Platform Vendors & Services

The hyperscalers, model providers, and enterprise platforms powering AI consulting

Competitive Landscape

Dimension	Leaders	Strategy
Model quality	Microsoft/OpenAI, Google, Anthropic	Frontier model R&D as moat
Enterprise integration	Microsoft, Salesforce, SAP, Oracle	AI embedded in existing tools
Data proximity	Snowflake, Databricks, Google	"AI where data lives"
Multi-model flexibility	AWS Bedrock, Google Vertex	Model marketplace / choice
Governance	IBM, Microsoft, AWS	Regulated industry certifications
Cost/infrastructure	AWS, Oracle, Google	Custom silicon, price competition

1. Microsoft

Azure AI + Copilot Ecosystem

Dominant enterprise AI vendor • 90%+ of Fortune 500 on Azure

Key Offerings

Azure OpenAI Service: Enterprise GPT-4/4o/o1 with Azure security, compliance, networking
Microsoft Copilot: Across M365, Dynamics 365, Power Platform, GitHub, Security
Copilot Studio: Custom agent creation ($200/month per 25K messages)
Azure AI Studio: Unified GenAI app building (prompt flow, model catalog, RAG)

Pricing

GPT-4o: ~$2.50/$10 per 1M input/output tokens
Copilot for M365: $30/user/month

Weakness

Platform complexity and high Copilot per-seat costs leading to slower-than-expected adoption. Copilot fatigue is real.

2. Google Cloud

Vertex AI + Gemini

Best-in-class context window (1M+ tokens) • Competitive Flash pricing

Key Offerings

Vertex AI: End-to-end ML/AI platform with Gemini + third-party models (Claude, Llama, Mistral)
Gemini for Workspace: AI across Gmail, Docs, Sheets, Meet ($30/user/month)
BigQuery ML: ML directly in the data warehouse
NotebookLM: AI research assistant

Pricing

Gemini 1.5 Pro: ~$1.25/$5 per 1M tokens
Gemini 1.5 Flash: ~$0.075/$0.30 per 1M tokens (very competitive)

Weakness

Smaller enterprise installed base than Microsoft/AWS. Reputation for killing products creates hesitancy.

3. AWS

Bedrock + SageMaker + Amazon Q

Largest cloud partner network • IaaS market leader • Custom AI silicon

Key Offerings

Amazon Bedrock: Managed foundation models (Claude, Llama, Mistral, Titan). Agents, Knowledge Bases, Guardrails.
Amazon Q: AI assistant for business ($20-25/user/month) and developers ($19/user/month)
Generative AI Innovation Center: Free advisory for qualified customers
Trainium/Inferentia: Custom AI chips for cost advantages

Weakness

Own first-party models (Titan) lag competitors. Perceived as infrastructure-focused, not application-focused.

4. OpenAI

ChatGPT Enterprise + API Platform

Best consumer AI product • Strongest developer mindshare • No consulting arm

Key Offerings

ChatGPT Enterprise/Team: SSO, SCIM, admin controls, unlimited GPT-4
API: GPT-4o, o1/o1-mini (reasoning), Assistants API, fine-tuning, batch

Consulting Strategy

No traditional consulting arm. Instead partners with: Bain (strategic), PwC (first reseller), Deloitte/KPMG/EY (implementation). The Frontier Alliance (Feb 2026) formalized this with McKinsey, BCG, Accenture, Capgemini.

Pricing

ChatGPT Enterprise: ~$60-70/user/month (custom)
GPT-4o API: ~$2.50/$10 per 1M tokens
o1 API: ~$15/$60 per 1M tokens

5. Anthropic

Claude Enterprise + MCP Protocol

Best for code, analysis, safety-critical apps • Dual-cloud distribution (AWS + GCP)

Key Offerings

Claude API: Claude 3.5 Sonnet/Haiku, Claude 3 Opus. 200K context, tool use, vision.
Claude Code: CLI-based AI coding assistant
MCP (Model Context Protocol): Open protocol for connecting to external data/tools

Consulting Strategy

Small Solutions team. Distributed through AWS Bedrock and Google Vertex AI. Key partners: Accenture (major strategic), BCG (advisory). Claude Partner Network (Mar 2026) with $100M commitment.

Pricing

Claude 3.5 Sonnet: ~$3/$15 per 1M tokens
Claude 3.5 Haiku: ~$0.25/$1.25 per 1M tokens

6. Salesforce

Einstein AI + Agentforce

Deep CRM data integration • "AI for revenue teams" niche

Key Offerings

Einstein 1 Platform: AI layer across Salesforce (Copilot, GPT, Trust Layer)
Agentforce: Autonomous AI agents for service, sales, marketing (major 2025 push)
Data Cloud: Real-time data platform with vector DB capabilities

Pivot: Moving from copilot to autonomous agents. Per-conversation pricing model for Agentforce.

7. IBM

watsonx Platform + IBM Consulting

$6B AI book of business since watsonx launch • ~160K consultants • Regulated industries focus

Key Offerings

watsonx.ai: Train/tune/deploy foundation models (Granite + third-party)
watsonx.governance: AI governance, risk, compliance (unique differentiator)
watsonx Code Assistant: For Ansible and mainframe modernization

Differentiator: Integrated consulting + platform play for banking, healthcare, government. Granite models positioned as efficient and transparent (not frontier).

8-10. Oracle, SAP, Snowflake

Vendor	AI Strategy	Key Product	Unique Angle
Oracle	Embed AI in existing Oracle apps	Database 23ai (vector search), OCI GenAI	OCI 30-50% cheaper; partnerships with OpenAI, xAI for training
SAP	Make business processes smarter	Joule (GenAI copilot across SAP apps)	"AI that understands your business" via SAP data grounding
Snowflake	AI where the data already lives	Cortex AI (SQL-callable LLMs), Cortex Analyst (NL to SQL)	Democratizes AI to SQL analysts; Arctic open-source LLM

Key Dynamics in Early 2026

Copilot fatigue. Per-seat costs of $20-50/user/month are being questioned when usage is uneven. Vendors pivoting toward usage-based or outcome-based pricing.

Agent era. The industry is shifting from "copilot" (human-in-the-loop) to "agent" (autonomous). Salesforce Agentforce, Microsoft Copilot Studio agents, and custom frameworks on Bedrock/Vertex are the new battleground.

Data gravity wins. Enterprises deploy AI where their data already resides. This benefits Snowflake, Databricks, and hyperscalers with existing workloads.

Enterprise Budget Allocation (Typical)

Category	Share
Cloud infrastructure & platform (Azure, AWS, GCP)	40-50%
Consulting & professional services	20-30%
Software licenses (Copilot seats, SaaS AI)	15-20%
Internal AI team headcount & training	5-10%

← AI-Native Firms Freelance & Independent →