Salesforce Insights for Regulated Industries | Vantage Point

Salesforce Multi-Cloud: AWS and Google Cloud Expansions Explained

Written by David Cockrum | Nov 7, 2026, 12:59:59 PM

At Dreamforce 2026, Salesforce did something it has never done at this scale: it announced major expansions with two rival hyperscalers at the same time. On the AWS side, Salesforce business context is coming to Amazon Quick, AWS frontier agents are coming to Slack, Agentforce gains model choice through Amazon Bedrock, zero-copy data access extends across more AWS data services, and Agentforce Voice will interoperate with Amazon Connect Customer. On the Google Cloud side, Hyperforce is going live on Google Cloud infrastructure, Gemini Enterprise connects to Salesforce's headless architecture, Commerce Cloud powers checkout on Google surfaces, and Data 360 zero-copy sharing with BigQuery expands.

For enterprise architecture teams, the message is unmistakable: the agentic enterprise is multi-cloud. Model choice, infrastructure choice, and data choice are all opening at once — and the decisions you make about cloud commitments, data gravity, and agent governance in the next two quarters will shape your CRM architecture for years.

This post breaks down what changed, who is affected, what to do now, and when each capability arrives — with capability claims attributed to Salesforce, AWS, and Google Cloud's own announcements.

Quick Answer

 

Salesforce announced expanded partnerships with both AWS and Google Cloud during Dreamforce 2026, turning its platform into a genuinely multi-cloud foundation for enterprise AI. Per the announcements: AWS gains native Salesforce context in Amazon Quick, AWS agents in Slack, broader Agentforce model choice through Amazon Bedrock, expanded zero-copy data access, and Agentforce Voice interoperability with Amazon Connect Customer. Google Cloud gains Hyperforce production workloads (North America GA November 2026), Salesforce capabilities inside Gemini Enterprise, Commerce Cloud checkout on Google surfaces, and expanded Data 360–BigQuery zero-copy sharing. This matters most for CIOs, enterprise architects, and RevOps leaders who hold AWS or Google Cloud commitments and run Salesforce. Vantage Point helps teams translate these announcements into a practical cloud and AI architecture plan.

Key Takeaways (TL;DR)

  • What changed: Salesforce deepened both hyperscaler partnerships at once — AWS gets Amazon Quick integration, Slack agents, Bedrock model choice, and zero-copy expansion; Google Cloud gets Hyperforce, Gemini Enterprise connectivity, and commerce checkout.
  • Why it matters: Salesforce multi-cloud means model choice (Bedrock, Gemini), infrastructure choice (Hyperforce on Google Cloud), and data choice (zero-copy across AWS and BigQuery) are all live decisions now, not roadmap items.
  • Who's affected: Any organization running Salesforce alongside AWS or Google Cloud commitments — especially teams planning AI agents, contact center voice, or cross-cloud data strategies.
  • Key dates: Most AWS integrations are available now, with voice and additional Slack agents landing Fall 2026. Hyperforce on Google Cloud reaches North America GA in November 2026, with regional expansion (including Germany) in 2027.
  • How Vantage Point helps: Our Salesforce implementation and advisory services help teams map these announcements to their existing cloud commitments and build a sequenced adoption plan.

What changed on the AWS side

The AWS expansion, announced September 16, 2026, covers four fronts — agents, models, data, and voice — and per Salesforce and AWS, most of it is available now.

Agents where employees already work

  • Salesforce context in Amazon Quick. Salesforce data, context, and skills are natively available inside Amazon Quick with no custom integration, via Salesforce's MCP-based headless architecture: pipeline health, account summaries, service cases, and operational analytics in a single brief. Available now.
  • AWS frontier agents in Slack. The AWS DevOps Agent works inside Slack now, with the AWS Security Agent, AWS FinOps Agent, and AWS Partner Central agents following in Fall 2026 — with full conversational context, so teams can investigate incidents or review findings without leaving the channel.
  • Agentforce model choice through Amazon Bedrock. Agentforce customers can now access Bedrock's selection of frontier models, including models from Anthropic and NVIDIA, within Salesforce agents and agentic workflows. OpenAI models — already generally available on Bedrock — are coming soon to Agentforce. Per the announcement, Amazon Bedrock enforces Zero Data Retention (no customer data stored or used for model training), Salesforce's Trust Layer adds dynamic grounding and toxicity detection, and the stack carries HIPAA, PCI, SOC 2, and ISO 42001 compliance with audit trails.

Data without migration

  • Data 360 zero-copy expansion. Zero copy now extends to AWS Glue-managed Apache Iceberg tables, S3-backed Apache Iceberg tables, Amazon Aurora, Amazon RDS, and SageMaker Lakehouse — building on existing Amazon Redshift and S3 support. Available now. The companies are also exploring bidirectional semantic-information sharing with the AWS Glue Data Catalog.
  • Informatica headless expansion. Informatica's MCP servers now support platform administration, catalog discovery, data-quality scores, and master-data retrieval, accessible from Amazon Bedrock AgentCore and Amazon Quick. Available now.

