Press and Media

Sean

— 3 min read
Share
Press and Media

Second Order by TKN is a research desk on the AI infrastructure trade. It maps who else gets paid past the headline name: memory, packaging, rented GPU hours, hyperscaler capex, data-center power, and the markets that supply them. Every page states what would change its view. Explained pages carry a "last reviewed" date and a source list. The free daily Wire gives one dated number, who invoices and who waits. Second Order is made by TKN, the team behind The Kopi Notes. It is education about how the stack works, not investment advice: no tip sheets, no buy or sell calls, no price targets.

Who writes it

Sean writes Second Order by TKN. He covers AI infrastructure for international readers:

  • chips and packaging
  • memory
  • power and cooling
  • neoclouds
  • hyperscalers
  • the landlords and suppliers who get paid when capacity binds

His method is to look one link further down the chain than the headline (who sells the memory, who powers the hall, who owns the building), then write down what would prove that read wrong. Sean is not a licensed adviser and does not publish buy lists.

Short bio

Sean writes Second Order by TKN, a research desk on the AI infrastructure trade. He covers chips and packaging, memory, power and cooling, neoclouds, hyperscalers, and the landlords and suppliers who get paid when capacity binds. Sean is not a licensed adviser and does not publish buy lists.

Long bio

Sean writes Second Order by TKN (sotkn.com), a research desk on the AI trade made by TKN, the team behind The Kopi Notes. He covers AI infrastructure for international readers: chips and packaging, memory (including HBM), power and cooling, neoclouds, hyperscalers, and the landlords and suppliers who get paid when capacity binds.

When a story is loud, he wants to know who else gets paid. Everyone watches the chip; he asks who sells it memory, who packages it, who rents the GPU hours, and who writes the capex cheque. His method is to look one link further down the chain than the headline, then state what would prove the read wrong. Pages are built on primary sources first (company filings, regulator releases, grid-operator data), with every figure dated and labeled as Reported, Guided, Our calculation or Our read.

Read our media note on the AI infrastructure research desk.

Sean is not a licensed adviser. Second Order does not publish tip sheets, buy or sell calls, or price targets; its work is education about how the AI infrastructure stack works and where money can stick.

What Sean can comment on

Background or on-the-record comment, focused on industry structure and supply chains:

  • AI chips and accelerators: GPUs, TPUs and custom chips, and who designs, makes and packages them
  • High-bandwidth memory (HBM): supply, allocation, and why memory and packaging have to arrive matched
  • Advanced packaging: CoWoS and the packaging gate
  • Neoclouds: rented GPU fleets, contracts, utilization, and who carries idle hours
  • Hyperscaler capex: what the lines in Microsoft, Alphabet, Amazon and Meta results show, and who gets paid down the chain
  • AI Data Centers and power: grid connections, large-load tariffs, on-site power, cooling, and why the building is the quick part
  • Asia's supply chain: Taiwan, Korea, Singapore and Malaysia as markets in the AI build-out

What Sean doesn't do: stock picks, buy or sell calls, price targets, or comment on whether any share is cheap or expensive.

Free to cite

How we work

Contact

  • Media and interview requests: [email protected]
  • Corrections: [email protected], with the page URL and what looks wrong. This is the address published on our editorial standards page.