The second group of pessimists is those who do not see revenue growth in the application layer. Dominant finance and tech players are accustomed to sharp revenue spikes in software companies. Despite booming data centers and cloud revenues, their portfolios lack growth, fueling bubble fears. From their viewpoints, this must correct, if not through their companies’ growth, then a sharp reversal in hardware.
Historically, hardware players, mostly from outside the Valley and still in their early stages, could not exercise pricing power. In the era typified by Steve Jobs and led by Apple, even a whiff of oversupply had them sacrifice margins at negotiation tables despite being numerically few. In any competitive intensity analysis, most hardware segments would feature in the least competing top decile of global industries’ league tables for decades, but that was never their pricing behavior.
The application layer subsegments had far more competition and regionalization, but platform and network effects allowed the first movers to run away with the spoils. This dynamic established Silicon Valley's dominance, leading to the current expectations that, eventually, any hardware spend must result in amplified revenue growth amongst application companies or go phut.
Now, two significant changes are reshaping this landscape. Firstly, explosive growth in coding productivity and the loss of first-mover advantage due to instant copyability have structurally affected all software sub-industries. The changes are perhaps as hard-hitting as those on traditional mail and communication players with the arrival of emails.
Secondly, in the era typified by Jensen Huang and led by Nvidia, the hardware players are fast learning their advantage. This does not mean cycles are over, but these players’ behaviors may be less value-destroying than before when oversupply inevitably arrives.
The killer app, once the holy grail, may not exist in a world where application barriers also must fall for data to converge and develop intelligence through language models. Venture capitalists, influencers, and observers benefiting from the application layer value capture will need more evidence to adjust to the disruptive force. Until they do, expect them to leave their usual optimistic stance and issue dire warnings whenever markets come under pressure.
Absolutely, upstream spend must result in multiplied value creation downstream, but this could be through driverless cars, new consumer gadgets, robotics, drug discovery, and other use cases we have discussed before. This may take time or may not even happen, but it has a higher chance of coming through than increases in software companies’ revenues by the asked amounts.
A parting shot: Donald Trump, if he becomes president, may want cheaper chips and more expensive cars, using his negotiation skills to achieve this. But without such influence, one must rethink the hardware vs. software dynamic.
