Newsletter comparison
SemiAnalysis vs Fabricated Knowledge
Short answer: Choose SemiAnalysis for detailed AI-compute architecture and economics. Choose Fabricated Knowledge for a broader semiconductor and robotics view tied to company strategy. Keep both if your thesis crosses chips, systems, and end markets.[1][2][3][4][5][6]
An independent guide. Scottie isn't affiliated with, endorsed by, or sponsored by SemiAnalysis and Fabricated Knowledge. Each publication owns its name, writing, and subscription terms.
A brief from both newsletters · July 31, 2026
See both newsletters in one brief
See what made the brief, what didn't, and the original links behind every included story.
In this brief: Scottie included material from SemiAnalysis and left out the issues it read from Fabricated Knowledge.
Reader priorities
A technology investor or operator needs technically credible analysis of semiconductors, AI infrastructure, and the companies exposed to each shift.
These are illustrative priorities, not a customer’s data.
What made the brief
- SemiAnalysis2 read · 2 included
- Fabricated Knowledge1 read · 0 included
Scottie
Scottie example brief
July 31, 2026 · Executive brief
01 / The rundown
- Modular datacenter construction is surging to bypass labor shortages, cutting build times by over seven months and expanding vendor dollar content.
- AMD is closing its software gap with Nvidia to win major hyperscale deals, though internal testing bottlenecks pose short-term execution risks.
02 / The briefing
01 / main
Datacenters Turn To Modular Assembly To Beat Labor Shortages
The takeaway: Datacenter builders are switching to modular, factory-built rooms and steel shells to bypass crippling trade labor shortages. Prefabricated components pull repeatable electrical and cooling work offsite, cutting build times significantly while reducing unit capital costs across major hyperscale deployments.
Concrete details
- Modular construction cuts datacenter build timelines by 36% and lowers capital expenditures by roughly 8% per megawatt.
- Industry tracker covers over 61 gigawatts of modular capacity, projecting 30% live capacity penetration by 2028.
Why it matters for this reader: Directly impacts your infrastructure model as suppliers like Vertiv double their content value per megawatt while builders bypass critical labor shortages.
Original sourcesSemiAnalysis
02 / main
AMD Closes CUDA Gap But Internal GPU Bottlenecks Persist
The takeaway: AMD is making real strides against Nvidia software dominance, securing large deployments with Anthropic and Microsoft. However, severe internal shortages of GPU clusters for automated testing and complex rack manufacturing hurdles threaten to slow its momentum before full rollout.
Concrete details
- Anthropic committed to deploy 2 gigawatts of AMD chips, while Microsoft plans MI455X Helios rack deployments.
- Helios racks require over 550 Broadcom ethernet retimers per rack due to backplane signal challenges.
Why it matters for this reader: Evaluates AMD ability to erode Nvidia moat, informing hardware allocations and supply chain exposure across AI accelerator investments.
Original sourcesSemiAnalysis
Action items
- Audit AI infrastructure holdings for exposure to modular equipment suppliers capturing expanded per-megawatt content value.
- Monitor AMD software CI cluster allocations to verify release stability before adjusting accelerator portfolio exposure.
See every issue behind this brief
Modular Datacenter Construction Shifts
- The Wild Wild West Of LEGO DatacentersSemiAnalysis · Included
AMD Advances Against Nvidia's CUDA Moat
- Can AMD break the CUDA Moat? AMD Advancing AI 2026SemiAnalysis · Included
Fabricated Knowledge update
- Close but Not QuiteFabricated Knowledge · Read, not included
The useful difference
The difference that matters
SemiAnalysis goes deepest on frontier AI compute, accelerator economics, and infrastructure cost. Fabricated Knowledge ranges more widely across semiconductors, robotics, software analogies, and company strategy.[1][2][3][4][5][6]
What each is best for
SemiAnalysis
SemiAnalysis contributes detailed work on accelerators, inference cost, hardware roadmaps, and the constraints shaping frontier AI deployments.[1][2][3]
Fabricated Knowledge
Fabricated Knowledge contributes a wider company and industry lens, including robotics and the strategic implications of technical change.[4][5][6]
When it’s worth reading both
The pair is useful when a component-level shift must be translated into a company or market thesis. Scottie can join the original links without flattening different technical assumptions.[1][2][3][4][5][6]