I have never compared it to the Graphcore chip. It was uiux...

I have never compared it to the Graphcore chip. It was uiux claiming AKIDA can compete with it in the cloud and I said there is no way it can.

If it helps anyone to sleep at night trusting false claims, so be it.
I'll outline very clearly why AKIDA can't be used as a cloud.

It doesn't have RNN. Which means it is feed forward neural network.
If this still sounds foreign to people, that means it can only process real-time data.

Any manufacturer can say on the powerslide something that they know they will stand no chance in the real competition, but still able to just get by.
For example, Hyundai can say their car can participate in the Le Man race. But it doesn't mean it will be able to 'compete' at that level.

Let's assume for a second that AKIDA was used on the cloud.
So If I wanted to search for a simple thing as today's weather and ask "Hey AKIDA, what is today's weather like?", let alone more complicated queries that cloud computing processes, AKIDA will not know what to do because it can't comprehend speech. It can detect individual words as it gets pattern matched to pre-saved words, but won't be able to process it as there is no LSTM nor RNN capability.

Also if I wanted to input a video feed and see if it can convert the video into natural language processing (NLP), it will not be able to do it because it lacks LSTM.

Which means it can't go cloud. If it does go to the cloud with that capacity, there will be no point as no one will back Hyundai to win Le Man.