日 10:13 UTCEnterprise business contribution to OpenAI revenue growth10月9日 5:41 UTCChina’s rare-earth leverage in a trade confrontation10月8日 11:16 UTCChinese share of hybrid cars in the EU10月8日 6:57 UTCRevenue growth needed to support AI investment returns10月9日 5:41 UTCEU tariffs on Chinese battery-electric vehicles10月8日 6:57 UTCCNBCのみが報道Earlier estimate of OpenAI revenue and its accounting basis10月8日 16:42 UTCNvidia share movement on October 810月8日 16:42 UTCOracle share movement on October 810月8日 16:42 UTCCoreWeave share movement on October 810月8日 16:42 UTCOpenAI third-quarter run-rate growth10月8日 16:42 UTCNasdaq movement on October 810月8日 16:42 UTCOpenAI enterprise business growth10月8日 16:42 UTC各行は、そのメディアの報道から私たちが記録した主張です。2026年10月13日 22:04 UTC時点で比較。関連報道背景·10月15日 5:53 UTC韓国の取引停止は売りを一時停止させたが、大幅な損失は残った
