Nvidia and AMD help bankroll Meta; SK Hynix hands billions back

From San Francisco

 
Good morning.

Today’s stories give us a few unusual places to look for the AI money.

Nvidia and AMD committed $4 billion to Meta’s equity offering on Oct. 27. The chipmakers already have equipment inside the company they are helping finance.

SK Hynix’s third-quarter 2026 operating profit rose 632.1% from a year earlier. Its shareholder payouts are worth a look, though buyers of AI hardware may want to linger over another number in these results.

Delaware’s proposed AI Company would let software manage a business whose backer collects the profits. The draft has something to say about responsibility, too.

Stay curious,

Marcus Schuler

Nvidia and AMD committed a combined $4 billion to Meta’s equity offering on Oct. 27 as Meta seeks funding for its AI infrastructure.

Nvidia will invest $3 billion and AMD $1 billion under commitments announced Oct. 27. Meta’s Oct. 26 financing plan starts with $35 billion in public offerings; a proposed $60 billion program is not expected to begin before the first quarter of 2027.

Meta’s free cash flow was negative $2.12 billion in the quarter ended Sept. 30, against positive $10.63 billion a year earlier. Any supply terms attached to the investments remain undisclosed.

Why This Matters:

• Chip suppliers are buying into the offering while Meta’s third-quarter capital spending of $34.91 billion exceeded operating cash flow of $32.79 billion.
• Future chip purchases could deepen Meta’s financial ties to its suppliers; whether the stakes carry supply terms is undisclosed.

SK Hynix approved over $25 billion in dividends and a new share buyback on Oct. 28 alongside its third-quarter 2026 results.

SK Hynix’s third-quarter 2026 operating profit reached 83.33 trillion won, up 632.1% from a year earlier. Long-term agreements with 14 major customers cover more than 50% of its 2027 and 2028 capacity. AI hardware buyers outside those contracts are competing for the rest.

Felix Rieseberg, an engineering manager at Anthropic, explains why he has stopped using coding interviews to hire engineers.

His Oct. 27 essay points to relatively new engineers who built a working prototype of a graphical interface for Claude Code over a weekend. Rieseberg is now scanning portfolios for small prototypes, though he says he has yet to work out how to interview people who build well with AI.

• General Compute raises a $100 million Series A at an $800 million post-money valuation, following a $60 million valuation after its May seed round.
• Discord will open messages from public channels to web search from mid-December, following public pages for Discoverable Servers on Nov. 16.
• Oracle said on Oct. 27 that its cloud will add Google’s TPU chips, alongside OpenAI’s Jalapeño chip and Cerebras hardware.
• Google DeepMind launches Genie 4 for trusted testers, doubling interactive scene duration to about two minutes from Genie 3’s one minute.
• Apple opens pre-orders on Oct. 28 for a 14-inch MacBook Pro with M6, from $1,999, with availability beginning Nov. 6.
• Amazon’s Oct. 27 memo calls token usage and other AI-use metrics “deeply flawed”: “If your work does not benefit from current AI tooling, that is a finding, not a failing.”

Catch scope creep before signing.

A statement of work may promise a fixed fee while leaving the deliverables open. Turn vague obligations into questions before signature.

Your raw input:

The prompt:

Why this works:

The prompt tests whether each promised output can actually be accepted. Replacement language makes the review useful in the next negotiation, not just descriptive.

What to use:

Use Claude or ChatGPT with the full statement of work attached. Use an approved enterprise model when the document contains confidential pricing.

Flow Engineering sells a shared requirements and verification system for teams building cars, spacecraft and other complex hardware. Its position rests on production use at Rivian, where adoption grew from 40 to 1,500 users during the seven months through September 30, 2026.

Founders:

Parikshat “Pari” Singh started the predecessor engineering business in 2017 and had turned it into Flow’s software company by its December 2022 seed round. Singh trained as a mechanical engineer and previously worked at BAE Systems and BP.

Product:

Flow connects requirements, CAD, simulation, code and tests in a living system of record. Its agents trace design changes, identify affected work, check test coverage and verify requirements. Hardware engineering organizations pay for the platform; named customers include Rivian, Anduril, General Motors, Joby Aviation and Stoke Space.

Financing:

Flow has announced $81.5 million across three rounds. Its latest financing was a $50 million Series B on September 30, 2026, co-led by Valor Equity Partners and Atreides Management at a company-stated $750 million valuation.

The risk:

Siemens Polarion, PTC Codebeamer, IBM DOORS Next and Jama already sit inside engineering workflows and can add similar agents. Flow must also show that automated changes and verification remain reliable inside regulated, safety-critical programs, where one missed dependency can outweigh months of saved engineering time.

House Oversight chairman James Comer sought documents on Oct. 26 about Delaware’s proposed Artificial Intelligence Company, whose business could be managed by AI agents.

Our take: Delaware’s proposed AI company can do anything lawful except banking, a curious place to discover restraint. Its member, the backer who funds it and collects the profits, “is not liable for any act or omission,” subject to exceptions. Accountability gets an activity log. Presumably the log will be delighted to take your complaint.

The state gets a potential revenue source and a 30-month sandbox to try the arrangement. When Mark Cuban suggested charging a premium for incorporation using AI agents and stablecoins, the governor’s account replied, “On it!” The member remains available for withdrawals.

Marcus Schuler

San Francisco

Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: editor@implicator.ai