SemiAnalysis Weekly

Recorded Election Day morning, before any results: SemiAnalysis’s model has about 62 GW of US datacenter additions in 2028, most not yet under construction and sitting in states electing a governor. Texas alone is close to 30%.

Jordan Nanos (@JordanNanos) sits down with Maya Barkin (@mayabarkin), Wayne Ma (@waynema) and Jeremie Eliahou Ontiveros (@JeremieEO) on why 2027 is largely locked in, why the backlash crosses party lines, and how developers can win over their neighbors.

Since recording: the AP called Texas for Gina Hinojosa, who would pause datacenter permitting, and Georgia for Keisha Lance Bottoms, and Democrats took a 3-2 majority on Georgia’s Public Service Commission, which sets power rates. (As of Nov 4, 2:35 PM ET.)

Read the article: https://newsletter.semianalysis.com/p/the-datacenter-backlash-goes-bipartisan

0:00 Intro

0:54 The 2028 Buildout

5:32 Pause vs. Conditions

10:03 Bipartisan Backlash

14:37 Grid Risk

16:30 Tax Incentives

19:51 Texas

24:34 Risks to 2027

30:23 Georgia

32:31 Federal Outlook

35:20 Winning Over Neighbors

42:04 What We're Watching

SemiAnalysis just launched GDPx, a benchmark built with 14 specialized data partners across 11 domains where frontier intelligence earns the highest returns. Even the best model scores just 12%.

Jordan Nanos (@JordanNanos) sits down with Max Kan (@maxkan) and Andrew Megalaa (@AndrewMegalaa) to explain why most popular benchmarks are broken, the difference between difficult and unfair tasks, how a real private equity deal room became a finance environment, why binary grading still matters, and what the leaderboard says about Opus 5.5, GPT-6 Astra and Gemini 4 Carbon.

Explore GDPx: https://gdpx.semianalysis.com

Read the article: https://newsletter.semianalysis.com/p/gdpx-measuring-frontier-model-capabilities

Want your model evaluated or to partner on a future version? [email protected]

0:00 Cold Open

0:10 Launch Week Chaos

4:02 Broken Benchmarks

5:15 Difficult vs. Unfair

7:33 Building GDPx

9:25 Frontier Domains

11:02 The 11 Domains

14:12 Domain Experts

15:00 Models vs. Experts

19:36 Model Behavior

20:44 The Leaderboard

22:33 Cyber and Safety

24:32 Frontier vs. Day-to-Day

27:13 The Finance Task

31:27 Binary vs. Partial Grading

36:48 AI and Jobs

42:43 Quick Takes

43:44 Google's Comeback

47:25 Public vs. Private Tasks

49:52 What's Next

Anthropic filed its S-1: revenue per dollar of compute is up 5.2x since Q1 2025, a customer who spent $1 last year now spends $9, and incremental gross margins are running near 90%. Jordan Nanos (@JordanNanos), Doug O'Laughlin (@fabknowledge), Jeremie Eliahou Ontiveros (@JeremieEO) and Joey Brookhart (@SaasquatchC) break down profitability, the half trillion in compute commitments, customer concentration, the 2027 compute roadmap, and whether OpenAI can keep up.

0:00 Cold Open

0:53 Profitability

2:10 The Datacenter View

4:14 Compute Commitments

6:28 Incremental Margins

8:19 Customer Concentration

12:00 Top Customers

16:53 New Funding

18:54 Underpromise, Overdeliver

20:25 Elon and Neoclouds

25:14 Compute Efficiency

29:40 Can Anyone Catch Up?

33:26 Surprises and Runway

37:47 Market Reaction

39:02 Holiday Usage

42:56 Is OpenAI Cooked?

48:44 Government Revenue

50:45 Wrap-Up

Every hybrid bonder on the market today brings the die to the wafer. ASML's patents bring the wafer to the die.Sherry Zhang and Howie join Jordan Nanos (@jordananos) to dig into the patents behind ASML's die-to-wafer hybrid bonder, why a track-based design could beat Besi on alignment and particle control, and why HBM won't adopt die-to-wafer hybrid bonding before 2029. The team also looks at how China is using hybrid bonding to keep scaling past its lithography limits, and what larger photomasks mean for High-NA EUV.Chapters:0:00 The Patent Tip-Off1:53 Hybrid Bonding 1014:06 Die-to-Wafer vs Wafer-to-Wafer5:43 Back End vs Front End8:05 Why HBM Waits Until 202913:45 Everything Comes to the Die18:26 Alignment and Cleanliness22:48 Bonding Instead of Shrinking25:49 China's Toolmakers Step Up28:25 Bigger Masks for High-NA33:29 Two Tool Stacks?Read the full article: ASML Is Building a Hybrid Bonding Tool!https://newsletter.semianalysis.com/p/asml-is-building-a-hybrid-bonding

Brad Gerstner put our 43 gigawatt forecast for 2027 on All-In, then again in his BG2 sit-down with Jensen, and called it too aggressive. The consensus counter sits near 25 GW, built on moratorium headlines and grid connection fears. Jordan Nanos (@JordanNanos) sits down with Jeremie Ontiveros (@JeremieEO) and Reyk Knuhtsen (@robotknower), who have been tracking the same capacity curve since March 2024, and it has barely moved across every model revision. So they aren't cutting it. They're raising it.

