Daily Alpha · Substack
· Post-Market Alpha · by Buzzberg Research
The substantive non-crypto signals were regulatory risk to AI deployment, behind-the-meter power as the data-center fix, and a sharply higher optics TAM.
Themes on this desk
AI regulation
A bipartisan AI safety bill could introduce deployment approval or certification, threatening data-center and semiconductor forecasts.
AI power bottleneck
Behind-the-meter gas generation is framed as the default power solution for AI data centers.
Optics TAM
Goldman doubled its 2028 AI optics TAM to $148B, with 1.6T and 3.2T modules driving the revision.
Bipartisan AI safety bill in the works could impose a speed limit on frontier AI
The article cites a Semafor report that Sens. Amy Klobuchar, Ted Cruz and Majority Leader John Thune are working on bipartisan AI safety legislation that could be introduced soon, with details not yet public; other lawmakers are floating model certification, liability regimes, new regulatory authority, or an outright pause on superintelligence.
A credible legislative path to restricting frontier model deployment introduces a regulatory risk to the AI capital-spending complex that the author argues is not priced.
Watch Whether a bill is formally introduced and whether its text includes deployment approval, certification, liability, or training thresholds.
Source →SemiAnalysis claims first call on onsite gas generation as primary AI power solution
SemiAnalysis states that last year it was the first to call out onsite gas generation as the primary method adopted by AI labs and hyperscalers to solve power constraints, and that its positive view was far from consensus, with behind-the-meter primary power solutions having been dismissed under various names.
If behind-the-meter gas generation is becoming the default power solution, gas turbine, generation equipment and related suppliers are levered to data-center buildout independent of grid interconnection queues.
Watch Announced onsite gas generation capacity at AI campuses and order books at gas turbine and generator suppliers.
Source →Goldman doubled its 2028 AI optics TAM forecast to $148B in five months
The article states Goldman Sachs doubled its 2028 forecast for the AI optics total addressable market from $69B to $148B in five months, and that the piece breaks down what changed, how 1.6T and 3.2T modules affect the market, and who can supply it.
A rapid doubling of the optics TAM estimate signals accelerating demand expectations for high-speed optical modules tied to AI networking.
Watch Whether other sell-side firms revise AI optics TAM estimates upward and whether 1.6T module shipments ramp on schedule.
Source →