Use Case Intelligence RSS Search use cases Search
Google Cloud Software / SaaS San Francisco, CA, USA

Anthropic

Document Processing · IT Operations Automation · Real-time Analytics

During large-scale training runs, traffic is highly variable—dropping to near zero during model checkpointing before bursting to millions of events per minute. Anthropic needed a serverless messaging queue that scaled dynamically without requiring manual partition management. One of Anthropic’s top priorities was reducing latency.

Read the original story Explore all use cases
Handles peak throughput of over 1M events per second
Ingests over 500 billion daily rows per Pub/Sub topic
Processes 3 GBps of sustained model training metrics
Reduces consumer-side ingestion latency by 88%

Reported impact

Business outcome

Handles peak throughput of over 1M events per second; Ingests over 500 billion daily rows per Pub/Sub topic; Processes 3 GBps of sustained model training metrics; Achieves sub-second event latency from publish to query.

  • Cost Reduction
  • Productivity Improvement
  • Revenue Growth
  • Scalability

Solution

Producers serialize events into Protocol Buffers and apply zstd compression. If the compressed payload is under 8 MB, the client publishes the message directly to a Pub/Sub topic. If the payload is 8 MB or larger, the client writes the compressed bytes directly to a Cloud Storage bucket and publishes a lightweight JSON pointer notification containing the object URI through Pub/Sub. The downstream consumer uses this notification to fetch and decompress the payload from Cloud Storage. By reusing the pre-compressed bytes for the Cloud Storage upload, the pipeline avoids redundant compression cycles, keeping the messaging broker entirely unblocked. A block diagram mapping Anthropic's multi-cloud ingestion setup. It shows AWS-based and Google Cloud-based producers securely connecting over Private Service Connect to regional Pub/Sub topics. The diagram displays events fanning out to BigQuery, Dataflow, Cloud Storage, and external SIEM systems.

Data flow

Source data is captured through Cloud Pub/Sub, stored or unified in BigQuery, Cloud Storage, Bigtable

Solution architecture

4 components · 2 layers
  1. Ingest
    • Cloud Pub/Sub Moves events between services so the platform can process updates asynchronously at scale.
  2. Storage
    • BigQuery Stores and queries the analytics and AI-ready datasets that power the workload.
    • Cloud Storage Stores large files, lake data, or static assets that feed downstream analytics and applications.
    • Bigtable Stores high-volume low-latency operational data for real-time workloads.

Architecture clues

  • Achieves sub-second event latency from publish to query
  • BigQuery Dataflow Cloud Storage Bigtable
  • Reduces consumer-side ingestion latency by 88%

Evidence from the source

Achieves sub-second event latency from publish to query
Handles peak throughput of over 1M events per second
Ingests over 500 billion daily rows per Pub/Sub topic
Processes 3 GBps of sustained model training metrics
Reduces consumer-side ingestion latency by 88%
Transparency Evidence & methodology 90/100 reference value · 79% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value90/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth20/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity20/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact25/25
    5 metrics describe measurable impact.
  • Deployment scale5/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence79%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence20/20
    5 direct source quotes retained.
  • Date provenance0/10
    The publication date has no documented provenance.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity10/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. Google Cloud

    Anthropic customer story

    First-party
    https://cloud.google.com/customers/anthropic-pubsub
    Source type
    Vendor-published customer story
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes

Date evidence

  • Sitemap last modified Oct 9, 2026 · Google Cloud · Publisher metadata Publisher-reported modification timestamp; not a publication date.
    Open date source
  • Unverified extracted date Oct 9, 2026 · Google Cloud · Unknown Legacy extracted date without source-page confirmation.
    Open date source
Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Related

Similar use cases

See all like this
Google Cloud Archived · Mar 14, 2026 ComplyAdvantage Software / SaaS United States For ComplyAdvantage , a financial crime prevention platform, this transformation presented an opportunity to scale its AI-driven services to get ahead of increasing and evolving threats. 100% improvement in merge requests per developer Document ProcessingIT Operations Automation Gemini · Gemini Code Assist Google Cloud Modified · Jun 26, 2026 Hallam Commercial and Professional Services United Kingdom Hallam builds custom no-code agents with Gemini Enterprise, removing development bottlenecks and scaling customer insights. For digital marketing agencies, capturing client attention is harder than ever. Online… Accelerates agent deployment from months to 3-4 hours with Gemini Enterprise Document ProcessingIT Operations Automation BigQuery · Gemini · Gemini Enterprise Google Cloud Archived · Oct 16, 2026 Alta Software / SaaS United States Alta needed a transactable Google Cloud Marketplace path for its AI and technology spend management platform without pulling its product team off core forecasting work, while also helping finance and IT leaders… Saved 125 engineering and 60 project management hours IT Operations AutomationReal-time Analytics Cloud Pub/Sub · Cloud Run · Cloud SQL