The GAME dynamic: High fixed-cost firms to gain most from AI adoption

The golden age of margin expansion (GAME) was a theory most publicly expounded by Brad Gerstner of Altimeter Capital in various tweets and podcasts. Mr Gerstner is best known for his open letter to Meta in October 2022, near the stock’s bottom, urging the company to become fit, cut costs, and refocus on its core business. Having followed much of this advice, Meta has since become one of the best-performing technology stocks of the past three years. 

A big believer in artificial intelligence (AI) and an investor in OpenAI, Mr Gerstner believes that now the time has come for the largest tech companies specifically, and all companies more broadly, to use AI tools to increase productivity. He thinks the productivity release afforded through the effective use of AI in business processes will lead to substantial margin expansion as it will allow companies to lower headcount while still growing revenues robustly.  

AI has already proven its value in customer support and programming, and its applications will only broaden as the technology matures and enterprises grow more comfortable with the tools. Every day, we see new applications and further proof of concept, from healthcare to defence and everything in between. The costs of AI adoption and use are also dropping rapidly. Large language models are on the path to commoditisation, with at least six models of similar capability, half of them open source. 

As evidence of this productivity and margin dynamic, Mr Gerstner points out how the Magnificent-7 technology giants have dramatically slowed headcount growth. He shows a chart indicating that between 2011 and 2021, these companies were growing headcount at about 25 per cent per annum. This peaked in 2020 at 45 per cent growth, when Covid supposedly changed consumer behaviour forever. In the last three years, however, this same metric has slowed to just about 2 per cent growth. 

However, these last three years, the Magnificent-7 companies are still growing their revenues by 15–20 per cent, despite almost no headcount expansion, hence the scope for margin expansion. Just for context, in their latest quarter, Alphabet grew revenues by 14 per cent, Meta by 22 per cent and Microsoft by 18 per cent. For operating margins, Alphabet at 32.5 per cent, Meta at 43 per cent and Microsoft at 45 per cent all reported an increase. They have been able to increase operating margins despite raising capex guidance to $75–80 billion each and surging research and development (R&D) costs. The margin increase has come from lowering the selling, general and administrative (SG&A) expenses as a percentage of revenue. 

Corporate boardrooms are seized of this dynamic and most surveys show that top management believes their companies can see a productivity release of at least 20 per cent from the use of AI. Most believe they will hire fewer workers. 

What is the implication of this potential headcount reduction? It is now clear that every company will have to adopt and embrace AI tools in order to compete. They will have to become more productive, otherwise they will fail and close. The winners will be those companies that have the pricing power and stature to not need to fully share these cost savings with their customers.  

If you can keep some of these cost savings and show the margin expansion, you will see your valuations improve and strong investor following. Only a few companies have this pricing power. Most companies, however, will be forced to compete away these cost savings through price reductions. If an entire industry is becoming more productive, it is by no means certain that companies will be able to keep any of the cost savings. AI will become in all likelihood table stakes, just to be in the game.

It is likely that we will see deflation across many industries and products. Customers will benefit and inflation should trend down. You need to bet on those companies that have price elasticity so that falling prices trigger volumes to compensate and deliver top line growth. Markets will disproportionately value volume growth, as most will believe that in the age of AI-driven deflation, pricing as a lever will be available to only a very small minority with strong customer lock-in. 

A good investment strategy may be to target companies with high fixed overheads, or that have a large SG&A budget. It might actually make sense to target companies with large existing headcount. They have the most to gain. If they have market share dominance and pricing power, they will be able to show strong margin expansion over the coming years as they adopt AI tools in their business processes and become fit. 

It is also a fact that if you are an incumbent fighting an AI-native startup, how do you compete? The startup will have a cost structure fundamentally different from the incumbent. Their process architecture will also be different. As an example, Alphabet has to compete with OpenAI. They are the two dominant players in LLMs with Gemini and ChatGPT. The fact is that Alphabet has 180,000 people, OpenAI has about 7,000. Which organisation will be more nimble and have lower overheads? If a startup is able to use AI tools from day one in how they build the business and structure their staff, they could be at a structural cost advantage. 

This dynamic will also have serious implications for younger generations just entering the workforce. In most developed countries, for the first time, the unemployment rate for fresh college graduates is higher than the population average. Entry-level roles are where AI is taking root initially. This will only heighten societal tensions between the haves and have-nots. Most fresh graduates cannot find jobs, while a small cohort of AI talent can name their price. 

All this talk of productivity release through the use of AI also brings to the fore the outlook for the Indian IT services industry. It is now a well-established fact that AI has a clear use case in coding. At a minimum, it improves the productivity of programmers by 25–30 per cent. Enterprise customers will ask for this 25 per cent productivity gain to be passed on to them. If customers are going to shed staff, will they have 2,000-person IT service project engagements? While I believe that our IT service giants are adaptive and will find ways to build new service lines around the use and adoption of AI, we may have an air pocket of a few years where the new service lines are not large enough to compensate for the deflation in the existing business. 

Every company will have to use AI tools and fundamentally rejig their business processes. They will not survive otherwise. We are already seeing signs of this in corporate India. The early adopters are showing strong cost control with robust revenue traction. Everyone will have to follow.

 

 

The author is with Amansa Capital