President Donald Trump’s OGE Form 278-T for August 2026 lists 32 transactions in nine large AI companies, 23 purchases and 9 sales, by our count of all 517 lines. The largest is a purchase of Meta Platforms of $5,000,001 to $25,000,000 on 21 August, the only line in the report in that band.
The 23 purchases add up to $8,081,023 to $34,540,000 and the 9 sales to $2,006,009 to $7,090,000, by our arithmetic from the ranges the form gives. The 32 lines are 6.2% of the report’s 517 transactions. The form gives value ranges, not share counts or prices.
Business Pill · Concentration Explained
An evergreen Business Engineer explainer on concentration, where a few holdings account for most of the value. It teaches the general idea only and does not discuss any company named here or the President.
The key insight: As we read the count, the AI lines in the report are mostly purchases, 23 of 32, but the totals rest on one line: a single Meta Platforms purchase of at least $5,000,001 is 61.9% of the low end of all 23, by our arithmetic, so the mix of line counts and the mix of money tell different stories.
We checked every line for ten companies: Nvidia, Microsoft, Alphabet, Meta Platforms, Amazon, Broadcom, Advanced Micro Devices, Taiwan Semiconductor Manufacturing, Oracle and Apple. Nine appear. In the 517 rows we read, there is no line for Taiwan Semiconductor Manufacturing.
Nvidia and Microsoft have six lines each, five purchases and one sale. Meta Platforms has five lines, four purchases and one sale. Oracle has four lines, three purchases and one sale. Alphabet has three lines, two purchases and one sale. Amazon has three purchases and no sale. Broadcom has one purchase and two sales. Advanced Micro Devices and Apple have one sale each and no purchase.
By combined purchase range, Meta Platforms leads at $5,151,004 to $25,365,000, followed by Microsoft at $1,616,005 to $6,315,000 and Nvidia at $767,005 to $1,580,000, by our arithmetic. The largest sale among the nine is Advanced Micro Devices at $1,000,001 to $5,000,000 on 21 August.
Number of purchase and sale lines for nine AI companies in the President’s OGE Form 278-T for August 2026, by our count of all 517 lines: 23 purchases and 9 sales. Ten companies were checked; Taiwan Semiconductor Manufacturing has no line. The form gives value ranges, not share counts or prices.
One Line Sets the Total
The single Meta Platforms purchase dated 21 August is at least $5,000,001, which is 61.9% of the $8,081,023 low end of all 23 purchases, by our arithmetic. Without that line, the other 22 purchases add up to $3,081,022 to $9,540,000.
The same date carries three of the nine sales in this group: Nvidia and Broadcom at $500,001 to $1,000,000 each and Advanced Micro Devices at $1,000,001 to $5,000,000. Together those three come to $2,000,003 to $7,000,000, by our arithmetic. The report lists other 21 August sales as well, including Micron Technology, Marvell Technology and Applied Materials at $250,001 to $500,000 each.
How the Disclosure Works
The form’s summary page says the 278-T “discloses purchases, sales, or exchanges of securities in excess of $1,000 made on behalf of the filer, the filer’s spouse, or dependent child”. It adds that transactions “are required to be disclosed within 30 days of receiving notification of a transaction but not later than 45 days after the transaction”.
The report was signed on 17 September, received by OGE on 25 September and posted on the OGE index on 8 October. Our earlier piece on the Nvidia lines in the same report has been corrected: it listed five Nvidia lines and missed line 261, a purchase of $1,001 to $15,000 dated 24 August.
The Structural Read
As we read the 21 August lines, the same date carries the largest purchase in this group, Meta Platforms at $5,000,001 to $25,000,000, and three of its nine sales, Nvidia, Broadcom and Advanced Micro Devices at $2,000,003 to $7,000,000 combined, by our arithmetic.
As we read the spread of names, purchases cluster in the companies that sell cloud and AI software, Microsoft, Meta Platforms, Amazon, Oracle and Alphabet, while the two sales-only names, Advanced Micro Devices and Apple, each have one line; the form gives no reason for any of them.
OGE Form 278-T, summary of contents
“Transactions are required to be disclosed within 30 days of receiving notification of a transaction but not later than 45 days after the transaction.”
Three Implications
FOR INVESTORS The 278-T gives value ranges only, so the 23 purchases in this group are between $8,081,023 and $34,540,000, by our arithmetic; exact amounts are not disclosed.
FOR GOVERNANCE WATCHERS The form records an agency ethics official’s opinion dated 25 September and an OGE certification digitally signed on 6 October; the comments box is blank.
FOR READERS OF THE NVIDIA STORY Our earlier piece on the six Nvidia lines lists each Nvidia line in the same report.
The framework says of Layer 9, Governance: “Layer 9 answers to geopolitics, not to engineering.”
As we read it, a disclosure regime that publishes the head of government’s trades across the chip, cloud and model companies of the stack is part of that governance layer. The framework predates this report and does not discuss it.
Combined purchase range per company in the President’s OGE Form 278-T, August 2026, by our arithmetic from the ranges. Meta Platforms includes one purchase of $5,000,001 to $25,000,000 on 21 August, at least 61.9% of the low end of all 23 purchases. Advanced Micro Devices and Apple have sales only.
What Is Not Established
The report does not say who decided any of these trades or whether an outside adviser manages the accounts, and we found no statement on that in the document. Because each line is a range, the true totals can sit anywhere inside the ranges given here. The publication did not contact the White House for this piece.
Our list of ten companies is our own choice of large AI names; other technology companies in the report, such as Micron Technology or Intel, are not in the count. The report does not link any transaction to later company events, and this piece draws no such link.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
The President’s August 278-T lists 32 lines in nine large AI companies, 23 purchases and 9 sales, by our count. Purchases add up to $8,081,023 to $34,540,000, and one Meta Platforms purchase of $5,000,001 to $25,000,000 is at least 61.9% of the low end, by our arithmetic.
A note on sourcing. We read the President’s OGE Form 278-T (signed 17 September 2026, posted 8 October) in full and checked every counted line against the page images. The choice of ten companies is ours. Counts, totals and percentages are our arithmetic from the ranges in the form. Under our editorial policy this piece covers governance and markets only and implies no motive from the timing of any transaction. The White House was not contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
