Stephen Smith's Blog

Musings on Machine Learning…

Archive for August 2020

The Downside of Monopolies in Tech

leave a comment »

The Downside of Monopolies in Tech


I’ve previously blogged on the potential for nVidia buying the ARM CPU business from Softbank, which would solidify a duopoly of CPUs and GPUs between the ARM/nVidia world and the Intel/AMD world. The purchase of ARM by nVidia isn’t a done deal and other purchasers such as a consortium led by Samsung are appearing, but regardless if one of these deals goes through, will this be good for consumers and technology innovation? In this blog we’ll look at some of the reasons and downsides of all the M&A activity that happens in the tech world.

The Bigger get Bigger

A few enterprising people create a hot startup that succeeds, grows and becomes profitable. Why is it considered a success to sell out to a larger company and become a small division in that company? The answer of course is money. Usually as a company grows, the first step is to go public on the stock market, this usually makes the founders of the company billionaires. Next the stock grows for a bit, but flattens out. At this point it becomes a takeover target as larger companies like Apple, Microsoft, Amazon, Google or Facebook which have mountains of cash lying around look to spend it, to move into a market that didn’t previously exist. They make an offer that gives the current shareholders a quick windfall profit and gobble up and assimilate the startup.

This business model isn’t new, but when it was abused at the turn of the previous century, the government instituted all sorts of anti-monopoly regulations and oversight. Over the last few decades, all this oversight has been gutted and pretty much any acquisition is approved.

It’s funny that the US Congress is lamenting the monopoly positions of Google, Apple, Amazon and Facebook when it was Congress’s job to oversee all their acquisitions that put them in this position. The government has to actually use their regulatory powers or accept the consequences. But governments tend not to accept accountability for their actions and just hold pointless hearings to complain.

The government also incentivises this behavior with the way corporate income tax works. Income tax is strictly on profits, so reducing profits, reduces taxes.  If they don’t spend their cash on acquisitions, then they have to pay a good chunk of it in income tax that is just a useless expense as far as corporate accountants are concerned. Many people think companies are crazy at what they spend on M&As, but they don’t consider how much they would need to pay in tax if they didn’t. From the company’s point of view, a gamble on an acquisition is better than a guaranteed loss from paying taxes. The stock market also incentivises this by the way companies report earnings, spending money on M&A isn’t penalized as reducing profits, further companies don’t like earnings to grow too fast as they are expected to beat them next year and don’t want to set the bar too high.

Why is This a Problem?

Sometimes things seem to work out, for instance YouTube seems to have thrived under Google or Instagram under Facebook. But have they really? And what about all the acquisitions that were just buying up competition to shut them down? The big problem is that once a company achieves a certain high percentage of market share, they have no incentive to innovate. They have no real competitors and can just milk their install base for as much money as they can while not incurring R&D expenses.

Consider nVidia buying ARM. What are some downsides:

  1. Less competition in the CPU/GPU market. This usually leads to less R&D spending and reduces innovation.
  2. Too much corporate bureaucracy. A combined nVidia/ARM will be twice the size. Can the current nVidia executives effectively manage such a large company? Or will it be bogged down by committees and worrying too much about competing between nVidia products?
  3. ARM is used by many current nVidia competitors. How well will nVidia work with these competitors like Samsung.
  4. nVidia is an American company. Will this cause problems with working with Chinese companies or other non-US companies? Will it force ARM to add NSA/CIA backdoors into its CPUs?
  5. Will nVidia ownership change ARMs CPU directions, concentrating effort on markets of interest to nVidia and ignoring others?
  6. Will ARM and nVidia become bogged down in regulatory and other legal matters that drag on for years?
  7. Will ARM employees remain at the new company? Will they make enough money from the acquisition to retire or form their own startup? Will they be moved to other nVidia division? Or will they just not like the new corporate environment and move on?

Basically will ARM thrive or will ARM be stifled and frustrated?

A Better Alternative

It would be nice if it made financial sense for companies like ARM, Instagram or YouTube to remain independent standalone companies, listed on the stock market and free from the threat of being gobbled up. I believe ARM would be better off as an independent company not owned by Softbank or nVidia. Perhaps it’s too late for ARM, but for future successful startups, regulatory, tax and stock market rules would all need to change to reward successful smaller companies and put a higher cost on the largest companies gobbling up everyone else. Here are a few things that would need to change:

  1. Governments need to take regulatory oversight of acquisitions seriously actually block a few of them.
  2. Tax law needs to change. Perhaps don’t allow acquisitions to be expensed the way they are now, so they don’t have such a positive effect on income tax. This would also go a ways to solving the problem of the largest and most profitable companies avoiding so much tax.
  3. Put some curbs on the percentage of a company’s shares that can be bought by an outside entity. Require diverse ownership of a company’s shares.

These are probably just a start, as large companies are good at bending rules to their favour. After putting these rules in place, they would need to be tweaked to prevent new imaginative abuses.

I tend to think this approach is better than the alternative of identifying monopolies and then breaking them up. This rarely seems to happen and the couple of cases like AT&T resulted in a big mess.


I don’t know that nVidia buying ARM will be a disaster for ARM customers, but I do feel that it would be better for everyone if ARM was a strong independent company, listed as itself on the stock market, but with some protection against opportunistic acquisition and interference.

Written by smist08

August 7, 2020 at 11:17 am

Posted in Business

Tagged with , , , ,