How IBM is Using AI to Stay Competitive in the Tech Sector
Companies worldwide are increasingly adopting AI technologies to enhance productivity and streamline operations.
Yet amongst mass AI adoption, comes great competition.
This urgency has led to a surge in demand for enterprise-focused AI solutions, prompting major technology firms to develop and refine their offerings.
As businesses across various sectors seek to harness the power of AI, the competition among tech giants to provide accessible and effective AI tools has only intensified, shaping the future of corporate technology adoption on a global scale.
Amongst this competition, IBM, the multinational technology corporation known for its hardware, software and cloud computing services, has made a significant move in the enterprise AI market.
It has released the latest version of its AI models designed specifically for businesses, aiming to capitalise on the growing trend of enterprise AI adoption.
What does Granite 3.0 offer?
IBM is a long-standing player in the technology industry and amongst its new AI models, known as the Granite 3.0 family, it is remaining competitive in the constantly changing field of enterprise AI solutions.
Highlighting this competitive edge, Arvind Krishna, CEO of IBM says in an interview with CNBC about IBM’s moves to capitalise more on its AI efforts:
“AI is not meant to displace, it’s meant to augment labour. If we can make every enterprise process more productive, we will create a better quality of life for everyone.”
Granite 3.0 is a family of AI models specifically designed for business applications. These models are categorised into three main types:
- General Purpose/Language models: These are versatile AI systems capable of understanding and generating human-like text across various tasks.
- Guardrails & Safety models: These are designed to ensure AI systems operate within defined ethical and safety parameters, addressing concerns about AI misuse or unintended consequences.
- Mixture-of-Experts models: These are more complex AI systems that combine multiple specialised models to tackle diverse tasks efficiently.
Granite 3.0 will also be made open-source and this is a key feature.
By making these models freely available under the Apache 2.0 licence, IBM is adopting a different approach compared to some competitors who charge for access to their AI models.
This strategy could potentially accelerate AI adoption and innovation across various industries.
However, IBM's open-source approach does not mean the company is forgoing revenue opportunities in the AI space.
Reuters reports that IBM offers a paid tool called Watsonx, which assists businesses in running AI models within their own data centres after customisation.
This tool provides a means for IBM to generate income from its AI offerings while still maintaining an open-source model distribution.
The Granite models are released under the permissive Apache 2.0 licence, providing a unique combination of performance, flexibility and autonomy to enterprise clients and the broader community.
How does the open-source approach set IBM apart?
IBM claims that its Granite 3.0 models can match or outperform similarly sized models from other leading providers on many industry benchmarks.
The company has trained these models on an extensive dataset, including 12 trillion "tokens" (units of text or code) from 12 natural languages and 116 programming languages.
To make these models more accessible, IBM is offering them through various platforms.
The models will be available on IBM's own Watsonx platform, as well as through partnerships with major cloud providers like Amazon Web Services (AWS) and Google Cloud.
This wide availability aims to give businesses flexibility in how they implement and use these AI models.
Additionally, a selection of these models will be accessible through Nvidia's software tools, which enable businesses to incorporate AI models into their operations.
Nvidia, a leading manufacturer of graphics processing units (GPUs) and other computer hardware, has become a key player in the AI industry due to the computational power its products provide for AI model training and deployment.
According to Reuters, Dario Gil, IBM's director of research, says that the new Granite models were trained using AI chip leader Nvidia's H100 graphics processor units (GPUs).
These advanced processors are specifically designed to handle the complex computations required for AI model training and inference.
Kate Soule, Program Director, Data and Model Factory at IBM and Dave Bergmann, Data and AI writer, summarise in IBM’s AI in Action Report:
“Reflecting our focus on the balance between powerful and practical, the new IBM Granite 3.0 models deliver state-of-the-art performance relative to model size while maximising safety, speed and cost-efficiency for enterprise use cases.”
******
Make sure you check out the latest edition of Technology Magazine and also sign up to our global conference series - Tech & AI LIVE 2024
******
Technology Magazine is a BizClik brand