Transcript
0:00Now, the second problem with AI
0:01is that even if every company in America trusted these AI companies,
0:05the money still does not make sense.
0:07Because the business model is broken
0:08in a way we've never seen from tech companies.
0:11Because here's how software is supposed to make money.
0:14Software is the greatest business model ever invented
0:16because you spend a lot of money building the thing at once, right?
0:20And then every new customer is basically free money.
0:22That's why tech stocks have done so well over the past few decades.
0:26When you buy Microsoft Excel,
0:28like, Microsoft doesn't spend anything extra to sell you that copy.
0:31Their costs stay the same,
0:33but their revenue goes up.
0:34And it's the gap between those two lines that is their profit.
0:38And that gap is why tech companies,
0:39the most valuable companies on earth now,
0:41A I, broke that model.
0:44And how they broke it
0:45was every single time you ask chat GPT a question.
0:48Right now, it costs open AI money cause it uses electricity.
0:52Chips are being worn down.
0:53So more customers does not translate to free money anymore.
0:58More customers means more cost,
1:00dollar for dollar. So AI is not a software business.
1:03It's more like a restaurant.
1:05And every time a meal gets served,
1:07somebody has to buy the ingredients.
1:08Every time.
1:09Except this is a restaurant that loses money every time it serves food.
1:13And its plan to fix it is to serve more food.
1:16And Let me give you some context.
1:17In 2025, open AI burned over $20 billion in just one year.
1:22They'd be the first to be this bad other than we work.
1:24And even then this is so much worse than that.
1:26Open AI burned $20.9 billion in 2025.
1:29That's the original financials that the FT and I reported.
1:31And the problem with these companies is
1:33their margins are getting worse
1:34and they actually their costs increase linearly with their revenues.
1:37So he's basically saying that costs increase linearly with revenues
1:41where the two lines are going up together
1:42and the gap never really opens up.
1:45for 25 years every investor has been trained to be patient with this
1:50cause they say well they're losing money now, right?
1:52But at scale their margins will get better, right?
1:55Amazon lost money for years.
1:57Except with AI we just keep on waiting.
1:59And the margins are getting worse
2:01cause every new model cost more to run than the last one
2:05and the market is starting to notice it.
2:07There is no proof that they can improve their margins.
2:09No amount of specialist silicon
2:10or supposed Vera Rubens will bring these costs down.
2:13And we're at a point now
2:14where open AI is now potentially pushing their IPO to 2027
2:17because they couldn't get a trillion dollar valuation.
2:19It's clear that people are waking up to the problem of generative AI,
2:22which is there's not really a business there.
2:24All of this, by the way,
2:25is Not just open AI, cause look at who's paying to build all of this.
2:29This is from Oracle's annual report.
2:30Oracle is a particularly scary one
2:32because they are building 7.1 gigawatts of capacity
2:35just for one customer. And they even said in their annual report
2:37that the risk was they might not get paid.
2:39Open AI only loses money,
2:41and I think I estimate it's like $75 billion of revenue annually.
2:44Now all of this, by the way,
2:45is not just open AI, cause look at who's paying to build all of this.
2:59Open AI only loses money, and I think I estimate it's like
3:02seventy five billion dollars of revenue annually.
3:04That they will have to pay for the full Stargate data center project
3:06in annual compute revenue.
3:08Open AI can't afford that.
3:09And if they can't, Larry Ellison can't afford to pay back those bills.
3:11And Oracle stock will be in jeopardy
3:13along with the margin loans that mr.
3:14Ellison holds. It's genuinely dangerous.
3:16So Oracle is building the equivalent of several nuclear power plants
3:20worth of electricity for basically one customer.
3:23A customer that just lost $20 billion.
3:26Here's my favourite one though.
3:27Nvidia sells its chips to a group of smaller cloud companies.
3:32They're called Neo Clouds.
3:33Now those companies borrow billions of dollars to buy Nvidia's chips
3:36and Nvidia rents them back.
