One competitor increases advertising and the others respond. One technology giant builds more data centres and rivals accelerate their own investment. One employer raises compensation for scarce specialists and the market follows. In the end, everyone may become stronger while nobody gains an advantage proportional to the extra spending.
The photograph that inspired this topic is dominated by massive horns. In many ungulates, such weapons are used in male competition. But if larger weapons provide an advantage, rivals also have an incentive to enlarge theirs.
In wild Soay sheep, this trade-off has a measurable cost. A Nature study found that genetic variants associated with larger horns were linked to greater reproductive success, while variants associated with smaller horns were linked to better survival. The “largest weapon” was not a free improvement.
A competitive advantage can be real while forcing every participant to spend more simply to avoid losing relative position.
When everyone runs faster, the ranking may barely change
The Red Queen metaphor has long been used for systems in which constant movement is required merely to stay in place. Business competition often works this way: one firm’s move creates an incentive for rivals to respond.
A classic AOM study analysed more than 4,700 competitive actions. Aggressive moves could improve a firm’s performance, but they also provoked faster and more frequent responses from rivals, weakening the original advantage.
Company A spends more and gets stronger. Company B sees the threat and increases spending too. Both now operate more expensive systems, yet their relative positions may remain almost unchanged.
Advertising is one of the clearest examples
If two brands sell similar products and one sharply increases advertising, the other risks losing attention. It has an incentive to respond. Now both spend more.
Advertising is not waste by definition: it informs buyers, builds awareness and can create demand. The arms race begins when a significant share of new spending is required mainly because the rival is also spending heavily.
Empirical research on advertising competition found that brand activity can move with rival actions. Competitive pressure can push budgets upward even when part of the result is a redistribution of attention.
AI has turned an old mechanism into a race worth hundreds of billions
Frontier AI requires chips, servers, data centres, electricity, networks and scarce technical talent. If one large player sharply increases compute capacity, rivals face a real risk of technological lag.
In July 2026, Meta narrowed its full-year 2026 capital-expenditure outlook to $130–145 billion, tying much of the infrastructure build-out to AI and future compute needs.
Microsoft reported $41 billion of capital expenditures in its fiscal 2026 fourth quarter. After an accounting-related shift in the treatment of future data-centre leases, the company said its calendar-year 2026 CapEx expectation was approximately $175 billion. Microsoft also stressed strong demand and capacity constraints, so it would be wrong to describe these investments as pure waste.
The competitive mechanism still matters: if one participant expands capacity while another stops, the risk of falling behind becomes strategic.
The most expensive scarce resource may be people, not hardware
If a small group of specialists can materially accelerate a technology, one company can offer them higher compensation. A rival must either lose the talent, find an equivalent replacement or raise its own offer.
The number of top specialists does not instantly increase, but the price of access to them can. What gives one firm an edge forces others to spend more to neutralise that edge.
WOW:
Logistics and patents can become defensive races too
After supply disruptions, a company may increase inventory, reserve transport capacity or add backup suppliers. For that company, the decision is defensive. If everyone does the same, scarce resources are booked earlier and the cost of protection rises.
A similar mechanism can emerge in patent portfolios. Patents protect inventions, but in dense technology fields they can also support cross-licensing, negotiation and defence against litigation. The portfolio becomes both an asset and a shield.
The most interesting race is one nobody can safely stop alone
Imagine two firms each spending $100 million on advertising. Both might earn more if they simultaneously cut spending to $70 million and kept roughly the same market shares. But neither can be sure the other will cut first.
If one reduces spending and the other does not, the second may gain. So continuing the race is individually safer even when a cheaper collective outcome exists.
That is why these races are difficult to stop: nobody needs to design them. Each participant is simply responding to the previous move of another.
Competition does not necessarily burn money for nothing
Races can accelerate genuine progress. Processor competition improves performance, cloud competition can reduce computing costs, and AI rivalry can produce more capable tools.
The important distinction is between a productivity race, where spending creates new absolute value, and a positional race, where spending is mainly needed to avoid falling behind.
Real markets usually contain both. The key question is the proportion.
How to recognise when healthy competition becomes an arms race
Use a simple test. Imagine every competitor cuts a particular expenditure by 30% tomorrow. What happens to the customer?
If products become meaningfully worse, innovation disappears or quality falls, the spending was creating real value. If customers notice almost nothing while firms retain similar market shares and higher profits, the industry may be in a positional race.
Sometimes the strongest strategy is to leave the race
If every carmaker competes on engine power, one can compete on efficiency. If major brands fight through advertising scale, a smaller firm can build a narrow community. If AI companies compete on ever-larger general models, another player can specialise in a cheaper model for a specific industry.
There are two ways to beat a rival: build bigger “horns,” or move to a game in which horn size matters less.
A real advantage exists only when the resource creates more value than it costs. Otherwise competition produces a strange world: everyone is stronger, everyone pays more, and the distance between them barely changes.
Publication note: The biological “arms race” is used as an analogy for a competitive mechanism, not as evidence that business or technology markets literally reproduce biological evolution.
