Back to Thoughts
AI & TechnologySolution SeekerJunior ClassSenior Class

Wake up and smell the coffee [Part 2]

Strive MasiyiwaStrive Masiyiwa
December 10, 2025
Image credit:  KWB - Ubuntu Hope via Google Gemini Nano Banana Pro [GPU image purchased from stock photo company] // Strive Masiyiwa Social Media (Facebook & LinkedIn)

Image credit: KWB - Ubuntu Hope via Google Gemini Nano Banana Pro [GPU image purchased from stock photo company] // Strive Masiyiwa Social Media (Facebook & LinkedIn)

Originally posted on:LinkedInFacebook

__Why I launched GPU as a Service with YOU in mind

At the Africa Tech Festival last month, I overheard someone asking: "What exactly is a GPU anyway?!" I was a little surprised, but I know once upon a time someone asked, "What is a light bulb?" or even "What is a car?" Imagine!

A "Graphics Processing Unit" [GPU] is basically a specialized processor "supercomputer" with highly complex superfast circuitry that can do thousands of complex mathematical computations in parallel. They can do analysis and make algorithmic predictions almost faster than you can blink your eye. Billions of microscopic transistors and "interconnectors" on silicon chips make this possible!

Why does this matter? All this makes #SolutionSeeking faster than ever before in human history!

Now these GPUs are expensive, very, VERY, expensive! When I saw how much, I almost choked! And you can't build an AI factory with just a handful! They come in clusters of thousands, installed in specially built data centres with excellent cooling systems and computer engineers staffing them 24/7.

Even most large companies like mobile operators and banks cannot afford this infrastructure, unless, of course, it is a behemoth from Europe, Japan, the USA, and some Asian countries.

This is why so-called experts expected Africa wouldn't get access to GPUs... because they couldn't imagine we'd come up with a viable business model! But of course we did.

__Our #Solution is called "GPU as a Service" [GPUaaS].

Cassava and its investors have taken on the high infrastructure costs to install GPUs in our specially outfitted AI factory data centres which will be located in several hubs starting in South Africa. Then we simply rent out access [via the Cloud] to those who need it and charge based on usage, which can be a few months or even years.

Yes, we rent out #Compute capacity, designed so you can access it from anywhere in Africa. We call this "AI Compute".

So what does this mean for an African #StartUp and other enterprises?
If you're developing an AI-based service in your own mother tongue, for example, you'll likely first "train" and create the AI model using high-quality African language data.

Once trained, you "execute" via the "inference process" where your model starts making predictions and output with new data. The guys who really understand this simply call it "running an AI workload on a high-speed computer".

__Financial services companies, digital content creators, scientific researchers, AI and machine learning developers all need to run these workloads, and they are, all over the world.

In Africa, GPUaaS involves contacting our sales teams to discuss your needs and our usage-based payment model. Of course there are a few Ts&Cs, but it's straightforward.

University research scientists working on agriculture, pest control, wildlife management, and climate patterns have already begun signing on. Enterprises like banks and mobile operators have been especially quick to see the opportunity. I earlier shared about our social enterprise partnership to make access available to NGOs working in health, education, and agriculture.

Many tech-savvy governments have also moved from words to #Serious actions, especially creating what we call a "Sovereign Cloud" using a SuperComputer.

I'm oversimplifying, but your starting point as an #Entrepreneur is to acquire #Knowledge and #Skills without delay. You don't get that in WhatsApp groups discussing gossip! Get #Serious about this stuff.

My young brother and dear friend Ralph Mupita, who runs MTN, recently said he now sets aside several hours every night to "go to school on AI". That is what #Serious really means.

For me, you, and all of us!

What #Serious actions are YOU taking now? I've shared some of what we're doing.

We've made a very #Serious investment in AI infrastructure in Africa for you as #Entrepreneurs, our nations, our enterprises, our NGOs, and research institutions to have the tools needed to deploy new technologies to grow opportunity and prosperity and... launch our own tech unicorns!

We know it's all moving fast, but with GPUaaS, we've made access to the best #Compute in the world affordable, for you.

Share this thought

3 Key Business Lessons

Generated by AI analysing this thought
1

Investing in advanced technology infrastructure can unlock significant opportunities for businesses. By recognizing the need for access to resources that were previously considered unattainable—such as specialized computing power—entrepreneurs can create services that democratize technology and enable a broader spectrum of organizations to innovate and grow. In this case, GPU as a Service (GPUaaS) illustrates how taking on the initial high costs of infrastructure can yield dividends by providing scalable solutions to diverse industries.

2

Entrepreneurs should prioritize acquiring knowledge and skills relevant to their industry, especially in emerging fields like AI. As highlighted by the speaker, the importance of continuous learning cannot be overstated. By dedicating time to understand complex technologies and leveraging resources available through innovative services, leaders can better position themselves and their organizations to capitalize on new trends, fostering a culture that values education and adaptation.

3

Collaboration and partnership with various sectors can amplify impact and broaden reach. By engaging with NGOs, educational institutions, and governments, businesses can expand their service offerings and foster community development. This approach not only builds a supportive ecosystem but also enhances shared understanding and resource distribution, ultimately driving collective progress and creating avenues for sustainable growth.

Afterthoughts

1

"The best way to predict the future is to invent it". - Alan Kay

2

Some of you in the #SeniorClass have been following AI for a while, as I think I first started writing about it maybe 10 years ago. The techies amongst you will appreciate the details in this website of our partner Nvidia which produces our GPUs:

https://www.nvidia.com/en-us/about-nvidia/#About%20Us

3

"GPUs have become the workhorse of modern artificial intelligence, enabling the training and deployment of complex AI models that power everything from image recognition to natural language processing.

Their ability to perform massive numbers of calculations simultaneously makes them well-suited for the computationally demanding tasks at the heart of AI.

GPUs accelerate the training of AI models, enabling researchers and developers to iterate on models more quickly and unlock breakthroughs in AI capabilities
".

This article will help you understand GPUs in more depth:

https://cloud.google.com/discover/gpu-for-ai

4

As you all work to #KnowledgeUp, my team found this 1h,30+/-min documentary for you called “The Thinking Game”, a film made over five years about my friend, Nobel Prize laureate Demis Hassabis and his team at AlphaGo… Most of you will find it amazing!

https://www.youtube.com/watch?v=d95J8yzvjbQ

5

Any business can use AI to improve any aspect of the “3Ps” that I always talk about: Take #People for example: you can do 10-100 times more with a smaller team; you can #Scale your business by adding AI-based #Processes; and you can either improve your #Products or create new ones.

Personally, I’m doing all three! Those of you already very #Serious, what are you doing?

6

Compute intensive workloads require HUGE infrastructure investments, out of reach of most except the biggest countries and institutions globally. For the #JuniorClass, here are some use cases for GPU cloud computing that an #Innovator, #Entrepreneur, #Scientist, #Developer, or big #Enterprise could do affordably, utilizing GPU as a Service:

#Train large language models
#Develop deep learning models
#Power computer vision, predictive analytics and AI applications
#Process medical imaging data at lightning speed
#Develop autonomous vehicle algorithms
#Accelerate drug discovery
#Run climate and financial scenario simulations
#Develop risk modeling applications
#Detect fraudulent transactions
#Conduct real-time market analysis
#Render high res videos, even 3D animations
#Reduce production timelines and improve output quality.... and on and on and on!

Have questions about this topic?

Ask our AI, trained on Strive Masiyiwa's complete writings.

Ask Sahwira AI