CNBC Scott Wapner with Brad Gerstner — May 28th, 2026
13-minute CNBC Halftime hit. Two halves: the **Invest America 'Trump accounts' app launch** (Treasury child-investment accounts, $1,000 auto-funded per child under 2 on July 4, invested in 'the best 500 companies'), and a **dense run of trade signals**. **The token-flow thesis: data-infrastructure companies (Snowflake, Databricks, ClickHouse) are 'in the token flow' — more tokens consumed accelerates their core data business.** Gerstner called the Snowflake quarter two weeks ago and it printed a blockbuster: **33% growth (accelerating) vs 27% consensus, full-year guide raised by more than the Q1 beat.** Snowflake remains a large personal holding but he moved out of it in the funds — **Micron and 'AAM' are both up ~200% YTD vs Snowflake +10% YTD**, so capital was better deployed there. Portfolio discipline: the **3/6/9 heuristic (30/60/90% = small/medium/large)**; take profits on parabolic moves that exceed price targets, keep dry powder. **SpaceX IPO rumoured June 12, raising ~$75B, ~$1.5T market cap, early index inclusion** — Gerstner says the **SpaceX↔Anthropic data-centre deal + the Cursor-team acquisition** is what changes the IPO from a launch/Starlink story to an AI-compute story: *'there's no human being on the planet better at turning electrons into tokens than Elon Musk.'* Notes ~$86B of near-term tech IPO supply (SpaceX + Anthropic + OpenAI) but frames it against ~$100T capital-market depth; flags a possible 10-20% SOX consolidation as 'run of the mill'.
Key points
- **Token-flow thesis — the cleanest framing of the week: 'data is the oil that runs the AI engine.'** Gerstner: *'data infrastructure companies like Snowflake, Databricks and ClickHouse, they're all in the token flow. As more tokens are consumed, it accelerates their core data business.'* Explicitly distinct from *'software apps like Salesforce that more directly compete with the AI model companies.'* **This is the actionable inversion of the [model-commoditisation thesis in this week's All-In](/issues/2026-05-31)** — if models commoditise, the data layer underneath captures the token-volume growth.
- **Snowflake blockbuster, called two weeks early.** *'33% growth, accelerating growth versus consensus of 27% growth, they took up their full year guide by more than their Q1 beat.'* Two drivers: AI accelerating the core (more tokens → faster core growth) and Snowflake Intelligence + Cortex Code hitting critical mass in Q1. Same dynamic in Databricks and ClickHouse (*'may be the fastest growing of the bunch'*).
- **The portfolio-rotation receipt — why he's OUT of Snowflake in the funds despite loving it.** *'Year to date, Snowflake is up 10% even including today's moves. Compare that to what we moved the money into like Micron, which is up 200% year to date, or AAM, which is up 200% year to date.'* Snowflake stays a large personal holding. **Direct validation of [Gavin Baker's long-memory / DRAM thesis from Issue 05-06](/issues/2026-05-24)** — Micron +200% YTD is the memory-cohort trade paying off in real time.
- **Portfolio-sizing discipline: the 3/6/9 heuristic.** *'Think about managing your portfolio in three sizes — small, medium and large. 3, 6, 9 — 30%, 60%, 90%.'* When everything is washed out and despondent (the start of the Iran war), go large; when euphoric, take a large position down to medium. On parabolic names exceeding price targets: *'we absolutely are making sure we're taking down some of our exposure... rotating it into other names'* — but *'let your winners run.'*
- **SpaceX IPO — the AI-compute reframe is the whole story.** Rumoured **June 12, ~$75B raise, early index inclusion**. Gerstner: *'the launch business was great, Starlink was great, but that alone would not justify the valuation. Him building the data centres which are going to power all of AI, and using those to train up his own AI with the incredible team he acquired from Cursor, I think is a total game changer for the IPO.'* And the line of the hit: *'there's no human being on the planet that is better at turning electrons into tokens than Elon Musk.'* **Continues the [SpaceX S1 / Elon Web Services thread from Issue 06 All-In](/issues/2026-05-24)** — now with Gerstner sizing it as a buyer.
- **IPO-supply absorption: $86B of near-term tech supply vs $100T capital-market depth.** Josh Brown's question — can the market absorb SpaceX + Anthropic + OpenAI all at once? Gerstner: he'll have to **sell existing positions to fund SpaceX** (*'we don't have unlimited capital'*), but *'remember the depth of our capital markets — 100 trillion relative to 75 billion.'* The bigger near-term risk: *'we could have a 10 to 20% consolidation in the SOX... a run of the mill consolidation'* — so keep cash on the sideline as dry powder.
- **Invest America 'Trump accounts' app live in both app stores.** Treasury-backed child investment accounts: **$1,000 auto-funded for every child under 2 on July 4**; $250-500 for ages 2-10 (state-dependent); a 'free lifetime 401k' for 10-18. Money invested passively in 'the best 500 companies in America' (*'a little bit of Nvidia, a little bit of Microsoft, a little bit of Walmart'*). Framed as a Main-Street direct-ownership program; Gerstner personally funding $250 into 700 elementary kids' accounts in Durham, NC via QR code.
Notable quotes
Data infrastructure companies like Snowflake, Databricks and ClickHouse, they're all in the token flow. As more tokens are consumed, it accelerates their core data business. This is a very different story than software apps like Salesforce that more directly compete with the AI model companies.
Snowflake is up 10% even including today's moves. Compare that to what we moved the money into like Micron, which is up 200% year to date, or AAM, which is up 200% year to date. So for our investors, that was the right thing for us to do.
Think about managing your portfolio in three sizes. Small, medium and large. 3, 6, 9 — 30%, 60%, 90%. When everybody gets euphoric, you may want to take your portfolio from that large position to a medium sized position.
There's no human being on the planet that is better at turning electrons into tokens than Elon Musk. He's building the biggest data centers on Earth and eventually in space, of anybody.
Him building the data centers which are going to power all of AI, and using those to train up his own AI with the incredible team he acquired from Cursor, I think is a total game changer for the IPO and puts them in a really good position.
Today is not the day that you shove all in on the market. Go back to the heuristic. If you have zero in the market today and you want to put $100,000 to work, I would put 30% to work because you have to start at some point and then you wait for your moments.
Themes
- Data-infrastructure token-flow trade
- Memory cohort (Micron) outperformance
- SpaceX IPO as AI-compute play
- Portfolio sizing & risk discipline
- AI IPO supply wave