A - Landscape Analysis

Statutory text

Government Code section 11546.8, subdivision (f)(2)(A), requires the Report to include:

(A) A landscape analysis of California’s current public, private, and nonprofit cloud computing platform infrastructure.

Source: SB-53 Gov Code 11546.8.

Reading list

This list groups the library’s sources for this element by tier. It does not say what the Report will recommend.

Primary (34)

• 2026-10-14 SB-53 Assembly Floor analysis: The analysis gives a cost figure for the consortium’s own work.
• California Report on Frontier AI Policy 2025-06-17: The working group report does not mention CalCompute in the excerpt.
• CDT About the Department: CDT oversees state IT planning, projects, procurement and standards; runs the State Data Center; Director is State CIO.
• CDT homepage: Thin homepage: Poppy, middle mile, Digital ID; training and AI community links. Lists the pages taken.
• ESP Action CalCompute public cloud release 2024-02-08 (web): The co-sponsor calls CalCompute a “public cloud”. Teri Olle: people with ideas “do not have access to the tools they need”.
• ESP Governor signs SB-53 release 2025-09-29 (web): The co-sponsor says compute is “dominated by just three companies: Amazon, Google, and Microsoft” and CalCompute is “a public option”.
• Fei-Fei Li Senate testimony 2023-09-14 (web): Fei-Fei Li tells a U.S. Senate committee that compute and data needs leave frontier AI to “a select few industry players”. She backs the National AI Research Resource.
• Governor AI working group release 2025-03-18: The release announces the draft of the Governor’s working group report.
• Governor California 4th largest economy 2025-04-23 (web): California becomes the fourth-largest economy, $4.1 trillion against Japan’s $4.02 trillion (IMF and BEA). No CalCompute.
• Governor economy tops 4.5 trillion 2026-10-05 (web): California output tops $4.5 trillion in 2025 (BEA) and “remains” the fourth-largest economy. No CalCompute.
• Governor ICYMI 32 of top 50 AI companies 2025-03-12 (web): The origin of “32 of 50”, linked to the Forbes AI 50. It says “fifth largest economy”. No CalCompute.
• Governor TIME Person of the Year made in California 2025-12-12 (web): “33 of the 50 top AI companies”, 15.7% of U.S. AI job postings, and over half of global AI venture capital in the Bay Area. No CalCompute.
• GovOps more efficient California
• Massachusetts AI Hub release: $31M state grant to MGHPCC for AI Compute Resources; about $120M over five years
• Newsom SB-53 signing release 2025-09-29: The release describes CalCompute in one line, under “Innovation”.
• NSF 26-508 TechAccess solicitation (web)
• NSF 26-513 solicitation: NSF AI Infrastructure Hubs: one award per state, first deadline Nov. 4, 2026; partners fund five years of compute, NSF funds staff and training
• Pathways to College in California: C2C data story on students from high school to college and work, by district
• Poppy overview: Poppy is a generative AI (GenAI) platform that CDT built for state employees.
• Poppy service page: CDT charges state departments for Poppy.
• Poppy Terms of Use (web): These are the terms between CDT and each “Ordering Entity” for Poppy, the state’s generative AI (GenAI) platform.
• SB 1047 CalCompute section 2024-08-29: SB 1047’s CalCompute section: the same elements as SB-53, report due Jan. 1, 2026; vetoed
• SB-53 Assembly Privacy analysis 2025-07-08: The committee’s summary of the bill (items 17 to 28) already matches the enacted § 11546.8 almost item for item.
• SB-53 Assembly Privacy analysis 2025-09-05: This is the last committee analysis of SB 53, on the version the Governor signed.
• SB-53 Gov Code 11546.8: The statute: Gov. Code § 11546.8, subdivisions (a) to (k), word for word. The base of every definition.
• SB-53 Senate GO analysis 2025-03-25: Excerpt of the CalCompute passages of the Senate Governmental Organization analysis. The full copy is the (web) page.
• SB-53 Senate GO analysis 2025-03-25 (web): The first committee analysis of SB 53. It has the enacted mission, UC placement and seven elements, with the March appointment scheme. It says “This bill does not establish CalCompute” and records no opposition.
• SB-53 Senate Judiciary analysis 2025-04-08
• UC Regents B4 Update on AI 2026-09-16: UC is scoping its own “Sovereign AI Cloud”, UC-owned and UC-governed. It does not name CalCompute.
• Wiener SB 1047 introduction release 2024-02-07 (web): SB 1047’s release names CalCompute users as “startups, researchers, and community groups” and asks cloud firms for transparent pricing.
• Wiener SB 294 announcement 2023-09-13
• Wiener SB-53 expansion release 2025-07-09: The author’s office describes CalCompute as more settled than the statute does.
• Wiener SB-53 introduction release 2025-02-28 (web): SB 53 “takes the first step toward establishing CalCompute”, with “low-cost compute”. It compares New York’s $90 million AI consortium.
• Wiener SB-53 signing release 2025-09-29 (web): The author calls CalCompute “industrial policy” for startups and researchers. Teri Olle calls it “a public option for computing power”.

