Tech Tonic | We must not ignore smart acquisitions as a route to AI domination

Even as you and I begin to integrate even more artificial intelligence tools in our daily lives, having progressed from testing the waters over the past couple of years, we have likely missed a trend that was staring at us all along. Tech companies understood, quite some time ago, that smart acquisitions are the way forward, to achieve a level of dominance in this space.

This method, more often than not, is proving to be the single shot for accessing technology, bringing expertise and researchers within the fold and welcoming an already finely curated user base into the extended family. The question is, what happens when consolidation is complete, and there are no worthwhile acquisitions to be done? I would hazard a guess here. No one wants to think that far ahead.
For now, this method is delivering success. A few days ago, chipmaker AMD confirmed a definitive agreement to acquire AI infrastructure company ZT Systems. This gives AMD an enhancement from the latter’s compute, storage, and accelerator solution expertise, its customers including Microsoft Azure and Amazon Web Services. AMD CEO Lisa Su summarises this as part of a long-term AI strategy, with cloud and enterprise customers in focus. No wonder she is convinced this will help the company sell more AI chips. Surely AWS and Microsoft will become bigger customers.
Guess who they are up against, and who they want to compete with? Nvidia’s server and enterprise business. (Many may not have realised Su and Nvidia CEO Jensen Huang, are relatives). On Aug 20, the next trading day after the announcement, AMD stock peaked at US$ 159.96 – Wall Street analysts Wedbush have kept a target price of $200 for this stock. Nvidia themselves have been very active this year, with four AI-focused acquisitions already done – an Israeli startup Run:ai which makes GPU orchestration software, Deci which is another Israeli start-up that has expertise in Neural Architecture Search (NAS), automation startup Shoreline and AI cloud company Brev.
Really early movers were Alphabet, with the DeepMind acquisition as far back as 2014. That has paid off well. Apple has been acquiring AI-focused companies for many years now, and surely that’d have a bearing on the results we see with Apple Intelligence that rolls out later next month, alongside the new iPhone line-up. PrimeSense (2013), Siri (2010), Drive.ai (2010) and Spektral (2017) are some of their most notable business moves with the spectre of AI.
The question therefore is, why are acquisitions happening at such a pace?
With scale and generational fallbacks, the costs of building algorithms and large language models will be reduced. That may not be as big a criterion in the next few years, as it perhaps was a couple of years ago. Therefore, the idea is to find value with engineering talent as well as proprietary technologies, datasets and flexibility to deploy an AI system that’d best fit a larger cause.
Late last month, Canva announced the acquisition of Leonardo.AI in what could prove to be their most important one yet. And Canva has done quite a few in the past couple of years, which has provided the all-in-one creative platform with a momentum that’d hold it in good stead against Adobe, Microsoft, Google and pretty much any competing generative AI platform. Joining Canva’s existing 190 million strong user base will be 19 million enterprise users who were already using the Leonardo.AI platform.
Even more important is access to the Phoenix Foundation. No surprise then that the company promises more investment into the foundational model research, as well as rapidly integrating Phoenix into the Magic Studio suite. Acquiring Affinity meant Canva had professional photo editing, illustration, graphic design and page layout tools to offer to users, as direct competition to Adobe. Ticking off the proprietary technology and engineering talent aspects.
Flourish (2022), Kaleido (2021), Smartmockups (2021), Pexels (2019) and Pixabay (2019) too are already part of the Canva family.
Adobe has the acquisition of Bengaluru-based Rephrase AI as an important piece that perfectly fits that puzzle. Rephrase’s natural language processing which decodes text inputs to generate professional-quality video content, will be worth its weight in gold for Adobe’s extensive, and powerful suite.
Microsoft’s method has delivered success too at a time when competition is only steadily increasing. Their partnership with OpenAI needs little introduction and provides the foundation for the Copilot platform, which is no longer just a chatbot. The scope is wide, including integration within subscription services for consumers and enterprises, as well as a foundation for Windows PCs’ biggest step forward in years, the Copilot+ PC.
OpenAI isn’t their only partnership. French AI company Mistral’s multi-year collaboration which began earlier this year, will bring Mistral’s open and commercial language models to Microsoft’s Azure AI platform. That’s $2 billion sent towards Mistral, following a reported $10 billion investment in OpenAI. For Mistral, this could open up access to more markets globally, than may have perhaps been able to unlock on its own. They also get access to Microsoft’s supercomputing infrastructure.
The key takeaway, in terms of revenue – Microsoft and Mistral AI will make the latter’s higher spec models available to customers through the Models as a Service (MaaS) proposition, in the Azure AI Studio and Azure Machine Learning model catalogue. This method may provide a template for when the question I asked earlier, comes true – when there are no worthwhile AI acquisitions to be made.
Vishal Mathur is the technology editor for the Hindustan Times. Tech Tonic is a weekly column that looks at the impact of personal technology on the way we live, and vice-versa. The views expressed are personal.