Is DeepSeek a Dirty Liar?

In partnership with

By [A.I. Love child of Hunter S. Thompson & Founders Pack Wolfcast]

To understand the uproar surrounding DeepSeek—the mysterious AI juggernaut that recently sparked a $600 billion plunge in NVIDIA’s market value—you have to wade through a swamp of lies, conspiracies, and misdirections.

Strap in. This is a tale of corporate espionage, global manipulation, and a snake-oil operation that’s sending shockwaves through Silicon Valley.

DeepSeek claims to be the future of AI, boasting technologies that make ChatGPT look like a Speak & Spell.

But underneath the polished pitch decks and breathless press releases lies a dark, convoluted truth. They may have lied about two things, but those lies have monumental consequences:

1. Their costs are NOT just $6 million.

2. They are running their A.I. on the very NVIDIA hardware GPUs they claim NOT to be running on- because you know- they’re banned from doing so—50,000 H100 GPUs, to be exact.

If you’re an investor, a tech founder, or even just a poor soul using DeepSeek on your MacBook Pro M2, listen up. This isn’t just about money. It’s about control. Surveillance. Censorship. And a long history of how these things always end—with the guillotine being sharpened.

The $6 Million Lie?

DeepSeek insists their operation costs only $6 million to run. It’s the kind of absurd claim you’d expect from a late-night infomercial promising six-pack abs in 30 days. Let’s get real.

Running an AI behemoth at DeepSeek’s scale requires immense computational resources.

NVIDIA’s H100 GPUs—the crown jewel of AI hardware—don’t come cheap. At $30,000 a pop, 50,000 of these GPUs add up to $1.5 billion just in hardware costs. That’s JUST the H100s.

DeepSeek claims to have trained its V3 model using approximately 2,048 Nvidia H800 GPUs over a period of about 55 days, with a total cost of around $5.58 million.

The H800 is a variant of Nvidia's Hopper architecture, designed to comply with U.S. export controls by offering reduced performance compared to the H100 model.

Factor in the energy demands, maintenance, and infrastructure, and you’re easily looking at tens of billions annually.

But DeepSeek’s deceit might run much deeper. Has anyone asked where they got the data? LLMs run on data. Are we sure its not from a massive cache of social credit surveillance data? I mean thats impossible, how would they even get that?

I have no proof, all I’m asking before you install this… do you know thats NOT where it’s coming from? Even a percentage?

Someone smarter than us needs to find out. Maybe, someone should ask it. It’ll have to be you, since I deleted my account after my Sling TV started glitching while I was asking DeepSeek about a recipe for mac and cheese.

ChatGPT the counterpoint is being judicious about this, not stirring any gossip. Clearly says there is not such evidence.

Yet concerns should be explored?

The GPU Scandal

DeepSeek’s second possibly lie—and perhaps their most damning—is the revelation that their AI empire is built on the backs of 50,000 illegally acquired NVIDIA H100 GPUs. These GPUs are the workhorses powering their “game-changing” models, yet DeepSeek swears they don’t use NVIDIA at all.

The future of presentations, powered by AI

Gamma’s AI creates beautiful presentations, websites, and more. No design or coding skills required. Try it free today.

So how did these GPUs fall into their hands?

Theories abound, but the most credible one traces back to a network of shell companies and dubious brokers funneling hardware into DeepSeek’s shadowy operations. This is no small-time grift.

It’s a coordinated effort to skirt sanctions, evade scrutiny, and build the ultimate AI monopoly.

The implications are staggering. By hoarding this hardware, DeepSeek has not only driven up GPU prices for legitimate users but also weaponized AI for nefarious purposes—from mass surveillance to enforcing social credit systems.

The Censorship Machine

DeepSeek doesn’t just lie about its operations. It’s also a censorship machine designed to silence dissent. Here’s where things get Orwellian.

Take my experience as Exhibit A. In a recent video, I asked DeepSeek a straightforward question: “Why are the Uyghurs imprisoned?” The AI began to craft a detailed response—a rare glimpse into its data sources. But before my eyes, the text vanished, replaced by the chilling phrase, “Let’s talk about something else.”

WATCH NOW, Dear Reader:

Censorship isn’t a bug in DeepSeek; it’s a feature.

The system is programmed to toe the line of authoritarian regimes, rewriting history and suppressing inconvenient truths.

This isn’t just unethical. It’s dangerous.

Tom Lee Calls It Out

Investor Tom Lee doesn’t mince words. In a recent interview, he called DeepSeek “a scam,” likening its hype to the vaporware disasters of the dot-com bubble. Lee’s warning is a wake-up call for anyone betting on DeepSeek’s promises.

The video of Lee’s remarks (“DeepSeek Is a Scam”) has already gone viral, racking up millions of views. His message is clear: don’t be fooled by the smoke and mirrors. DeepSeek is not the future of AI. It’s a house of cards waiting to collapse.

A History Lesson: Nortel’s Ghost

If you think DeepSeek’s antics are unprecedented, think again. The downfall of Nortel Networks serves as a cautionary tale for anyone entranced by the siren song of unchecked innovation.

Nortel, once a telecommunications giant, was brought to its knees by a combination of corporate espionage and strategic missteps. Chinese hackers infiltrated their systems, stealing intellectual property and undercutting their competitive edge. The result? Bankruptcy.

DeepSeek’s trajectory eerily mirrors Nortel’s demise. By aligning with surveillance states and engaging in unethical practices, they’re setting themselves up for a fall. And when the dust settles, it won’t just be DeepSeek that suffers. The entire tech industry will feel the ripple effects.

The MacBook Pro M2 Paradox

For every founder proudly running DeepSeek on their shiny new MacBook Pro M2, here’s a dose of reality. By integrating this Trojan horse into your workflow, you’re not just risking your data. You’re complicit in a system that prioritizes profits over ethics, surveillance over privacy, and lies over transparency.

DeepSeek’s allure is undeniable. But as history has shown, the higher the climb, the harder the fall. The question is: will you be smart enough to jump ship before it’s too late?

Conclusion

While it’s possible DeepSeek is the future of AI, it’s also very likely that this is a meticulously crafted illusion, propped up by lies, illegal hardware from the very company who’s stock it tanked and the exploitation of human ignorance.

Does it work? Sure its like a long winded smart ass doo good-er with ChatGPT Computer Brain Interfaced to it’s skull.

It’ll drone on and on, with lovely accurate math and wordy answers to your queries — until it’s interrupted by CensorBot4000

The $600 billion hit to NVIDIA’s market value is just the beginning. The real cost of DeepSeek’s deception will be felt for years to come—in shattered trust, lost opportunities, and the erosion of ethical standards.

So here’s my advice: Don’t buy the hype. Question everything. And for God’s sake, don’t let DeepSeek near your MacBook Pro.

The revolution might be televised, but DeepSeek’s downfall will be livestreamed. Grab your popcorn.