Assuming we all agree that KF is reprehensible, the question is where is the line of moral obligation to stop supporting them. In the most reductive case, you could argue that anyone selling food or water to white supremacists is supporting white supremacy. Or that firefighters who put out a fire at a white supremacist's house are supporting white supremacy. But I don't think people generally consider that to be providing support, whereas they would consider, say, hosting a conference and paying white supremacists to speak at it as supporting white supremacists. So where is the exact line where it becomes "support"? It's ambiguous.
Cloudflare is in an unenviable position of being exactly on the line where moral obligation rests, further complicated by providing different products around hosting and DDOS mitigation. IMO, there is a clear distinction between hosting content and providing DDOS mitigation services, as the article suggests. And just as you wouldn't want your electric company unilaterally deciding that you were a white supremacist and cutting your power, I agree with CF's stance that they shouldn't be making extrajudicial decisions about which customers to use their DDOS protection.
This, incidentally, is why the term "racism" (and similar supremacist terms) must be understood to refer to embedded social structures that may include honest, and often honorable, people unwittingly perpetuating them and not just being mean to the target. So, the question you are in effect raising is how Cloudflare can be ANTI-transphobic and not simply trans-supporting, and whether it has a moral or other obligation to do so. To me, it's a very easy answer with an extremely difficult execution.
When you speak of honorable people unwittingly perpetuating racism, do you really mean grocery stores that don't perform ideological purity tests on their customers?
The harm of kiwi farms is centralized and immediate and obvious, the harm of centralized control is insidious, long-term, and surprises you without warning once you establish the precident years later.
Someone said The Daily Stormer and 8Chan are online today.[1] But this would mean Cloudflare can stop doing business with Kiwi Farms without limiting speech.
How is it the opposite?
Whether Cloudflare is hosting the content or mitigating DDoS attacks against it, they bear some responsibility for the content being accessible. I see no good reason for them to have different policies between their hosting and DDoS mitigation services if they actually care about not propagating the content they refuse to host.
> This policy seems to only apply to content that they host.
Yes, which is exactly what the parent commenter was saying.
> Our decision to disable access to content in hosting products fundamentally results in that content being taken offline, at least until it is republished elsewhere. Hosting products are subject to our Acceptable Hosting Policy. Under that policy, for these products, we may remove or disable access to content that we believe: [...] Is otherwise illegal, harmful, or violates the rights of others, including content that discloses sensitive personal information, incites or exploits violence against people or animals, or seeks to defraud the public.