Here's the contrast that breaks your brain:
$MU could buy all of $AMD in about 5 years of earnings.
$AMD would need 42 years, FOUR DECADES, to buy $MU.
But here's the part that makes no sense. $AMD's market cap is ~$800B. Only 20% below $MU's ~$1,000B.
Vicinity valuation but $MU earns +800% more than $AMD does.
A company's job is to generate cash and return it to shareholders in the form of dividends or buybacks or future Capex or investments. So if both started buybacks tomorrow, Micron could return capital 8x faster, because it earns 8x more against a nearly identical market cap.
Make it make sense.
"BNP Paribas analyst Karl Ackerman expects the average selling prices (ASPs) of DRAM and NAND — which surged on AI demand — to peak in mid-2026, much earlier than his original forecast of mid-2027, before turning into quarter-over-quarter declines starting early next year. He argues that this is driven by aggressive capacity expansion from Chinese memory makers such as CXMT and YMTC, combined with reduced memory consumption from other industries due to soaring prices, including a 14% decline in global smartphone shipments this year. He also noted that growing concerns over future oversupply and margin deterioration within the next one to two years are beginning to weigh on the sector in advance."
I’m trying to remember the name of this analyst.
🚨 THE INTEL ERA IS QUIETLY ENDING. AND ALMOST NOBODY IS TALKING ABOUT IT.
Reuters just reported that ByteDance is developing its own CPUs. Not GPUs. The boring chips that run everything behind the scenes.
Intel and AMD have jacked up CPU prices by 10% to 35% in just a few months. Intel is even warning customers their orders will arrive late.
So the biggest buyers are walking.
And ByteDance isn't alone. Far from it.
Google builds its own CPU. So does Amazon. So does Microsoft. Even Nvidia is moving in, with Jensen Huang calling it a $200 billion opportunity.
Now ByteDance joins the list.
Here's what almost everyone is missing.
The AI story everyone tells is about GPUs and training models. That era is fading. The new era is inference, where AI actually does the work, and inference leans hard on CPUs.
That shift created a CPU shortage. Intel and AMD saw it and cranked prices. And in doing so, they handed every major player the same reason to build their own.
This is bigger than one company or one country.
The West is escaping Intel and AMD to protect their margins.
China is doing it to survive sanctions.
Different reasons.
Cerebras $CBRS doesn't buy from Micron $MU . That is the whole thesis.
Every GPU-based inference provider is paying Micron's 80-85% margin on HBM. Cerebras isn't.
Cerebras uses on-chip SRAM - memory etched into the wafer during fabrication, not purchased separately. No HBM dependency. No margin stack. No shortage exposure.
They are also on TSMC's 5nm node, not the 3nm node where most GPU production concentrates. Less congested part of the same foundry.
That is a structural cost advantage at a time when every competitor is absorbing an 80-85% margin tax on a critical component, per @andrewdfeldman on @20vcFund with @HarryStebbings
How Cerebras's SRAM architecture sidesteps the entire HBM supply chain - and why it matters for valuation:
https://t.co/th46XFBTR0
At current levels, $AMD represents a high-risk, high-beta investment. In the event of a market correction or macro shocks, its lack of a fundamental "safety cushion" could trigger a sharp 20–30% drop toward sector medians—falling faster than the more resilient and higher-margin $NVDA.
In my latest article for Seeking Alpha, I break down AMD's core growth drivers, the competitive advantages of its business strategy, key takeaways from the Q1 2026 earnings, and long-term forecasts up to 2030.
The analysis addresses critical questions: Is AMD a buy right now? And if a market correction hits, can the company’s structural catalysts outweigh macro headwinds?
Investors must strictly separate the brilliant execution of AMD's engineers from the market's valuation of the stock. Trading at a massive 56.5% Forward EV/EBITDA premium to its peer group, coupled with an inferior margin profile compared to Nvidia, the stock presents a dangerous asymmetric risk for buyers at these levels.
👇Link to the full article in the pinned comment below.👇
Hey Grok, Take Me Home
After finishing meeting with a client earlier today I got into the car, asked Grok to navigate me home, and then I pressed the button to start Self-Driving.
Nearly an hour later I was home and FSD had handled the whole drive without any issue.
Tesla is the only car that can do this.
NEWS: California trucking firms have applied to buy 1,200+ Tesla Semis, a deal worth ~$348M.
That is more applications than all other electric trucks combined since the state's incentive program launched in 2019.
