Karabo Tema
AI Investment and Dominance
The progress achieved by OpenAI’s ChatGPT has revolutionised Artificial Intelligence (AI) and the technological landscape re-establishing a tech market cap dominance that seemed to be fading. Since its launch in November of 2022, ChatGPT has caused a burst of activity introducing large investments in AI. At the end of September 2024, AI funding rounds had reached over 3,680. A large portion of the capital raised was seed funding indicative of an AI ecosystem still in its nascent state as many startups try to enter AI-mania.
AI native companies have benefited by taking advantage of depressed investment themes and headwinds in Fintech, Metaverse and Blockchain. Dealmakers and investors have shown keen interest in companies focused on the fundamentals of AI. Thus, investment has been mostly in hardware and foundational segments of AI. This is predominantly attributed to the cost of training AI models and the capital required to perfect and produce highly sophisticated models.
In 2024, the median pre-money Series A valuation was US$34m, Series B was US$150m and Series C was US$558m. AI is capital intense and will continue to attract strategic industry investors, Venture Capital companies, and Private Equity firms with deep pockets. Banks will show greater appetite for sizeable equity and debt as AI players begin to disrupt entire sectors of the economy and improve financial inclusion. In the long run it will be of greater benefit to invest in businesses that are able to develop and execute sound business models for AI rather than the AI companies themselves.
Microsoft through OpenAI has enjoyed big leaps in becoming the global AI leader beating Google which declared itself to be an “AI-first” company in 2016. AI development is ultra-competitive, and the next phase of competition will be the rapid build of supercomputers optimised for AI. Colossus, a supercomputer by xAI currently has a 100K H100 training cluster making it the most powerful AI training system globally. The power comes from CPUs and GPUs that serve as accelerator in training large language models. Securing AI chips will be key in determining winners and losers. However, the progression from AI training models to inference or deployment models could be the next frontier in the cost efficiency battle.
OpenAI has raised US$6.6bn in a recent capital raise improving their valuation to US$157bn. xAI, Musk’s AI startup has managed to raise $6bn in series B funding. Anthropic has received investments worth $6bn from Amazon and Google intensifying the battle for AI supremacy.
According to a report published by PWC, McKinsey, Economic Potential of Generative AI and Genesis Analytics the current AI market is valued at US$16.5trillion with Africa representing 2.5% of the global market equivalent to US$0.4trillion.
Africa’s struggle with data democratisation has slowed down the continent’s capital attraction. Notwithstanding the challenge of digital infrastructure that can handle hefty AI workloads. Over 60% of AI startups are early stage and unable to attract financial investors with short to medium turn return requirements. Private capital currently does not have incentives to invest in African AI startups, as most African governments have been slow out the blocks with AI strategies and regulation creating uncertainty for investors. African countries have a large backlog in physical infrastructure and will have to find budget within their fiscus to prioritise digital infrastructure and data democratisation. Barriers to primary research data should be removed giving African researchers access to the significant amount of data on Africa held outside the continent.
Threat to Revenue
The commercial potential of AI has caused a fierce contest for AI dominance, surpassing that of the first moon landing. Microsoft, Google, Meta and Hauwei are all fighting a relentless battle to become the market leader. Google is concerned about OpenAI replacing the search function and wiping out their $160bn revenue gold mine. Meta has an equally difficult situation that threatens their advertising revenue accounting for $27.3bn (70%) of Meta’s total revenue. Online advertisers are concerned about the growing number of AI generated bots on Meta platforms such as Facebook.
Geopolitical contestation and Regulation
AI has become the latest technological prize in the geopolitical battle between China and the US. Washington instructed Nvidia, to cease the sale of advanced semiconductors to China in efforts to ensure that Beijing has inferior chips when training their AI models.
Opaque and fragmented regulation is under pressure to enhance and consolidate into meaningful legislation. Valuation multiples have been high, however proposed regulation will reduce valuations as regulators provide investors with clearer investment perimeters.
The global economy is in the early-stage process of regulating and investing in AI. Regulation will be hit and miss as disputes are raised. The importance of regulation is paramount to the safety of humans however a delicate balance is required to avoid stifling innovation which will impact multiple tech valuations and industries.
AI investment has become a formidable variable in modern economics. Unlocking venture funds and improving the circa 1% AI investment in Africa requires clear strategies on AI infrastructure and supportive policies. Majority of AIs value has been created and realised in the U.S with an investment of US$249bn, China with US$95bn, Europe and Israel. Africa is on the back foot in global AI with datasets not fully representing the continent in current training models. These challenges will not curtail Africa’s AI prospects as it boasts a young population and opportunities to create bespoke solutions. Digital infrastructure complimented by energy infrastructure is paramount to the success of AI startup growth in Africa.
Karabo Tema, Associate for Telecoms, Media & Technology at Standard Bank.He writes in his personal capacity.
