IBM Inks $240M Infrastructure Deal with AI-optimized Cloud Operator Together AI
AI Executive Summary
IBM has secured a $240 million infrastructure deal to deploy a large cluster of Nvidia HGX B300 systems, featuring Blackwell Ultra chips and ConnectX-8 SuperNICs, on IBM Cloud for Together AI by the first half of 2027.
This partnership follows Together AI's recent $800 million funding round and aims to scale its cloud platform, which currently processes 400 trillion token per month.
Together AI will leverage this 15-petaflop hardware to host production-grade inference workloads for open-source AI model.
Why It Matters
Strategic TakeawayThis partnership highlights a shifting cloud infrastructure ecosystem where legacy enterprise giants like IBM act as specialized hardware enablers for high-throughput AI startups, bypassing traditional hyperscaler dominance. By anchoring Together AI's 400-trillion-token monthly workload on next-generation Blackwell Ultra architecture, the deal establishes a blueprint for cost-optimized, decentralized open-source model inference at massive scale.
Multi-Vector Implications
- TECHNICALDeploying HGX B300 clusters with ConnectX-8 SuperNICs and BlueField-3 DPUs will drastically reduce latency and interconnect bottlenecks for multi-node open-source model inference.
- MARKETIBM's $240M deal positions it as a key alternative to AWS and Azure for AI startups seeking guaranteed access to scarce, high-demand Nvidia Blackwell Ultra silicon.
- GOVERNANCERelying on a 2027 hardware roadmap requires Together AI to implement strict capacity planning and SLA guarantees for enterprise clients during the multi-year transition.
Strategic Outlook
12-18M HorizonOver the next 12-18 months, expect a surge in secondary cloud providers and legacy IT giants securing long-term hardware reservation agreements to capture the booming open-source inference market. Together AI will likely transition its highest-throughput workloads to hybrid environments, balancing immediate AWS capacity with the incoming 2027 IBM Blackwell Ultra clusters to optimize token-delivery margins.
Referenced Coverage & Sources
Read the full coverage below for original reporting, technical benchmarks, and complete primary source details.
Vibe Coding Startup Lovable Doubles Valuation to $13.3B with $400M Raise
Lovable Labs Inc. said today it has raised $400 million in Series C funding at a $13.3 billion valuation. The Swedish artificial intelligence coding startup was worth $6.6 billion in December. The company sells a vibe coding service.
Nvidia Taps Wall Street for a Half-trillion Dollars to Fuel Global AI Infrastructure Buildout
Some of Wall Street's biggest financial firms are partnering with Nvidia Corp. to pour a half-trillion dollars of funding into the artificial intelligence industry's massive infrastructure buildout.
Former Apple Engineers' Physical AI Startup Lyte Raises $165M at $1.6B Valuation
Lyte, a physical AI startup building sensing and perception technology for robots, has raised a Maverick Silicon-led $165 million Series C funding at a $1.6.
NVIDIA AI Factory Compute Is Becoming an Investable Asset Class
We announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish independent financing platforms designed to.
Artificial Intelligence
Artificial Intelligence (AI) is a broad field of computer science dedicated to building systems capable of performing tasks that typically require human cognitive function, such as visual perception, speech recognition, decision-making, and translation.
Agentic AI
Agentic AI refers to artificial intelligence systems designed to act autonomously, make decisions, plan workflows, and execute tasks without constant human intervention. Unlike traditional models that only respond to queries, agentic systems use an agentic loop to perceive environments, reason over goals, use tools, and iterate to achieve outcomes.
Explore technical glossaries, weekly market briefings, and editorial research articles related to this story:
Get top 5 high-signal AI news, venture funding rounds, and research papers auto-routed to dedicated channels every 3 hours.