Is AMD a Buy After the $1,250 Call? That's $2 Trillion

TL;DR

• Yes at $514.39, and on management's number rather than Baird's. AMD closed Friday up 6.50% at $514.39, its best session in weeks, after Baird's Tristan Gerra doubled his target to $1,250 from $625 and the company priced its largest ever bond.
• The widely quoted 159% upside was struck against Thursday. From Thursday's $482.99 close, $1,250 is +158.8%. From where the stock actually finished, it is +143.0%.
• $1,250 is a $2.04 trillion market cap on AMD's 1.63 billion shares, against $839.73 billion today.
• The target needs about $27 of earnings, not $20. AMD trades at 46.3x forward. Hold that multiple and $1,250 requires $26.99; Lisa Su guided 2027 EPS "significantly above $20", which at $1,250 would be 62.5 times.
• AMD raised $4.75 billion of notes on August 13, roughly doubling total debt to $9.03 billion. Against $13.11 billion of cash it stays net cash, which is why the market treated this debt differently from the $370 billion attached to Broadcom's financing platform on the same day.

Is AMD a Buy After Baird's $1,250 Target?

Yes, at $514.39, on a view that runs to the end of 2027. My own frame is $800, which is management's guided $20 of 2027 earnings at 40 times, or about 56% above Friday's close. That is a target I can defend line by line. I cannot defend $1,250.

Start with what happened. Baird's Tristan Gerra doubled his price target to $1,250 from $625 and kept an Outperform, taking the Street high. Separately, the company priced $4.75 billion of senior notes, its largest US dollar bond ever. The stock rose 6.50% into a tape where the S&P 500 fell 0.17% and Broadcom dropped 5.92%.

The Board

The target, the multiple it needs, and the balance sheet underneath it.

The 159% Belongs to Thursday

Most of the weekend coverage carries an upside figure near 159%, and the arithmetic behind it is fine. It is just a day old. $1,250 against Thursday's $482.99 close is +158.8%. Against Friday's $514.39, the same target is +143.0%.

Fifteen points of headline upside disappeared into the session that the note itself helped cause. Anyone reading the number on Saturday and buying on Monday is buying the smaller one.

What $1,250 Costs in Earnings

Two ways to price the target, and they land in different places.

The first is the multiple. AMD trades at 46.32x forward earnings at $514.39, which implies a forward number near $11.10. Keep that multiple and $1,250 needs $26.99 of earnings.

The second is the guide. On the Q2 call Lisa Su said AMD is tracking materially ahead of its long-term model, with 2027 data center revenue more than doubling, server CPU growth above 70%, and annual EPS "significantly above $20", earlier than the Financial Analyst Day plan. Take $20 literally and $1,250 is 62.5 times it. Take $23 and it is 54 times.

So Baird's target does not merely require the 2027 doubling to happen. It requires the doubling, then a further beat of roughly a third above the guided EPS, or a multiple that stays near today's after two years of the growth being delivered rather than promised. Gerra's model reportedly gets there via $147 billion of AI GPU revenue by 2030 at about 15% share, which implies a roughly $980 billion accelerator market that year. AMD's entire trailing twelve-month revenue is $41.3 billion.

I am not calling the number wrong. Nobody can falsify a 2030 TAM from here, which is exactly the problem with underwriting a share price off one.

AMD Borrowed $4.75 Billion in the Same Week

The bond deal is the more interesting event and it got a fraction of the attention. AMD priced four tranches on August 13: $1.25bn of 4.600% notes due 2029, $1.5bn of 5.000% due 2031, $1bn of 5.250% due 2033 and $1bn of 5.500% due 2036, settling August 17. Demand was strong enough that pricing tightened about a quarter point during the day.

Set that against AMD's balance sheet. Total debt was $4.28 billion and cash $13.11 billion. The deal roughly doubles the debt to $9.03 billion and leaves the company net cash by about $4 billion, funded at investment-grade coupons on maturities anyone can read.

Broadcom fell 5.92% the following day on a Bank of America estimate of debt inside an AI financing vehicle. Same theme, opposite reaction, and I think the market got the distinction right: a disclosed corporate bond with a coupon is a different instrument from an off-balance-sheet platform whose size is an analyst's projection. What it does mean is that the AI buildout has moved from being funded out of cash flow to being funded in the credit market, and AMD is now a participant in that. The same rotation is visible across the AI supply chain.

The Part I Would Not Underwrite

The bull case is real and it is in the numbers AMD published on August 4. Revenue of $11.536 billion, data center revenue of $6.718 billion up 107% year over year and now 58% of the company, and a Q3 guide of about $13 billion plus or minus $300 million at a 56% non-GAAP gross margin. Those figures come straight from AMD's own results release.

The problems sit either side of it. Gaming fell 31% year over year to $779 million, so the non-AI business is not carrying anything. The stock has run 178.92% in 52 weeks and sits 58% above its 200-day average of $324.33, though at a 50-day average of $510.40 it has gone roughly nowhere for two months. And at a 20.3x price-to-sales multiple with a beta of 2.49, a disappointment in the 2027 doubling does not cost a few multiple points.

My invalidation is specific: a Q3 print in late October with data center revenue below roughly $7.5 billion, or any walk-back of the 2027 doubling, and the $800 frame goes with it. The nearest outside read on the same demand curve is Nvidia's quarter on August 26.

The Options Angle

No live chain could be sourced on a Saturday, so nothing below is priced off a quoted premium. The structural point does not need one. At $514.39 a single options contract carries about $51,400 of notional, which puts most defined-risk structures on this name out of reach for a normal account and makes the shares the practical way to hold the view. Buying calls the session after a 6.50% gap also means paying an implied volatility that just repriced upward.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Conviction	Breakeven
1	Bullish	Long common stock	No expiry; scored on the Q3 2026 reaction session	n/a	$514.39, Aug 14 close	n/a	6/10	$514.39
2	Pass	Long calls into the gap	No live chain sourced	n/a	$514.39, Aug 14 close	not sourced	3/10	n/a

Both rows sit on the Track Record ledger against the SPY close of $776.34, so the call is scored against the index rather than on its own.

The One-Line Read

AMD is worth owning at $514.39 for the 2027 data center doubling. It is not worth owning for $1,250, which needs a beat on top of the guide and a $2 trillion valuation to arrive on schedule.