brianletort.ai
Back to Issue 11

Capital flow drill-down

Money in, revenue out — the full breakdown.

Issue 11 · Week 27 of 2026.

The four-category capital scorecard expanded with named transactions, burn-to-revenue context, and a per-category trend across recent issues. Each row corresponds to a row in the summary table on the issue page.

Category

Frontier Labs.

OpenAI, Anthropic, Google DeepMind, xAI

Capital in

~$95B

vs ~$95B

Revenue out

~$21B

vs ~$21B

Burn / rev

~1.3x

Lower means more capital out than in.

The read

Both S-1s remain at confidential-draft stage, with Deutsche Bank framing each offering at ~$60B. The strategic tell is pricing: closed labs are cutting rates ahead of the listings rather than harvesting margin. Investors should read the price cuts as demand-generation for the capacity buildout — and as pressure on every model-layer competitor's unit economics going into IPO season.

Trend across recent issues

W20W21W22W23W24W25W26W27
Capital inRevenue out/$B per issue

Category

Hyperscaler-Hosted.

Azure-OpenAI, AWS-Anthropic, Google Cloud-Gemini, Oracle-OCI

Capital in

~$187B

vs ~$187B

Revenue out

~$62B

vs ~$62B

Burn / rev

~3.0x

Lower means more capital out than in.

The read

Aggregate 2026 top-4 capex of ~$725B (+77% YoY) is now colliding with operating cash flow, and the Epoch crossover analysis became the consensus frame this week. Meta selling raw GPU capacity and SoftBank standing up a 10GW-target neocloud are best read as responses to that pressure. Buyers gain negotiating leverage; investors should watch Alphabet's Jul 28 print as the first test of the crossover narrative against actuals.

Trend across recent issues

W20W21W22W23W24W25W26W27
Capital inRevenue out/$B per issue

Category

Neoclouds.

CoreWeave, Nscale, Crusoe, Lambda, Fluidstack, IREN

Capital in

~$13.5B

vs ~$12.7B

Revenue out

~$5B

vs ~$5B

Burn / rev

~2.7x

Lower means more capital out than in.

The read

The category absorbed two opposing signals in one week: fresh capital at rising marks (Together AI, Crusoe talks) and a violent public-market repricing of customer-concentration risk once Meta's compute-rental plan surfaced — Nebius carries a ~$27B Meta deal and CoreWeave ~$21B. Operators should diversify anchor-customer exposure in any multi-year neocloud contract; investors should now price hyperscaler competition into the category's cost of capital.

This week’s transactions

  • 2026-07-01

    Together AI Series C at $8.3B valuation, led by Aramco Ventures, plus >500MW of investor-capitalized compute commitments

    Together AI

    $800M

Trend across recent issues

W20W21W22W23W24W25W26W27
Capital inRevenue out/$B per issue

Category

On-Prem / Hybrid.

Enterprise GPU clusters, sovereign and national programs, Cisco / Dell / HPE

Capital in

~$94B

vs ~$94B

Revenue out

~$36B

vs ~$36B

Burn / rev

~2.6x

Lower means more capital out than in.

The read

A reported ~$590B Korean government-plus-chaebol chip-hub plan needs primary confirmation before it moves this category. For hybrid buyers the real news is procedural: the Jul 9 FERC intervenor deadline and PJM's July filings will determine whether large-load interconnection accelerates or becomes a regional litigation map. Site strategy should still start with power and grid process, not GPU SKUs.

Trend across recent issues

W20W21W22W23W24W25W26W27
Capital inRevenue out/$B per issue

Methodology

Capital-in and revenue-out figures are quarterly or annualized as noted in the row, sourced from public earnings disclosures, SEC filings, and lab and vendor announcements. Burn-to-Revenue is revenue divided by committed capital. The trend chart shows up to eight prior issues. Data is updated weekly; revisions in future issues do not retroactively edit this one.

Back to the full issue