Whalesbook
Depreciation Ka Jhol?
Dekho, sabko pata hai ki AI mein paisa bahut lag raha hai. Bade bade companies $600 billion tak kharch karne ka plan bana rahe hain 2026 tak. Lekin Burry ka kehna hai ki companies jis tarah se apne hardware (khaas kar GPUs) ki value kam kar rahi hain (depreciation), woh sahi nahi hai. Unke hisaab se, ye chips sirf 2-3 saal chalte hain, par companies shayad 3-6 saal ka hisaab laga rahi hain. Is 'depreciation trick' se companies ka reported profit 20% tak badh sakta hai, jo ki $176 billion tak ka jhol ho sakta hai 2026-2028 ke beech! Yeh toh seedha earnings ko phulane wala kaam ho gaya, hai na?
Paisa Kahan se aa raha hai aur Kitna?
Yeh sab AI infrastructure banane ke liye jo kharcha ho raha hai, woh bhi kamal ka hai. Companies ne 2025 mein $108.7 billion ka debt liya hai sirf tech sector mein. Poori duniya mein non-financial companies ne 16.7% debt tech companies ko diya hai. Unka kehna hai ki companies $1.5 trillion tak aur debt le sakti hain! Aur haan, Amazon jaise giants ne toh $200 billion CapEx ka plan kiya hai, jiske chakkar mein shayad 2026 mein negative free cash flow ho jaye. Tension waali baat hai!
Nvidia Par Special Dhyan
Michael Burry ne khaas taur par Nvidia ko target kiya hai kyuki woh AI ka "purest play" hai. Matlab, unki earning ka zyada hissa AI se hi aata hai. Agar AI ki demand kam hui ya hardware jaldi out of date ho gaya, toh Nvidia par seedha asar padega. Abhi Nvidia ki market cap lagbhag $4.66 trillion hai aur uska P/E ratio 60-75 ke aas-paas chal raha hai. Analysts ne target price $254.54 diya hai, jo lagbhag 25% ka upside dikhata hai. AMD bhi $320 billion ki market cap ke saath 75 ke P/E par hai, aur uska target price $283.69 hai, matlab lagbhag 45% upside. Lekin sawaal yeh hai ki kya yeh valuations, hardware ke jaldi purane hone ke risk ko sahi se pakad pa rahe hain?
History Se Seekh Lo!
Burry ne isko 1920s ke radio craze se compare kiya hai, jab RCA ka share 200x bhaga tha aur phir 1929-1932 ke beech 98% gir gaya tha! Us time bhi P/E ratio 72x tak pahunch gaya tha. Matlab, technology chahe kitni bhi achhi ho, agar valuation galat ho toh stock crash ho sakta hai. Ab AI ke saath bhi wahi risk ho sakta hai.
Final Verdict Kya Hai?
Nvidia ki sabse badi problem yahi hai ki woh AI ke upar bahut zyada depend karta hai. Agar hyperscalers ne spending kam kar di ya hardware jaldi purana ho gaya, toh Nvidia ka revenue aur profit dono gir sakte hain. Yeh kehna ki purane GPUs ko adapt kar sakte hain, woh alag baat hai, par Burry ko lagta hai ki innovation itni tez hai ki yeh kaam nahi aayega. Aur jab itna sara paisa debt se lagaya ja raha hai, toh thoda bhi gadbad hone par poora ecosystem hil sakta hai. Jab tak AI se real returns nahi aate, tab tak yeh sab sustain karna mushkil hai.