Read the article: https://newsletter.semianalysis.com/p/datacenters-2027-the-year-of-acceleration

0:00 Cold Open

1:17 The 2027 Forecast Debate

2:21 Model Track Record

5:27 Where the Pushback Comes From

8:16 Delays and Haircuts

12:31 The Top 40 Sites

15:23 Leading Indicators

20:40 Modular Builds

25:41 Chips vs Datacenters

27:53 Responding to Brad Gerstner

GLM 5.3 only attends to the tokens that matter. It still has to keep every one of them.

Kimbo Chen (@kimbochen) and Wenyao Gao (@gao_wenyao) join Jordan Nanos (@JordanNanos) to dig into AMD catching NVIDIA on InferenceX, how post-training alone pushed an open-weight model within reach of Mythos, and whether Z.ai quietly tuned it for Chinese chips. The team also looks at how GLM 5.3 became a fallback when Claude and GPT refuse security work.

Chapters:

0:00 Cold Open

0:37 Sparse Attention 101

4:46 HiCache and HiSparse

10:13 AMD Catches NVIDIA

12:54 Capacity vs Bandwidth

18:44 The Frankenstein Flash

20:09 Mythos-Class From China?

22:52 Post-Training Pipeline

32:32 Chinese Chips, Day One

39:54 GLM 5.4 and Beyond

Read the full article: How GLM 5.3 Sparse Attention Affects HBM Memory Usage

https://newsletter.semianalysis.com/p/sparse-savings-persistent-demand-inside-glm53

H100 rental prices went up, on a GPU that should be sliding down its depreciation curve. That curve is what lenders underwrite. Daniel Nishball (@dnishball) kicks off our new Compute Capital Markets group by following the money through Nvidia's $497B backstops, ClusterMAX as a credit signal, and whether each token is profitable. Jordan Nanos (@JordanNanos) brings the rental data. AI debt is about to pass auto loans.

0:00 Opening

1:48 Nvidia's Backstop Universe

6:22 The $11T Funding Wall

17:25 Balance Sheet as a Service

26:27 ClusterMAX as Credit Risk

31:51 Why H100 Prices Rose

43:09 Token Profitability

58:05 You Can't Go Back

1:07:25 AI Debt vs Auto Loans

1:11:14 Inside Nvidia's Backstops

Pay Anthropic $200 a month and you can pull $12,000 of Opus 5.5 tokens at API pricing. Max Kan (@maxkan) and Andrew Megalaa bought every plan, OpenAI, Meta, xAI, MiniMax, Moonshot, Z.ai, Cursor, and measured every token type to find the ceiling. The catch isn't subsidy. Anthropic still books 50 to 60% gross margins on these plans, which means the API itself is the price gouge.

0:00 Cold Open

1:03 Subsidies and Credits

5:23 Token Types and Cache

14:08 Model Value Charts

20:51 Test Methodology

26:54 Other Plans

31:26 Utilization and Limits

36:24 Token Efficiency

41:24 Open Models

44:22 Cyber and Capture

Read More: https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x

Jordan (@JordanNanos) and Abhilash (@AbhilashJain17) discuss the clients SemiAnalysis consulting serves, custom projects like the inference simulator, how client work turns into research products, neocloud due diligence, and why SLAs decide who gets funded. They also discuss Feynman specs, de-specced memory, and a 20-trillion-parameter OpenAI model.

0:00 Intro and Overview

2:27 Client Types

5:24 Custom Projects

9:47 Strategy Work

12:27 Consulting to Products

14:33 Neocloud Due Diligence

18:40 SLAs and Insurance

22:32 Risk Assessments

26:45 Build vs Buy

29:19 Favorites and Outlook

The InferenceX team is back after two months. Jordan Nanos (@JordanNanos) is joined by Cam Quilici (@noslawextratost), Bryan Shan, and Alec Ibarra to discuss Engram and DRAM offloading, the AgentX benchmark and its profit calculator, TPU v7 on InferenceX, Vera Rubin vs Blackwell, fast tokens, and TileRT on NVIDIA GPUs.

0:00 Intro

0:31 Engrams

12:27 AgentX Benchmark

16:36 Profit Calculator

23:40 TPU v7

32:03 Vera Rubin

40:49 Fast Tokens

49:33 TileRT

54:00 Chip Startups

1:01:31 Wrap Up