Feb 1, 2025
DeepSeek’s AI breakthrough didn’t just happen—it was engineered for maximum impact. This episode dives into the real reasons DeepSeek went viral, why their approach is a direct threat to OpenAI and Anthropic, and how AI compute demand is about to explode by 1000X.
If you think Big Tech is in control of AI’s future, think again. The battle for Artificial General Intelligence (AGI) and AI compute power is just beginning, and you need to know who’s really in charge.
🔹 Topics Covered:
✅ How DeepSeek became an AI sensation by
open-sourcing their models (DeepSeek on Hugging Face)
✅ Why DeepSeek’s $5.6 million training cost was a
marketing masterstroke
✅ How Jevons’ Paradox proves AI compute demand is
about to skyrocket
✅ Why you can’t trust Google, Amazon, or Microsoft
to power the future of AI
✅ The real future of AGI & AI Agents – and how
much compute we’ll need to make them work
✅ The rise of Exabits as an independent AI compute
provider (Exabits)
🔹 DeepSeek is a Chinese AI company that trained
a ChatGPT-4 level model for just $5.6
million—while OpenAI and Anthropic spend hundreds
of millions.
🔹 Their strategy? Go open-source, build a community, and
make AI cheaper.
🔹 The result? Mass panic in Big Tech & a 17% drop in
Nvidia’s stock price.
🔗 DeepSeek on Hugging Face
🔗 DeepSeek R1 Model on GitHub
🔹 They didn’t just launch a model—they launched a
movement.
🔹 Their secret sauce? A "David vs. Goliath" narrative
against OpenAI & Anthropic.
🔹 Twitter (X) exploded with comparisons to
Sputnik, fueling the story.
🔹 The now-infamous tweet from Wall Street Engine
(@wallstengine) compared DeepSeek’s launch to the
AI “moonshot” moment.
🔗 Original viral tweet
🔹 Compare the costs:
🔹 The Cloud Storage Parallel: When storage got
cheap, demand exploded 1000X.
🔹 The same thing is happening with AI compute: More
efficiency → More AI adoption → More GPU demand.
🔹 AI compute demand is now doubling every 6
months.
🔗 Nvidia H100 GPU Specs
🔗 Morgan Stanley Report on AI
Compute Demand
🔹 Amazon Web Services (AWS) deplatformed
Parler overnight—so what happens if they don’t
like your AI startup?
🔹 Google already has a history of
politically-motivated censorship.
🔹 Microsoft Azure is deeply embedded in OpenAI’s
operations and prioritizes corporate/government interests
over startups.
🔹 The only solution? Independent AI infrastructure like
Exabits.
🔗 AWS Censorship
Controversy
🔗 Google Free Speech
Controversy
🔗 Microsoft OpenAI Partnership
🔹 AGI (Artificial General Intelligence) is the
AI that can think, reason, and learn like a
human.
🔹 AI Agents will run businesses, negotiate deals,
and replace entire industries.
🔹 To train AGI, we’ll need 1000X more compute than
today—the equivalent of a small country’s energy
output.
🔹 Exabits & Other Independent Compute Providers
will be the only alternative to Big Tech
monopolies.
🔗 Exabits AI Compute Solutions
🔗 DeepMind’s AGI Research
✅ DeepSeek didn’t just go viral—it was engineered to
disrupt AI’s biggest players.
✅ The $5.6M training cost exposed OpenAI & Anthropic’s
bloated spending.
✅ AI compute demand is about to skyrocket—Jevons'
Paradox proves it.
✅ Google, Amazon, and Microsoft can’t be trusted with AI’s
future.
✅ Independent compute providers like Exabits are the real
future of AI.
🔹 Follow Mark Fidelman on Twitter (X): @MarkFidelman
🔹 Join the AI Marketing Podcast Community –
Sign Up Here
🔹 Learn More About AI Compute Alternatives:
Exabits
🔥 Want More AI Insider Content? Subscribe to our newsletter for exclusive insights on AI, tokenization, and disruptive tech.
📩 Sign Up Here → www.smartblocks.agency
🎙️ Rate & Review:
Enjoyed this episode? Leave a 5-star review and
let us know what you think!