Voice and collaboration

  • Agentforce Voice with Amazon Connect Customer. Using Agent2Agent (A2A) support, the two voice platforms will communicate through real-time, bidirectional audio over WebSockets, letting AI agents coordinate across both contact center stacks. Arriving Fall 2026.
  • Dictation for Slack built on AWS Trainium. Salesforce AI Research optimized its proprietary speech models on AWS Trainium chips for ultra-low-latency voice dictation in Slack. Available now.

There is also a procurement angle: Salesforce Clouds are available in AWS Marketplace across 33 countries, and per the announcement, customers can apply eligible AWS spend commitments toward Salesforce products, including Agentforce.

What changed on the Google Cloud side

The Google Cloud expansion, announced September 15, 2026, goes further up the stack — into Salesforce's infrastructure itself.

Hyperforce on Google Cloud

Hyperforce is already handling live production customer traffic on Google Cloud, and Salesforce will begin migrating select U.S. customers in Q4 2026. North America general availability is set for November 2026, with phased feature and regional expansion in 2027, including Germany. Per Salesforce's Chief Trust and Infrastructure Officer, the trust bar — security, compliance, and resilience standards — stays the same; what changes is the infrastructure underneath.

Gemini Enterprise meets Salesforce's headless architecture

Google Cloud is one of the first hyperscalers to implement Salesforce's headless architecture, which exposes thousands of Salesforce capabilities to any model, agent, or interface over MCP. Per the announcement, enterprises can surface Salesforce data and capabilities directly inside Gemini Enterprise — pipeline health checks, account summaries, case triage — without custom integration. Tableau analytics connect through Tableau MCP, enforcing governance, row-level security, and semantic models on every agent query.

On the model side, the Agentforce Reasoning Engine with Gemini is GA now, and Google's latest models remain available for Prompt Builder. The Salesforce Federated Connector for Gemini is in private preview with GA planned for late October, and the Agentforce Sales Agent for Gemini Enterprise is in beta on Google Cloud Marketplace.

Commerce Cloud checkout on Google surfaces

Starting this Fall, shoppers can discover and buy products directly in Google Search (including AI Mode) and the Gemini app, powered by Commerce Cloud and the Universal Commerce Protocol (UCP). Merchants keep payments, compliance, and order management on their existing commerce infrastructure. The catalog feed integration with Google Merchant Center is GA now; UCP checkout reaches GA in October 2026.

Data 360 and BigQuery

Salesforce Data 360 and Google BigQuery already support zero-copy data sharing. The expansion adds more regions, adopts Iceberg REST Catalog open standards, and enables Private Connect — with expanded zero copy reaching GA in Fall 2026.

AWS vs. Google Cloud: what each expansion delivers

Layer AWS expansion Google Cloud expansion
AI models Bedrock model choice in Agentforce (Anthropic, NVIDIA now; OpenAI soon) Gemini in Agentforce Reasoning Engine (GA now) and Prompt Builder
Agents in work surfaces Salesforce context in Amazon Quick; AWS DevOps/Security/FinOps agents in Slack Salesforce capabilities in Gemini Enterprise; Agentforce Sales Agent beta on Google Cloud Marketplace
Data access Zero copy to Glue Iceberg, S3 Iceberg, Aurora, RDS, SageMaker Lakehouse (now) BigQuery zero-copy expansion, more regions, Iceberg REST Catalog, Private Connect (Fall 2026)
Infrastructure Salesforce Clouds purchasable in AWS Marketplace (33 countries) Hyperforce running Salesforce on Google Cloud (NA GA November 2026)
Voice / commerce Agentforce Voice ↔ Amazon Connect Customer via A2A (Fall 2026) Commerce Cloud + UCP checkout on Google Search and Gemini app (October 2026)

Who's affected

Three groups should treat this as a planning trigger, not a news item:

  1. Organizations with existing AWS or GCP spend commitments. If you hold an AWS Enterprise Discount Program commitment, Salesforce purchases through AWS Marketplace may now count toward it. If you are a Google Cloud shop, Hyperforce gives you a credible path to running Salesforce on your preferred infrastructure for the first time.
  2. Teams building agentic AI roadmaps. Model choice inside Agentforce is no longer theoretical. Per Salesforce, Agentforce now spans Bedrock's frontier model catalog and Google's Gemini models — which means your agent design process needs a model-selection step, with cost, latency, and compliance criteria attached.
  3. Contact center and commerce leaders. A2A voice interoperability between Agentforce Voice and Amazon Connect Customer, and UCP checkout on Google surfaces, both arrive in Fall 2026. If your 2027 roadmap touches either, architecture decisions made now will determine whether you can adopt them cleanly.