3:39I think companies like core even,
3:40especially nebulous and iron and Cypher Mining and all of them.
3:43Terror Wolf as well. They are all very.
3:45They're basically outgrowths and sub their subsidiaries of Nvidia.
3:49Nvidia is now what, according to the information,
3:51going to be paying them to rent back their GPUs
3:53when they install them in the data center?
3:55This is the. This is something that only happens in an industry
3:58without diverse and real demand.
3:59So what he's saying there
4:00is that Nvidia's sales are partially funded by Nvidia.
4:04And that's like a car dealership lending you money to buy a car
4:07and then paying you to borrow the car back for the weekend
4:10and then reporting all of it as demand.
4:12Now look how much profit we're making, right?
4:14Yeah, because you are buying back your own equipment.
4:16There's a name for when an industry starts doing this.
4:19It's called not enough real customers.
4:21So for companies investing trillions in AI like Microsoft, Google,
4:24Amazon, meta,
4:26what is the ROI from all their spending?
4:29They won't tell you. These companies report everything.
4:31Cloud revenue, ad revenue,
4:33YouTube revenue. But AI revenue.
4:50They're not Telling us that Microsoft,
4:52Google, and meta and Amazon are all doing a funny little.
4:54I wanna call it a scam, but it's a.
4:56A trick where.
4:57Because their other businesses are still growing,
4:59but they never disclose their AI revenues.
5:00Everyone conflates that with AI driving their growth.
5:02In reality, their other businesses are growing,
5:04and AI is losing the money across the board.
5:06You'll notice that neither Microsoft or Amazon,
5:09who both share their run rate of AI,
5:11will share the actual AI revenues.
5:12That tells you that these companies are afraid.
5:15Public companies love good news.
5:16If they had good news, why wouldn't they share it?
5:17That's because they've only got bad news here.
5:19Now, as of right now,
5:20the stock market is still patient,
5:21and investors are saying, okay,
5:23give it time. Still.
5:24But all of it really depends on one big Assumption,
5:28which is that if and when the profits do come,
5:32it's going to be the American companies that will make the profits.
5:35Because the world has no other option.
5:37But the third problem with AI is that the world has another option.
5:41That option is called China.
5:42So let me show you what China is really doing.
5:44Remember this chart from the beginning of the
Mind Map
Loading mind map…
Viral Breakdown
Hook (أول 3 ثوانٍ)
- ما يحدث حرفياً في السطر الافتتاحي: "Now, the second problem with AI is that even if every company in America trusted these AI companies, the money still does not make sense."
- نوع نمط الخطاف: ادعاء جريء (bold claim) – يعلن أن المال لا معنى له حتى لو وثقت الشركات بشركات الذكاء الاصطناعي.
- لماذا يجعلك تتوقف عن التمرير: يبدأ بتحدي افتراض شائع (أن الذكاء الاصطناعي استثمار مربح) ويخلق فجوة معرفية فورية – المشاهد يريد معرفة "لماذا المال لا معنى له؟".
Emotional Rhythm
- فضول → توتر → صدمة → سخرية → قلق → غضب → دهشة
- الفضول: يبدأ بادعاء أن نموذج الأعمال مكسور.
- التوتر: يشرح أن كل سؤال لـ ChatGPT يكلف المال (عكس البرمجيات التقليدية).
- الصدمة: يكشف أن OpenAI خسرت 20.9 مليار دولار في 2025.
- السخرية: يشبه الذكاء الاصطناعي بمطعم يخسر المال مع كل وجبة.
- القلق: يتحدث عن Oracle وNvidia و"القروض الهامشية" و"الانهيار المحتمل".
- الغضب: يصف "الحيلة" التي تفعلها Microsoft وGoogle بعدم الإفصاح عن إيرادات الذكاء الاصطناعي.
- الدهشة: يختتم بـ"المشكلة الثالثة" – الصين – لخلق تشويق للمقطع التالي.
