Thunderf00t: BUSTED!

Starlink Edition

Introduction

On July 13, 2021, Thunderf00t released this video attempting to debunk Starlink. If you haven’t read my analysis of Common Sense Skeptic’s similar attempt, I suggest starting there. I won’t go into as much detail about the system itself as I did there. This is a lazy article, I’m not breaking down the video point by point with quotes and timestamps.

Thunderf00t has a rambling, repetitive way in his videos, so I’m going to start with a breakdown of Starlink’s situation, according to him.

Starlink, According to Thunderf00t

Thunderf00t takes quite a bit of time to bring everything together, and changes quite a few numbers back and forth along the way. I have brought the final numbers he uses for everything together. (Costs are all listed as totals per 5 year period as the satellite constellation needs to be replaced every 5 years.)

• $5 billion - Cost of data.

  • $0.01 per GB.

  • 3 million global customers.

  • (Calculated) 2,777 GB per user, per month.

• $15 billion - Cost to manufacture and launch 10,000 satellites.

  • $500,000 - Cost to manufacture each satellite.

  • $1 million - Cost to launch each satellites.

    • Calculated with a $60 million Falcon 9 launch cost divided by 60 satellites.

• 1,000 customers per satellite.

  • 20 Gb/s throughput per satellite divided by 20 Mb/s per user.

• $15 billion - Revenues from 3 million customers.

• ($5 billion) - Loss per 5-year cycle.

Now, there are a lot of problems here, but I’m going to show how far off Thunderf00t is in his analysis by only correcting one number.

His assumption(not explicitly stated, but calculated using his data costs and maximum number of customers) is that customers will use, on average, 2,777 GB per month.(100 billion GB per year / 3 million customers / 12 months) This is frankly absurd. According to OpenVault, around the time of Thunderf00t’s video, the average household in the US was using around 435 GB per month. In Europe, that figure appears to be around 200 GB per month.

Thunderf00t’s calculated data capacity for the entire Starlink system, adjusted from 2,777 GB per user to 435 GB per user, increases his maximum number of subscribers from 3 million globally, to 19 million globally. At $99 per month, that’s $22.5 billion per year in revenues, or $112.5 billion per 5-year cycle.

Comparing that to Thunderf00t’s calculated $20 billion in expenses per 5-year cycle.

Starlink, in Reality

But why stop there? Let’s correct more. I won’t, however, touch Thunderf00t’s data cost numbers, because frankly it doesn’t matter, internet peering and transit are complicated, and I’m no expert.

For the maximum number of customers, rather than relying on the GB used per month calculation, I will instead focus on the maximum throughput of each satellite divided by the average per-second usage of customers, per the FCC. (This is during peak-traffic hours)

• $5 billion - Cost of data.

  • $0.01 per GB.

  • 24.5 million global customers.

    • 7,407 customers per satellite.

    • 10,000 satellites.

    • 1/3rd time spent over land.

  • (Calculated) 337 GB per user, per month.

• $10 billion - Cost to manufacture and launch 10,000 satellites.

  • $500,000 - Cost to manufacture each satellite.

  • $500,000 - Cost to launch each satellites.

    • Calculated with a $30 million Falcon 9 launch cost divided by 60 satellites. (Using internal cost for F9, not customer pricing)

• 7,407 customers per satellite.

  • 20 Gb/s throughput per satellite divided by 2.7 Mb/s per user. (Source: FCC, page 6)

• $146 billion - Revenues from 24.5 million customers.

• $131 billion - Profit per 5-year cycle.

Do I need to say more? I could, there are lots of other little things in the video that are just plain wrong, but I think this shows just how far off Thunderf00t was, just like CSS.