Atlasresearch

Workspace

WorkspacePokedexHPC MapPortfolioCalendarTaiwan RevenueMemory
Calculators
  • 01AI Factory
  • 02CPO Stack
  • 03HBM Bridge
  • 04GPU Cloud TCO
  • 05Inference Tokens
  • 06Power Bottleneck
  • 07Custom ASIC
  • 08OCS TAM
  • 09N3 Allocation
  • 10Optical TAM
Agent APISubscribe
Sign in
Calculator
Atlasresearch

Workspace

WorkspacePokedexHPC MapPortfolioCalendarTaiwan RevenueMemory
Calculators
  • 01AI Factory
  • 02CPO Stack
  • 03HBM Bridge
  • 04GPU Cloud TCO
  • 05Inference Tokens
  • 06Power Bottleneck
  • 07Custom ASIC
  • 08OCS TAM
  • 09N3 Allocation
  • 10Optical TAM
Agent APISubscribe
Sign in
Calculators/OCS TAM

Jason's Chips

Subscribe or sign in with your Substack email for live inputs, outputs, and saved scenarios.

Subscribe to Jason's ChipsSign in

Inputs

Google TPU superpod

High-radix scale-out

Medium-radix scale-up

Supplier economics

Last calculated: Not yet

Calculator 08Optical circuit switching

OCS TAM Calculator

Convert TPU superpod geometry, high-radix spine adoption, and medium-radix optical scale-up attach into merchant OCS units, supplier revenue, gross profit, port count, and EPS power displacement.

Sensitivity

External OCS market

$2.06B

22.5K external boxes at $91.5K avg ASP

Lumentum OCS revenue

$1.24B

56% gross margin assumption

Lumentum units

13.6K

$55.1K LITE revenue / external box

Total OCS ports

9.77M

38.5K total OCS boxes including internal

Power savings

105.1 MW

$69.0M annual energy value

OCS gross profit

$695M

4.6 MW OCS vs 109.7 MW EPS

Revenue bridge

Three OCS demand buckets

Google TPU Superpod OCS

$835M

TPU superpod and Google external merchant OCS demand.

High Radix Scale-Out

$126M

Spine, super-spine, and scale-across high-radix OCS.

Medium Radix Scale-Up

$280M

60-80 radix OCS used around Phase 1 optical scale-up.

BucketTotal boxesMerchant boxesLITE boxesASPExternal marketLITE revenue
Google TPU Superpod OCS28.6K12.9K7.7K$108.0K$1.39B$835M
High Radix Scale-Out2.1K1.8K1.1K$118.0K$218M$126M
Medium Radix Scale-Up7.8K7.8K4.8K$58.0K$451M$280M

OCS versus EPS

Optical switching power displacement

Total OCS boxes

38.5K

Merchant plus internal deployments

OCS power

4.6 MW

125W per optical box

EPS equivalent

109.7 MW

3,000W per packet switch

Annual energy savings

$69.0M

$75.00 / MWh at 95% load

Sensitivity analysis

Google merchant share x Lumentum share

Merchant \ LITE45%53%60% base68%75%
30%$824M$893M$963M$1.03B$1.10B
38%$928M$1.02B$1.10B$1.19B$1.28B
45% base$1.03B$1.14B$1.24B$1.35B$1.45B
53%$1.14B$1.26B$1.38B$1.50B$1.62B
60%$1.24B$1.38B$1.52B$1.66B$1.80B

This grid isolates the two variables that matter most while Google is still the dominant buyer: how much of TPU superpod demand moves external and how much of that merchant demand Lumentum captures.

Preset comparison

Same framework, different adoption curves

PresetExternal boxesExternal marketLITE revenueGross profitPower savings
CY26 Google6.2K$652M$330M$182M51.5 MW
CY27 run-rate22.5K$2.06B$1.24B$695M105.1 MW
Conservative8.9K$883M$460M$244M63.3 MW
Scale-up Phase 129.9K$2.38B$1.42B$798M123.1 MW

Assumptions

CY27 run-rate

Demand caseLumentum run-rate support case
Preset noteAnnualized run-rate case that lets Google externalization, non-Google high-radix OCS, and Phase 1 scale-up all contribute.
Total OCS boxes38.5K
External boxes22.5K
Average external ASP$91.5K
LITE revenue / box$55.1K
LITE gross profit$695M
Total ports9.77M

Saved scenarios

Google-heavy CY26

6.7K boxes / $748M market

$412M

CY27 run-rate

16.0K boxes / $1.72B market

$1.03B

Scale-up Phase 1

14.5K boxes / $1.25B market

$760M