The Semi starts at ~$290,000 for the 500-mile range version. Competitors from Daimler and Volvo start at $400,000+ for less range. Tesla is also bringing a 350-mile version at a lower price. Both options out-range every other electric truck on the market.
Jennie Abarca, owner of King Fio Trucking in Long Beach, ordered 20 Semis. "This is something new coming to the market that kind of answers all those problems."
The Semi is doing to trucking what Model Y did to SUVs. Better range, lower price, massive demand. The incumbents never saw it coming.
We are no longer overbought so it is possible we can rally once more led by compute-ai and financials and even travel & leisure and manufacturing on the Middle East rebuild
🚨 $AMD
🧵Shocking calamities inside the monstrous $580 billion A.I. Chip Producer’s balance sheet;
Class,
Let us take a gander shall we.
Photo 1.) You can see here Net Revenue declined QoQ along with diminished Operating Income. 🔻
Followed by Increased Cost of Sales, R&D, and SG&A. 🔻
Thus reveling a story of:
- spending more on marketing and sales to generate less 🔻
- playing keeping up with the joneses with R&D 🔻
- corporate/execs are paid more for less operating income 🔻
Photo 2.) Shows us an additional $1.7 Billion in short-term investments QoQ (Stocks/Corporate Bonds). Also shows Accounts-receivable missed QoQ.🔻Then reveals Inventories accelerated further. 🔻Finally we see the misfortune of Goodwill & Acquisition- Related Intangibles of $41.48 Billion Dollars onto the Asset Sheet. 🔻(Now granted this has been on here since 2022 with the acquisition of Xilinx)
What Song does this Sing:
- They are making less revenue but gambling 25% more QoQ on short-term investments 🔻
- Despite Execs/experts calling for supply-constraints, Inventory keeps rising QoQ 🔻
- Goodwill & ARI >50% of their entire Asset sheet. (One small slip and impairments start seeping through and there goes EPS)🔻
Photo 3.) Cash Flow portion looks concerning in a few ways. First thing is this multi-billion corporation filed this section under Three Months Ended but clearly the dates show Twelve Months Ended. ‼️ So which is it. If it is a YoY or QoQ Accounts Receivable still shows a >50% decrease. Also it shows a positive number for accrued liabilities of $713 Million. (Weird)
This story is just not legible:
- Major Accounting error on Dates and Reporting QoQ or YoY?? 🔻
- Accounts either paid prior or less was sold 🔻
- accrued liabilities showing almost $1 Billion is not particularly ‘normal’ unless they just offloaded a butt load debt or inventory of some sort. (maybe old inventory but if its old inventory then we have a bigger problem because inventory still rose) 🔻
Photo 4.) This has Red Flags all along it. Mainly because experts are touting Data Centre Revenue. But this tells a different story. Look closely;
- Datacenter revenue grew at 57% Twelve Months Ending and then the prior Twelve Ending from March 2025-2024 it grew at 57% as well. So steady growth but not accelerating and unsustainable. Most Data Centers are built/close to finalization. A ton are hitting road blocks such as regulatory and sustainability issues. Even civilian protest. 🔻
- Not only that but Operating Income declined QoQ for Datacenters. *Not sure why we are touting datacenters, they suck money, our water, and electricity *🔻
- the nail in the coffin is CPU & Gaming decelerated greatly🔻
*Bonus; GAAP Margins are declining* 🚨
Nothing good is going on at AMD.
I looked at $GOOG & $NVDA balance sheet for comparison on the Goodwill and their’s is not even >10% of Asset Sheet.
I will not be buying this company anytime soon.
- Not enough Cash to pay off Debts
- Asset/Liabilities risk
- Declining Margins
- Declining Revenue
- Increasing Inventories
- Decreasing Datacenter Net Income
- Decreasing CPU&Gaming Sales
- Increase cost of Sales
- Increased Gambling’s
I wouldn’t call this cooking the books.
I would just call this Ignorance is Bliss ✨
📚 Hope this helped
Micron is trading at roughly 9x forward earnings while $AMD trades near 60x.
Micron generates about five times more profit, yet trades at roughly one-fifth the multiple.
I believe Wall Street has this wrong, and $MU reaching $1,000 within two years is not out of the question.
Memory shortage stocks have to go to a higher place. It's very difficult to imagine it, but stocks do gallop to where they should be... WDC SNDK, STX, will be overheated until they get to where they have to go