top of page

GenAI, Google Cloud, Disruption, and the Global South

Without a doubt, generative AI (GenAI) is the biggest buzzword of 2023! Indeed, it seems difficult to have a conversation, even about boring, stodgy old telecoms without GenAI and its impact creeping into the conversation. Reams of copy and opinion has been produced on GenAI, with a specific focus on the tremendous disruption the technology will wreak across industry verticals. This has brought about significant debate around the degree of disruption to be expected but also about whether AI should be regulated. Disruption is already visible, with significant rounds of layoffs and cuts being attributed to the expected efficiencies to be garnered from GenAI investments. Not to be partial to telecom but Vodafone announced a new strategic plan that included 11,000 job cuts back in May. Many will argue that it’s a telecoms operator and the industry is ripe for disruption. But that is exactly the point. There are many such companies and industry verticals that are ripe for disruption. Regardless, it is still early days for GenAI. The universe of use cases that will be enabled by GenAI will continue to expand with several not even imagined as yet.

The other question that has not seen as much debate is the impact of GenAI in the so-called Third World, or perhaps we should use the new moniker, the Global South. China has been galvanized since the high-profile launch of ChatGPT last year, as argued in this article by Andy Mukherjee of Bloomberg, with the country’s heavyweights like Baidu, Alibaba, ByteDance and a host of others pivoting towards large language models (LLMs) and investing heavily in AI. China, though, is not a member of the Global South. The latter, somewhat nebulous grouping, includes several growth economies. The biggest of these is India, along with Indonesia, Brazil, and others. The Global South possesses neither the compute-intensive digital infrastructure nor the technology to create and deploy foundational LLMs. Rather, the best-case scenario is they will be restricted to building applications on top of AI platforms and foundational LLMs offered by the leading tech giants like Google and Microsoft.

The biggest of the Global South countries, India, would be a leading candidate to join the AI arms race. However, as Mukherjee argues correctly, despite its significant technology assets, Indian software and tech companies seem to be playing defense. Infosys is a good example of this trend with their new Topaz offering, which claims to offer enterprises “more than 150 pre-trained AI models running across more than 10 platforms with more to come.” Swiss army knife approach! What does this mean for enterprises in India and the Global South? Most have a few sunshine sectors that are strong digital adopters but nearly all have a combination of traditional industries with outdated practices and a really long tail of small and medium enterprises (SMEs). So will AI help these economies leap into a digital future, or will it cause a slow haemorrhaging of jobs as enterprises wise up to the potential of AI to drive “efficiencies” in their business?

There is another way to look at the onset of AI amongst the Global South countries. That is, do they have the technical capacity to build meaningful applications on top of foundational tech coming from the likes of the US? At least in India, the news would appear to be encouraging on this front. We recently had the opportunity to attend a Google I/O developer event in Bengaluru in the southern Indian state of Karnataka. Many will know Bengaluru by it’s previous name, Bangalore, as well as the fact that it is known as India’s Silicon Valley. While comparisons to Silicon Valley are overblown, Bengaluru has India’s highest pool of developer talent. No wonder Google would host a developer event here, where we witnessed hordes of interested developers getting turned away from workshops due to the event being over capacity. The event also featured a keynote by the Google Cloud CTO, Will Grannis, and amongst several interesting themes, he cited India as having 60 AI startups already at this stage. It is impossible at this early stage to posit how many of them will survive or thrive but even assuming that a significant number of them use foundational technology from the liked of Google Cloud, that would imply a fast start to adoption and awareness.

For Google Cloud, India is a priority market, and their cloud infrastructure has already seen significant traction with customers. However, Google is making several moves to make AI accessible to the developer community in India. To do this, Google Cloud is investing in several areas for AI. The biggest of these is the Cloud platform itself, with new products like Vertex AI combined with foundational models like Codey, a family of code models built on top of PaLM 2 (Google’s LLM) and Chirp, which does text to speech conversion. The approach seems to be offer AI foundational models like Codey and Chirp with a strong compute layer on top. Or, as the Google Cloud folks seem to be calling it, “AI-optimized infrastructure”. Overall, Google’s macro approach to AI seems to cut across the entire tech stack, with chips, LLMs, tools, frameworks and more. Ultimately, the strategy seems focused on giving away all this tech, with a subtle reminder to use Google Cloud’s “AI-optimized infrastructure”. So, AI tech + Compute layer = AI Optimized infrastructure. We shall be watching this space closely to see how Indian enterprises adopt GenAI, as well as examples from the broader Global South economies.

6 views0 comments

Recent Posts

See All


bottom of page