Ben Smith, chair of digital and analytics at global consulting firm Kearney, has witnessed several significant enterprise technology changes throughout his career. The first change came with the advent of the PC, followed by the client-server model, the HTTP browser, and the smartphone.
As he surveyed the market at the end of 2024, he said the AI boom “is bigger than all of them.”
“I think it’s going to lead to a lot of craziness, like we saw between ‘95 to ‘04,” Smith told TechRepublic, referencing the internet boom and the rise of Google. “So, from a vendor perspective, this is a unique time.”
What’s in store for tech vendor competition in the age of AI? And what should organisations do about it? Smith and Anshuman Sengar, Asia-Pacific’s lead for digital and analytics at Kearney, outlined key trends shaping the market into 2025.
Kearney’s experts argue five forces are shaping AI heading into 2025.
Incumbent and regulatory capture occurs when established companies in a market use their influence over regulatory agencies to shape rules that favor existing business models and restrict competition. Regarding big players in the AI market, Smith explained that there is “a lot of effort to create fear in the market so that the incumbents are better positioned” into the future.
While incumbents seek to reduce the chance of being disrupted by others moving more quickly, Smith said that regulatory agencies in the E.U. and U.S. are also showing “anti-acquisition” pushback. This may create an environment for start-ups and challengers to build businesses and grow rather than being acquired by larger organisations seeking to dominate.
Sengar noted that there is potential for start-up innovation in the Asia-Pacific region at the application layer, building on the AI engines embedded into major platforms. He said local innovators, as do corporates like banks, through their own venture funds, have a role to play in competition with large global tech firms.
Smith said that society is experiencing the “largest off-cycle boom in semiconductor manufacturing fabrication facilities, or fabs, that we’ve seen in a long time.” Nationalism and governments are driving this surge, as many are building large data centres for AI computing.
For AI users, this shift will cause “a massive decline in the next 12 to 24 months” in the cost of a token, making AI cheaper.
In APAC, the drive for capacity results in large data centre transactions, such as private equity firm Blackstone’s acquisition of regional data centre company AirTrunk for AUD $24 billion. Sengar said companies across Asia are investing in data facility developments in Indonesia, and he expects India to launch its fab capability soon.
Smith emphasised the importance of data in the emerging dynamics currently unfolding in the AI market. He noted that various players in the ecosystem are working to determine data ownership and usage in an AI context, whether the data resides within a company or externally.
“I think the market is much more sophisticated this time as opposed to when the internet happened the first time,” Smith said.
According to Smith, the capital intensity of AI — particularly in a high global interest rate environment — is a significant factor for technology players aiming to gain an advantage in the field.
“If you look at the telco build out, or the fiber build out from ‘95 to 2002 … the amount of capital being deployed on GenAI data centers, swamps that by multiple factors. I mean, you will see data centres that cost hundreds of millions of dollars relative to the fiber build-outs we saw.”
Smith predicts that, within the next few years, the world will see its first single data centre worth U.S. $1 trillion, which he expects to see built in either the U.S. market or in China.
AI is connected to national economic security and national defense, making it different from previous technology innovations. Smith said this has created a “weird geopolitical environment,” with AI more of a political issue than the internet because of these national security implications.
Kearney experts advise organisations to consider investing in and implementing AI based on their company and industry. Smith warns against getting caught up in the AI hype, while Sengar suggests data can be managed as a priority.
Smith said one dangerous path is for enterprises to test every use case for AI. Instead, he said organisations should prioritise only those use cases that are most important for the business to pursue today. While some business use cases might be effective, they may also be much cheaper to do when the cost of an AI token decreases, so organisations would be better off waiting in some cases.
Executive teams and boards should ask whether AI will fundamentally change the basis of competition in their industry. If it is, then they should move quickly, Smith said, giving the example of drug discovery and insurance markets, where he said AI is likely to dramatically impact how businesses compete. If it is not, he said investing heavily in AI, especially now at higher prices, may prove to be a mistake.
Organisations can act immediately to take control of their data, whether they are pursuing AI heavily now or choosing to wait. Sengar said companies in APAC are struggling with their data, including customer data quality.
“Getting control of your data, especially customer data, and ensuring you understand what you are capturing and that it is secure is something you don’t need to wait for,” Sengar said.
