The staggering scale of Big Tech

First, Big Tech is investing massively in R&D. Amazon’s annual R&D spend, in particular, is staggering:

  1. Amazon: 14% of revenue (US$73.2B)
  2. Alphabet: 14% (US$39.5B)
  3. Meta: 30% (US$35.3B)
  4. Apple: 7% (US$27.7B)
  5. Microsoft: 13% (US$26.6B)

The Apollo program, adjusted for inflation, cost US$19B per year. Every Big Tech company now spends more than this on research and development. Each year, Amazon spends nearly four times what the US government spent to go to the moon.

Not all of this investment is particularly productive. For example, some of this money goes towards acquihires (acquiring businesses for their teams, rather than their technology or revenue). Big Tech reportedly does this in part to hoard talent within the industry, a defensive move against other Big Tech companies as well as startups. A recent study outlined the inefficient talent allocation that results from this practice, which also negatively impacts startup formation.

P&Ls at Big Tech are changing in the AI-era. Google now spends more on compute than on people. This is a big deal. Talent is still the most important cost centre in a company. However, talent is no longer exclusively human. As AI grows increasingly capable, compute will become the key cost centre in businesses that operate models, redefining talent entirely.

Lastly, let’s talk electricity. Over the past few decades, Big Tech got big enough to require their own internet infrastructure, including undersea cables. Now, Big Tech is big enough to require dedicated energy infrastructure. Geothermal-via-fracking now running some Google data centres, and Microsoft has made commitments to purchase modular nuclear reactors to power its data centres.