The Crypto Hangover: AMD's 100% Rally Meets the Fed's Wrecking Ball
AMD doubled in 9 weeks on Ryzen and EPYC momentum, then round-tripped as crypto GPU revenue collapsed and the Fed crushed risk appetite.
The Setup
What the world looked like at entry
Executive Summary
In the summer of 2018, Advanced Micro Devices was the hottest stock in the semiconductor space. The Ryzen CPU lineup was stealing market share from Intel for the first time in a decade, EPYC server chips were landing deals with Amazon, Microsoft, and Tencent, and the stock had already doubled from its April lows. AMD closed at $16.50 the week of July 16 and proceeded to rocket to a weekly close of $32.72 by mid-September—a 98.3% gain in just nine weeks.
Then everything reversed. AMD's Q3 earnings on October 24 delivered a beat on the top and bottom line, but Q4 guidance disappointed badly. The crypto mining boom that had inflated GPU revenues through early 2018 had collapsed to "negligible" levels, and the market punished AMD with a 22% after-hours plunge. That earnings shock landed on top of the worst broad market selloff since 2011, as the Fed hiked rates in September and signaled more to come while US-China trade tensions escalated.
This case study follows a five-month trade that ended almost exactly where it began—AMD closed at $16.93 the week of December 17, a +2.6% return that masked a gut-wrenching round trip through a 98% gain and a 48% drawdown from peak. Meanwhile, the S&P 500 dropped 13.9% over the same period. AMD technically "outperformed," but anyone who held through the September highs and December lows experienced one of the most violent round trips in recent semiconductor history.
MACRO REGIME
- Fed Funds rate at 1.75–2.00% after June 2018 hike, with two more hikes expected before year-end
- US GDP growth strong at 4.2% annualized in Q2 2018
- US-China trade tensions building—first tariffs imposed in July 2018
- Tech stocks broadly in a bull market, NASDAQ hitting new highs
- Semiconductor cycle appeared healthy, with strong enterprise and data center spending
COMPANY SETUP
- AMD stock had surged from $9.56 (April low) to $16.50 by mid-July—a 72% move
- Q1 2018 showed crypto mining was ~10% of revenue, already declining from its peak
- Ryzen 2nd generation desktop processors launched April 2018 to strong reviews
- EPYC server processors gaining traction: HPE, Cisco, Tencent, and Baidu adopting
- CEO Lisa Su executing one of the most successful turnaround stories in tech
- The company was guiding for revenue acceleration in Q3 and Q4
SECTOR MOMENTUM
- Intel stumbling on 10nm process delays, giving AMD a rare window
- NVIDIA riding both gaming and data center AI demand to record revenues
- Crypto mining GPU demand collapsing—Bitcoin had fallen from $19,000 to $6,000
- Memory chip prices peaking, with early signs of cyclical downturn in DRAM/NAND
- Semiconductor stocks broadly outperforming the S&P 500 in the first half of 2018
SENTIMENT
- AMD was a retail investor favorite, heavily discussed on forums and social media
- Analyst consensus was divided—bulls pointed to Ryzen/EPYC share gains, bears warned about crypto GPU overhang
- Short interest was elevated, creating squeeze potential on positive catalysts
- Price targets ranged widely from $10 to $25, reflecting deep uncertainty about AMD's execution
Entry Point
The thesis and the position
A trader entering AMD in mid-July 2018 had a compelling growth story: Lisa Su's turnaround was delivering real results, with Ryzen taking desktop CPU share from Intel for the first time in years and EPYC starting to crack the lucrative server market. Intel's 10nm delays had created a once-in-a-decade competitive window. At $16.50, the stock had already run from April lows but appeared to have momentum heading into Q2 earnings on July 25. The risks were visible but easy to dismiss amid the euphoria. Crypto mining revenue, which had been 10% of AMD's top line in Q1, was evaporating as Bitcoin cratered. The Fed was tightening aggressively, and trade war rhetoric between the US and China was escalating—both threats to the global semiconductor cycle. And AMD's valuation was already stretched, pricing in years of server market share gains that hadn't materialized yet. Buying here meant betting that Ryzen and EPYC growth could more than offset the crypto GPU collapse, and that the macro backdrop would cooperate.
Before continuing: Consider what you would have done. Would you have taken this entry? What risks would you have been most concerned about?
