$AMD: Google picked EPYC. The stock is up 36.7% in four sessions. Here's my sell put.
The four sessions
• Thu Nov 5: $404.40 (−17.26%), the AI chip crash
• Fri Nov 6: $449.04 (+11.04%), a sector bounce
• Mon Nov 9: $508.58 (+13.26%). Google Cloud named EPYC as the host CPU for its TPU 8tx, and Oppenheimer upgraded to Outperform with a $600 target.
• Tue Nov 10: $515.21 (+1.30%). Anthropic's IPO filing showed AMD buying $2.5B of Anthropic shares, part of AMD's commitment of up to $5B linked to Anthropic deploying up to 2 GW of AMD chips.
• Wed Nov 11: $552.90 (+7.32%) on CoreWeave's results and a UBS note on growing CPU sales. About $574 after hours.
From $404.40 to $552.90 is +36.7%. The stock is still 14.9% below its $649.42 record close.
What Google actually said
AMD's 5th Gen EPYC chips run the host side of Google Cloud's TPU 8tx: workload orchestration, data preparation, and high-speed communication with the TPU chips. TPU 8tx is the first cloud TPU with hardware-enforced confidential computing. AMD also says EPYC is the host CPU of choice for more than 10 leading hyperscalers and frontier AI labs.
The topic's question: when does a design win become cloud revenue?
The bears have a point. Contract value, shipment cadence and margin weren't disclosed, and a one-day stock gain isn't booked revenue.
I'm watching for three things:
1. Google Cloud product pages listing TPU 8tx as generally available, with regions and pricing. Not press releases.
2. AMD's Q4 data center revenue against the company's ~$17B ±3% total revenue guidance. Data center was $8.5B in Q3, up 97%, and 64% of sales.
3. More named host-CPU wins. GPUs and custom chips both need a CPU next to them, and EPYC is the default at many of the biggest buyers.
If the first two land, the win is real money. Until then, it's a design win with no dollar figure attached.
The risks
• Hot October CPI pushed the 10-year yield to about 5% and raised the odds of rate hikes
• +36.7% in four days invites chasing. Some of it can give back.
• Nvidia's CEO investigation is pushing money toward AMD. If that resolves, some of that money may go back to Nvidia.
Checking it against my rules
✅ Trend: $552.90 against the 200-day (~$416), +33%
✅ Earnings: next report around early February 2027 (estimated), after expiry
⚠️ Volatility: IV is about 60% (my estimate after the rally). Against a 1-year range of 42–89, that's an IV rank around 38, just under my 40 bar.
Strike math (Dec 18, 36 days, 60% IV)
• 1 standard deviation down ≈ $449
• 50-day × 0.98 ≈ $535
• $450 has delta 0.11, too far out to be worth selling, so I move up to $470 (delta 0.16)
The pick
• Sell 1× AMD Dec 18 $470 put
• Model premium ≈ $9.90
• Break-even ≈ $460, 15% below Wednesday's close
• About 78% model odds of expiring worthless
• $47,000 cash per contract. Tier 2, so 2% max.
• Half size now. Add only if a pullback holds $515 (Tuesday's close).
• Take profit: GTC buy-back at ~$4.95 (50%). Stop: ~$24.75 (2.5× the credit).
Update on last week's pick
On Nov 9 I picked the Dec 18 $370 put at a model ~$9.90. With AMD at $552.90, that put is worth about $0.47 on the model. The 50% take-profit order would have filled days ago. That's model pricing, not an actual fill, but it's the method doing its job: sell into fear, buy back once the fear has passed.
The verdict: the trend is back and the news is real. Half size now at $470. Let the rest come to me.
Personal framework, education only, not advice. Premiums are Black-Scholes estimates, not live quotes. Options carry risk, including assignment.
#AMD #EPYC #GoogleCloud #SellPut #Options
Comments