Peer-reviewed (2)

• PEARC 2025 Cultivating Computational Equity Central Valley: CENVAL-ARC: Central Valley campuses use under 0.05% of ACCESS allocations; training and shared infrastructure
• PEARC 2025 Empire AI model for provisioning AI and HPC: Empire AI: $250M state capital, $150M from members, nonprofit governance, member-ranked proposals

Institutional (26)

• ACCESS Allocations policy: This is the full ACCESS rule set for who can get compute and how ACCESS decides.
• ACCESS Get your first project: ACCESS gives a first project in five steps.
• ACCESS homepage: This is a thin homepage of about 286 words.
• ACCESS Our impact: ACCESS reports 67,925 “Users with an allocation” and 2,055 “Research institutions served” since its launch in September 2022.
• ACCESS Project types: ACCESS has four project types: Explore, Discover, Accelerate and Maximize.
• Cal-ICOR About: Partners (UC Berkeley lead, Learning Lab, 2i2c) and five named staff.
• Cal-ICOR documentation
• Cal-ICOR homepage: 1,907 students this term and 4,369 all time (UC 275, CSU 1,497, CCC 134); Learning Lab supports it; “Free”. Lists the pages taken.
• Cal-ICOR hub usage: 13 institutions on 20 hubs; users by term 9 to 1,907; summer use drops about 90 percent.
• Cal-ICOR Partner With Us: 20 partner institutions (15 CSU, 2 UC, 3 CCC) and a four-step onboarding with a choice of support model.
• CENVAL-ARC homepage: UC Merced-led regional research computing network; NSF CC* award; symposium March 2027
• CITRIS UC Santa Cruz homepage: Four-campus UC institute since 2001; no compute or AI program at UC Santa Cruz. Lists the pages taken.
• CloudBank at Berkeley: 2020 UC Berkeley data science college (CDSS) news: NSF CloudBank (SDSC, UC Berkeley) brokers public cloud for research and teaching
• CMU Policy Innovation Lab: Heinz College course page: students build public interest technology for government clients; no compute facts
• CNCF landscape: The saved copy has no content.
• cncf landscape2 on GitHub: landscape2 is a free, open-source command line tool that builds a static, interactive landscape website from YAML (a plain-text data format) files.
• Columbia Empire AI page: 700,000 service units per institution; $0.50 per unit, half paid by Columbia; free low-priority overflow
• Epoch AI data centers: 96 of the largest AI data centers, with power, compute and cost estimates; no California site among them
• Epoch AI GPU clusters: Over 500 GPU clusters with performance, power, cost and owner; CC-BY; a cross-check and model for the survey’s fields
• LBNL 2024 US Data Center Energy Usage Report: US data centers: 176 TWh in 2023 (4.4% of electricity), 325-580 TWh in 2028; California second to Virginia in demand
• NAIRR Pilot resource allocations (web)
• NRP Getting started (web): NRP has no proposal review.
• Public AI Network 2024 white paper: Public access, accountability and permanent public goods; names CalCompute as a local public compute effort
• SDSC NAIRR Operations Center launch: NSF chose SDSC to lead the National AI Research Resource (NAIRR) Operations Center.
• UC Davis HPC clusters: Farm, Franklin and Hive: 21,776 cores and 139 GPUs in the table, “over 48,000” cores in the text
• UC Davis HPC rates

Know a source this list is missing? Use Suggest reading on the site, or write to team@californiacompute.org with the subject “Suggested reading”.