Photo Credit: Mikita Yo on Unsplash
IBM (NYSE: IBM) recently announced results for its third quarter that missed estimates. However, its Software revenue had good momentum driven by its AI business and a sound strategy. Its stock hit an all-time high this week despite the revenue miss.

IBM’s Financials
Revenues for the third quarter grew 1% to $14.97 billion, missing the Street’s estimates of $15.07 billion. Adjusted earnings of $2.30 per share were ahead of the analyst estimates of $2.23 per share.By segment, Software revenue grew 10% to $6.52 billion, ahead of the Street estimates of $6.37 billion driven by 14% growth in revenue from RedHat and 13% growth in automation revenue. Consulting revenue was down 0.5% to $5.15 billion. Infrastructure revenue was down 7% to $3.04 billion. Financing revenues fell 2.5% to $0.2 billion.IBM ended the quarter with $13.7 billion in cash and marketable securities, up $0.3 billion from the end of 2023. Debt, including IBM Financing debt of $10.4 billion, totaled $56.6 billion, flat year to date.IBM continues to expect over $12 billion in free cash flow for the year. It expects fourth quarter constant currency revenue growth to be 2%. 

IBM’s AI Platform Strategy
During the recent earnings call, IBM said that it continues to reposition its portfolio towards a higher growth, higher margin business that is well positioned to address client needs around hybrid cloud and artificial intelligence. It disclosed that its generative AI book of business stands at over $3 billion, up more than $1 billion quarter to quarter. The mix is roughly one-fifth software and four-fifth consulting signings. IBM’s AI strategy is a comprehensive platform play where Rhel.ai and OpenShift AI are the foundation of its enterprise AI platform. They combine open-source IBM Granite’s LLMs and InstructLab model alignment tools with full stack optimization, enterprise-grade security and support and model indemnification. On top of that, it has an enterprise AI middleware platform with watsonx and an embed strategy with its AI assistance infused through its software portfolio and those of its ecosystem partners. Earlier this year, IBM released code models with 8 billion to 34 billion parameters. This month, it updated its Granite models, making them approximately 90% more cost efficient than larger models. These models can be trained in weeks instead of months and are easier to fine-tune for specific tasks. Granite models are available on watsonx and Red Hat and are also integrated into offerings from partners like AWS, Salesforce, Qualcomm, and SAP. It alsoannounced new collaborations with NatWest, Telefonica, Samsung SDS, and Toyota Systems. I have recently done a series of interviews with AI investors where I discuss the relevance of smaller models for vertical AI or specific use cases that can also address the hallucination issue. IBM’s Granite models embrace this thought. Smaller language models are easier and faster to train and are great for Vertical AI. It is currently trading at $218.39 with a market capitalization of $200.7 billion. It touched an all time high of $237.37 this week and a 52-week low of $136.33 in October 2023.More By This Author:AI Unicorns: SambaNova’s Chip Competes With NvidiaAI Unicorns: Hugging Face’s Successful Pivot Away From Teenage Chatbot EngineAI Unicorns: Mistral AI Brings Stiff Competition To Other LLMs