Issue #1: Why Enterprises Are Lukewarm on AI Model Providers

-Hey High Stakers,-

Good morning and welcome to the 3rd issue of High Stakes!

Walk into the Canary Wharf Underground station, and you'll see massive digital ads, such as from Cohere, pushing enterprise AI solutions. OpenAI, Anthropic, and others are making similar plays — desperately trying to carve out their space. 

Their challenge? They're getting boxed out by the real winners: Hyperscalers and IT Services firms that own the enterprise IT stack.

Enterprise buyers don't want raw AI models. They want secure, compliant, integrated AI that fits seamlessly into their existing workflows. 

And hyperscalers (Microsoft, AWS, Google) and IT service firms (Accenture, Deloitte, TCS) are delivering just that — but at scale.

📊The Real Winners in Enterprise AI

The enterprise AI market is projected to hit $58.11 billion by 2025, but the biggest beneficiaries aren't model providers.

They are the companies embedding AI into enterprise IT:

• Microsoft dominates enterprise AI, capturing 30% of market share, followed by AWS (25%) and Google (20%).

• 65% of AI infrastructure spending goes to hyperscalers, bundling AI with cloud, security, and compliance (IDC).

• 64% of enterprises prefer AI solutions from existing vendors rather than standalone AI providers (McKinsey).

• Microsoft's Copilot has 4M business users, while OpenAI — despite 400M+ ChatGPT users — has just 2M enterprise customers.

The lesson? 

AI models’ superiority alone isn't enough to win enterprise AI. 

The companies that integrate AI seamlessly into existing enterprise stacks are the ones controlling the future.

🔍The Enterprise AI Game: A Seller’s Perspective

Imagine enterprise AI as a high-stakes football match. Who's really controlling the game?

• Enterprise IT (The Stadium Owners) → They set the rules and control access.

• Hyperscalers (Microsoft, AWS, Google) → They own the infrastructure and dictate pricing models.

• IT Service Firms (Accenture, TCS, Deloitte) → They coach the players, embedding AI into enterprise workflows.

• B2B SaaS Firms & Internal IT Teams → They adjust their plays based on market shifts.

• AI Model Providers (OpenAI, Anthropic, Cohere) → Despite superior technology, they struggle to find product-market fit in enterprise AI.

And here's the kicker: Enterprises don't buy AI models the way consumers do.

❌Why AI Model Providers Are Losing the Enterprise AI Battle

1. Enterprise buyers care less about AI benchmarks. 

B2C AI thrives on cutting-edge models. But enterprise buyers prioritize business impact over raw technical superiority.

• A 2% gain on MMLU or beating GPT-4o in reasoning tasks? Ok.

• The real question: Will this AI drive revenue, reduce costs, or make us more competitive?

2. Hyperscalers and IT Services firms are embedding AI into everything.

• AWS, Azure and Google Cloud bundle AI with cloud, security and compliance, making it an easy ‘yes’ for enterprises.

• Accenture, Deloitte and TCS don’t sell AI - they sell AI-powered business transformation.

Model providers sell AI experiments. Hyperscalers & IT Services leaders sell solutions with proven enterprise ROI tracking.

3. Enterprises prefer to build, not buy.

• Buying an external model means integration headaches, compliance risks, and security concerns.

• Many enterprises would rather fine-tune open-source models than rely on external vendors with unclear enterprise experience.

🚀How AI Model Providers Can Still Win

1. Own the Vertical AI Market 

General LLMs are great for consumer apps — but enterprises need domain-specific AI. Winning examples:

• Financial AI → 3x faster fraud detection with compliance-first models.

• Healthcare AI → HIPAA-native AI reducing diagnostic review time.

• Manufacturing AI → AI copilots optimizing supply chains.

The Play: Stop selling 'smart AI.' Start selling industry-specific AI that reduces risk and drives revenue.

2. Move from Automation to Revenue Intelligence 

Enterprises don't want AI that just automates — they want AI that makes them money. Winning examples:

• AI-driven sales acceleration → 20-30% shorter sales cycles.

• AI-powered procurement → Millions saved in supplier costs.

• AI in finance → Automated upsell/cross-sell predictions.

The Play: Tie AI directly to revenue impact. If it doesn't move the needle, enterprises won't buy.

3. Make Deployment Frictionless with AI Operating Systems 

Enterprises don't need just a model — they need AI integrated into their workflows. Security, compliance, latency, governance — these are non-negotiable.

Winning examples:

• Microsoft & Oracle: Shifting from model marketplaces to AI operating systems.

• NVIDIA's AIDE: Integrates AI into healthcare workflows, ensuring compliance with DICOM and HL7 standards.

• IBM's LinuxONE: Provides secure and scalable deployment of AI workloads in enterprise environments.

• Netflix's Use of AWS: Demonstrates how AI can automate content delivery and recommendation systems, enhancing user experience and optimizing infrastructure costs.

The Play: Sell AI ecosystems that plug into enterprise workflows seamlessly.

🤔What This Means for Different Roles

🔹 If you're evaluating AI vendors → You can look at 3 key aspects here.

1) Prioritize embedded AI solutions over standalone tools. Hyperscaler-native integrations matter - Microsoft’s Copilot already has 4 million business users because it’s embedded within Teams and Outlook.

2) Compliance is non-negotiable, so demand SOC 2 and ISO 27001 certifications, along with indemnification clauses for AI errors.

3) Flexibility in build-vs-buy decisions is key, as 64% of enterprises prefer fine-tuning open-source models. Ensure vendors support this without punitive licensing.

🔹 If you're an IT Services Leader → Hyperscaler partnerships are key. As AI embeds deeper into enterprise IT, services firms that integrate AI solutions will win the largest deals.

🔹 If you're a SaaS Founder → Find your moat. Don't compete on models! Build domain-specific, AI-powered SaaS that hyperscalers can't replicate.

🔹 If you're a GTM Leader → Align AI with procurement psychology. Selling AI means selling transformation, not technology. Understanding how hyperscalers influence procurement cycles will be crucial.

🔹 If you're shaping AI strategy (Alliances? Architect? Consultant?) → I guess it depends on where you as an influencer are sitting and who you are serving, but broadly three recommendations come to mind.

1) Accenture's $3B AI investment isn't R&D - it's pre-built industry workflows. Recommend partnering with such firms. Always be on the scout for new partnership opportunities. Better still? Work on and deploy such workflows yourself.

2) Negotiate cloud contracts with AI credits (AWS's $100M startup fund model). Bake AI into procurement!

3) Are you training sales teams on selling past the hyperscalers - e.g., "Our healthcare AI augments Azure's generic tools with pre-built prior-auth workflows." This sales enablement can be one of your core missions.

💬Final Thought: Can AI Model Providers Still Win?

Yes — but only if they stop acting like research labs and start competing like enterprise software companies.

Best,
Srini

P.S. If this sparked ideas about your AI strategy, share it with your team. These insights could reshape your 2025 roadmap.

👉 Who's best positioned to dominate enterprise AI? Reply and let me know.