What businesses should do next

A practical sequence, based on what is available now versus what lands in the next two quarters:

  • Inventory your cloud commitments. Document AWS and GCP spend commitments, marketplace eligibility, and data-residency requirements before choosing where new Salesforce workloads live.
  • Map your data estate against zero-copy support. If your data sits in Redshift, S3, Aurora, RDS, Glue, SageMaker Lakehouse, or BigQuery, you may be able to ground AI agents in it without migration. A system integration and data migration assessment can confirm which sources qualify.
  • Define a model-selection policy for Agentforce. Decide how your team will choose between Bedrock-hosted models and Gemini models per use case — including cost, latency, data-handling, and compliance criteria — before agents multiply.
  • Pilot where GA is already real. Amazon Quick integration, Data 360 zero copy on AWS, Gemini in the Agentforce Reasoning Engine, and the Google catalog feed are all available now. Pick one low-risk workflow and measure it.
  • Plan governance before Fall 2026. Voice interoperability and UCP checkout both expand the surface area where AI agents act on customer data. An AI governance and compliance review is cheaper before go-live than after.
  • Watch the Hyperforce migration window. If you are a U.S. customer interested in Google Cloud infrastructure, select-customer migrations begin Q4 2026 — talk to your Salesforce account team early.

Timeline: what is available when

Capability Availability
Salesforce context in Amazon Quick Available now
AWS DevOps Agent in Slack Available now
AWS Security, FinOps, Partner Central agents in Slack Fall 2026
Agentforce model choice via Amazon Bedrock Available now (OpenAI models coming soon)
Data 360 zero-copy expansion on AWS Available now
Informatica headless MCP expansion Available now
Agentforce Voice ↔ Amazon Connect Customer (A2A) Fall 2026
Slack voice dictation on AWS Trainium Available now
Hyperforce on Google Cloud Select U.S. migrations Q4 2026; North America GA November 2026; Germany 2027
Salesforce Federated Connector for Gemini Private preview; GA planned late October 2026
Agentforce Sales Agent for Gemini Enterprise Beta now on Google Cloud Marketplace
Agentforce Reasoning Engine with Gemini GA now
Data 360–BigQuery zero-copy expansion GA Fall 2026
Google Merchant Center catalog feed GA now
UCP checkout on Google surfaces GA October 2026

How Vantage Point Helps

Multi-cloud announcements are easy to read and hard to operationalize. Vantage Point is a boutique, senior-led Salesforce and HubSpot consulting partner that helps organizations turn platform news into architecture decisions. We can help you:

If your team is evaluating what the AWS and Google Cloud expansions mean for your Salesforce roadmap, Vantage Point can assess the right next step and build a practical implementation plan. Contact us to schedule a readiness conversation.

FAQ

Is Salesforce moving off AWS onto Google Cloud?

No. Salesforce expanded both partnerships at the same time. Hyperforce on Google Cloud adds infrastructure choice — North America GA is November 2026 — while the AWS expansion deepened agent, model, data, and voice integrations, most of which are available now. Per Salesforce, this is about giving customers choice in models, data sources, and infrastructure, not switching clouds.

What is Salesforce's headless architecture?

Salesforce's headless architecture is an MCP-based layer that exposes thousands of Salesforce capabilities — data, workflows, and actions — to any model, agent, or interface without custom integration. Per the announcements, it is what lets Amazon Quick and Gemini Enterprise surface Salesforce business context natively.

Do I need to migrate my data to use zero copy with AWS or BigQuery?

No. Zero copy lets Salesforce Data 360 query data in place, without migration or duplication. On AWS, per Salesforce, zero copy now covers Glue-managed and S3-backed Apache Iceberg tables, Amazon Aurora, Amazon RDS, and SageMaker Lakehouse, in addition to Redshift and S3. On Google Cloud, the BigQuery zero-copy expansion reaches GA in Fall 2026 with more regions, Iceberg REST Catalog standards, and Private Connect.

Which AI models can Agentforce use now?

Per the announcements, Agentforce customers can access Amazon Bedrock's frontier model selection — including Anthropic and NVIDIA models, with OpenAI models coming soon — and Google's Gemini models, which are GA now in the Agentforce Reasoning Engine and available in Prompt Builder. Salesforce's Trust Layer and Bedrock's Zero Data Retention apply on the AWS side.

Can AWS spending commitments be applied to Salesforce purchases?

Yes, per Salesforce and AWS. Salesforce Clouds are available in AWS Marketplace across 33 countries, and customers can apply eligible AWS spend commitments toward Salesforce products, including Agentforce, while consolidating procurement and billing. Confirm eligibility with your AWS and Salesforce account teams, since terms are governed by customer agreements.

When can U.S. customers run Salesforce on Google Cloud infrastructure?

Per Salesforce, Hyperforce is already handling live production traffic on Google Cloud, select U.S. customer migrations begin in Q4 2026, and North America general availability is scheduled for November 2026. Regional expansion, including Germany, follows in 2027. Customers interested in migrating should engage their Salesforce account team before the Q4 window.

What should contact center leaders do about Agentforce Voice and Amazon Connect?

Treat Fall 2026 as the decision window. Per the announcement, Agentforce Voice and Amazon Connect Customer will interoperate through Agent2Agent (A2A) support with real-time, bidirectional audio over WebSockets. Contact center leaders should map routing, compliance recording, and escalation requirements now so they can evaluate the integration the moment it ships.