- ذروة المشهد: "Open AI burned $20.9 billion in 2025" – لحظة الصدمة الأكبر.
Keyword Density
- "AI" – 20+ مرة. يدفع الخوارزمية (موضوع ساخن).
- "money" / "cost" / "revenue" – 15+ مرة. يربط الموضوع بالاقتصاد (جاذب عاطفي).
- "loses money" / "burned" – 5 مرات. يخلق إحساساً بالخطر.
- "margins" – 4 مرات. كلمة تقنية تزيد المصداقية.
- "business model" – 3 مرات. يحدد الإطار الفكري.
- "customers" – 4 مرات. يربط الجمهور بالمشكلة (كلنا عملاء).
- "scam" / "trick" – 2 مرات. يثير الغضب والفضول.
- "China" – 3 مرات. يضيف بُعداً جيوسياسياً (يثير القلق الوطني).
الخوارزمية: "AI" و"revenue" و"costs" ترفع الوصول (موضوع رائج).
الجذب العاطفي: "loses money" و"burned" و"scam" تخلق تفاعلاً عاطفياً (غضب/قلق).
Why It Spreads
- يكسر السردية السائدة: معظم المحتوى عن الذكاء الاصطناعي متفائل (إنجازات، توقعات). هذا الفيديو يقلب الطاولة: "الذكاء الاصطناعي ليس مربحاً". هذا التحدي ينتشر بسرعة لأنه نادر.
- الدليل: "AI is not a software business. It's more like a restaurant. And every time a meal gets served, somebody has to buy the ingredients. Every time."
- أرقام صادمة وسهلة التذكر: 20.9 مليار دولار خسارة، 75 مليار دولار إيرادات مطلوبة. الأرقام الكبيرة تجعل الفيديو قابلاً للاقتباس (shareable).
- الدليل: "Open AI burned $20.9 billion in 2025."
- يستهدف جمهورين في آن واحد: المستثمرين (يتحدث عن الأسهم والهوامش) والمستخدمين العاديين (يتحدث عن ChatGPT). هذا يوسع نطاق المشاركة.
- الدليل: "every single time you ask chat GPT a question... it costs open AI money."
- يستخدم "الفضيحة" كأداة انتشار: يتهم الشركات الكبرى (Microsoft, Google, Nvidia) بـ"الحيلة" أو "الخداع". الفضائح تزيد التفاعل (تعليقات، مشاركات).
- الدليل: "I wanna call it a scam, but it's a trick where... they never disclose their AI revenues."
- يخلق تشويقاً للمستقبل: ينهي بـ"المشكلة الثالثة" (الصين) دون حلها، مما يدفع المشاهدين للبحث عن الجزء التالي أو التعليق بفضول.
- الدليل: "But the third problem with AI is that the world has another option. That option is called China."
What You Can Steal
- استخدم "التشبيه المقلوب": شبه الذكاء الاصطناعي بمطعم يخسر المال. هذا يبسط فكرة معقدة ويجعلها لا تُنسى. في فيديو آخر، يمكنك تشبيه أي تقنية جديدة بشيء يومي (مثلاً: "الواقع الافتراضي مثل حذاء رياضي – يبدو رائعاً لكنه يؤلم بعد ساعة").
- ابدأ بنفي الافتراض السائد: بدلاً من "الذكاء الاصطناعي رائع"، ابدأ بـ"الذكاء الاصطناعي ليس مربحاً". هذا يخلق فجوة معرفية تجبر المشاهد على الاستمرار. في أي موضوع، ابحث عن الافتراض الشائع واقلبه.
- أضف "الفضيحة" كطبقة عاطفية: اتهم الشركات الكبرى بعدم الشفافية (مثل "لن يخبروك عن إيراداتهم الحقيقية"). هذا يثير الغضب والثقة في نفس الوقت، مما يزيد التفاعل. لكن تأكد من أن لديك أدلة (أرقام، تقارير).
