Every chip vendor is in a race to provide the best hardware to compete in the massively accelerating AI market. While hype may be a bit ahead of reality now regarding AI capabilities, the speed of AI advancement is closing that gap surprisingly rapidly. Chip vendors like AMD, Intel, Qualcomm and the current AI market leader, NVIDIA, are moving very rapidly to understand and embrace this opportunity.
What is quickly becoming evident is that just focusing on the hardware isn’t good enough. These vendors must also develop AI software competence which was demonstrated early on with NVIDIA’s massive two-decades-long effort to bring AI to the market at scale. Intel is significantly increasing its own software capabilities, and AMD, not wanting to be left behind, has been ramping up, as well.
AMD’s recent move to acquire Nod.ai is a case in point and should significantly and very rapidly advance AMD’s AI capabilities.
Let me explain.
NPUs and GPUs + Software
AMD is one of several vendors that will have a complete hardware set for AI shortly, including low-performance, highly efficient NPUs (Neural Processing Units) and high-performance, less efficient GPU solutions. Used in concert, they should provide the breadth of hardware needed to both run high-performance AI workloads as well as persistent low-performance edge AI workloads where energy efficiency is more important.
To do this effectively, they will need to both understand the AI engines intimately and better understand the AI software that will run on it, otherwise the solutions are likely to underperform hardware vendors who have developed a stronger AI software proficiency and are better able to tune their products to meet related demand.
If AMD doesn’t understand deeply the AI solutions and can’t provide the necessary performance advantages, the AI market will pass it by. But AMD isn’t a company that intends to let that happen, so it acquired the AI software-enabling company, Nod.ai.
This fits with AMD’s award-winning CEO Lisa Su’s IBM training. One of IBM’s historic advantages is co-developing hardware and software together. It’s still arguably the market leader in mature tested AI technology with its watson.x solution.
Nod.ai creates enabling software for future AI deployments. In other words, it makes AI run better, but doesn’t generally compete with AI vendors like OpenAI, allowing it to move between AI providers to create solutions that significantly enhance the performance of third-party AIs in a way that shouldn’t upset the AI providers.
Tying Nod.ai to AMD makes a lot of competitive sense. If an AI can be made to run better on a vendor’s hardware, those that buy that AI are likely to favor the hardware it runs best on, and Nod.ai will assure that AI software runs best on AMD.
This should provide a strong competitive advantage to AMD once the company is acquired and merged because not only will AMD have unparalleled access to Nod.ai’s tools, but the two companies will then work harder together to create a unique and powerful solution.
In addition, Nod.ai has been favoring AMD for some time in its solutions, suggesting that this greater integration resulting from the merger already exists due to Nod.ai seeming to favor AMD, providing a far faster path to market with a related AMD/Nod.ai blended solution.
In short, the two companies already seem to be working together well, suggesting the benefits to AMD for this merger should emerge unusually rapidly.
Vendors like AMD are rapidly adjusting to the AI opportunity, but those with enhancing software will likely have the strongest market advantage. Anticipating this, AMD has moved to acquire Nod.ai, an AI enabling software company that has the potential to give AMD a significant advantage in creating and providing advanced AI solutions.
This acquisition again shows that AMD is here to play and plans to aggressively go after and seize this new AI opportunity and make it theirs.