Serenity, the small-cap-focused stock-picker who back-maps the AI buildout to hunt the obscure upstream suppliers everyone quietly depends on, spent Wednesday tracking a theme he has pushed for months: co-packaged optics (CPO), the practice of placing the optical engine right next to the switch or GPU die to save power and boost bandwidth. His read today was that the long-promised CPO ramp is no longer theoretical — the money is starting to show up, and it is landing on the testing companies first.

Test and measurement leads the ramp. Serenity flagged commentary from Viavi Solutions (VIAV), a test-and-measurement firm whose gear checks yield and performance across optical and memory supply chains. Management pushed back hard on chatter that CPO is slipping on yield problems, insisting the technology is moving forward and that the company already has purchase orders for CPO-test work, with revenue expected to accelerate into December. Viavi also said 1.6T optical links are ramping fast and should reach parity with 800G next year — a data point Serenity treats as confirmation that the broader optical cycle is healthy. He extends the same logic to FormFactor (FORM), the wafer-probe and measurement supplier he views as a beneficiary of the memory supercycle, HBM4 yield ramp and the emerging silicon-photonics test bottleneck. His takeaway: because test vendors get paid as customers qualify new parts, seeing revenue hit VIAV and FORM now is an early signal the CPO wave is real, even as full "scale-up" volume waits until H2 2027–2028. It's worth noting he does not claim personal positions in either name; he treats them as ecosystem indicators rather than conviction bets.

The surprise: Aeva turns into an optical-source player. The day's genuine curveball, in Serenity's telling, was Aeva Technologies (AEVA) — a company known for 4D FMCW LiDAR that he already owns as a speculative bet on physical AI and humanoids. Aeva jumped more than 18% after hours on news of a joint development agreement with an optical-engine partner to deploy at a major hyperscaler, with production ramping in 2028. What matters to Serenity is the supply-chain read-through. Aeva's plan leans on external light-source products and on-chip integrated lasers — exactly the parts he believes come from Sivers (SIVE), the tiny Swedish maker of indium-phosphide continuous-wave DFB lasers that is his single highest-conviction holding.

Why Sivers is the payoff. Serenity's entire framework centers on finding the smallest merchant chokepoint the giants depend on, and for the CPO buildout he argues that is the laser light source — the piece that feeds silicon-photonics engines. Sivers already supplies CW lasers into Aeva's LiDAR effort, so an Aeva optical-connectivity program opens a fresh addressable market and a hyperscaler pathway for the same upstream supplier. He reads Aeva potentially becoming a CPO/near-package-optics customer as a clean positive for Sivers, which he has repeatedly called his top pick and a candidate for the "next Lumentum."

The net of the day for Serenity: the CPO timelines he has been mapping look intact, revenue is arriving on schedule at the test layer, and an unexpected new program reinforced his conviction in the laser chokepoint underpinning the whole thesis. As always, this is derived commentary, not investment advice.