The Journey
From entry to exit
Jul 16, 2018
Entry at $16.50, stock consolidating before earnings
Entry — Starting point
Jul 25, 2018
Q2 earnings beat: best quarter in 7 years, revenue up 53% YoY
Earnings — Stock surged to $18.94 weekly close (+14.8%)
Aug 22, 2018
Stock breaks above $20 as Ryzen momentum builds
Technical — Closed week at $23.98, up 45.3% from entry
Sep 5, 2018
Stock hits $29.94 intraweek high, crossing above $25
Momentum — Closed at $27.38, up 65.9%
Sep 12, 2018
AMD reaches $34.14 intraweek—all-time high territory
Peak — Weekly close $32.72, up 98.3% from entry
+98.3% from entrySep 26, 2018
Fed hikes rates to 2.00–2.25%, signals further tightening
Macro — Stock pulling back to $30.89
Oct 10, 2018
Broad market selloff begins; S&P 500 drops sharply
Macro — AMD falls to $27.35, down 16.4% from peak
Oct 24, 2018
Q3 earnings beat, but Q4 guidance disappoints on crypto GPU collapse
Earnings — Stock crashes to $17.63—down 46.1% from peak
Nov 6, 2018
AMD announces 7nm EPYC "Rome" at Next Horizon event
Product — Modest rebound to $21.03
Dec 19, 2018
Fed hikes rates again, signals more hikes coming
Macro — Broad market crashes; AMD closes week at $16.93
The Earnings Rocket (July 16 - August 2018)
AMD entered earnings season on July 25 and delivered a blowout: revenue of $1.76 billion was up 53% year-over-year, the best quarter in seven years. The Computing and Graphics segment hit $1.09 billion, up 64% on Ryzen desktop and mobile strength. EPYC server revenue jumped 50% sequentially. The stock surged from a weekly close of $16.50 to $18.94, then accelerated through August as the market digested just how fast AMD was taking share from Intel.
By the week of August 20, AMD had broken above $20 and kept going—closing at $23.98 for the week, a 45.3% gain from entry in just five weeks. Volume was explosive, averaging over 400 million shares per week. The narrative was simple and powerful: Intel was stuck on 14nm, AMD had the superior product lineup, and server market share was inflecting. Every data point confirmed the thesis.
The Parabolic Blow-Off (September 2018)
September was euphoric. AMD opened the week of September 3 at $25.62 and kept climbing. By the week of September 10, the stock touched $34.14 intraweek—more than double the entry price just nine weeks earlier. The weekly close of $32.72 represented a 98.3% gain from the $16.50 entry.
But the parabolic move was unsustainable. The stock had gone vertical on momentum and short-squeeze dynamics, not on fundamentals that justified a $32+ price. The week of September 17 brought the first cracks: AMD opened at $31.75 and closed at $31.02, failing to make new highs. The following week closed at $30.89. The Fed hiked rates on September 26 to 2.00–2.25% and signaled further increases. For a high-multiple growth stock like AMD, rising rates meant the market would discount those future earnings more aggressively.
The Double Crash (October 2018)
October was catastrophic—both for AMD and the broader market. The S&P 500 entered its worst month since 2011, falling from $290.72 to $265.33 as trade war fears and Fed tightening crushed risk appetite. AMD fell harder. The stock dropped from $30.89 to $27.35 in the first week of October, then to $26.34, then to $23.66.
The killing blow came on October 24. AMD reported Q3 earnings that technically beat expectations, but the Q4 guidance was ugly. Revenue was guided to approximately $1.45 billion, implying just 8% year-over-year growth after quarters of 50%+ growth. The culprit was clear: cryptocurrency mining GPU sales had collapsed to "negligible" levels after contributing 10% of revenue just two quarters earlier. Graphics revenue overall slid 14% sequentially. The stock plunged 22% after hours. By the week of October 22, AMD closed at $17.63—a 46.1% decline from the September peak and barely above the entry price. Five months of gains had been erased in five weeks.
The Failed Recovery (November - December 2018)
November offered a head-fake rally. AMD hosted its "Next Horizon" event on November 6, unveiling the 7nm EPYC "Rome" processor that would eventually become a game-changer for the server market. The stock bounced from $17.63 to $21.03 the following week. By late November, shares had recovered to $21.30 as investors bought the dip on the 7nm narrative.
But the macro backdrop was deteriorating. The Fed hiked rates again on December 19 and Chairman Jerome Powell signaled the balance sheet runoff was on "autopilot"—a comment that sent markets into a tailspin. The S&P 500 fell to $240.70 for the week of December 17, its lowest level of 2018. AMD, despite its company-specific growth story, got swept up in the broad liquidation. The stock opened the week of December 17 at $20.01 and collapsed to $16.93—wiping out the entire November recovery in a single week. The year ended with AMD essentially round-tripping back to entry.
Price Action
The trade in chart form
Results
The final accounting
During the same July 16 to December 24, 2018 period:
S&P 500 (SPY): -13.9% ($279.68 → $240.70)
AMD: +2.6% ($16.50 → $16.93)
AMD outperformed the S&P 500 by approximately 16.5 percentage points—but the headline number is deeply misleading. Anyone who held from entry through the September peak watched a 98.3% gain evaporate to just 2.6%. The stock's beta made it a leveraged bet on both directions: it outperformed dramatically on the way up and gave almost all of it back on the way down. The "outperformance" came entirely from the residual Ryzen/EPYC fundamental value that kept AMD from falling as far as the broader market in December.
Lessons
What the trade revealed
Parabolic gains demand a profit-taking plan
When a stock doubles in nine weeks, the risk/reward has shifted dramatically. AMD's move from $16.50 to $32.72 was a gift—but only if you locked in some of it. Trailing stops, partial sells at technical targets, or selling into strength are all tools that would have preserved meaningful gains from this trade.
Earnings beats don't protect you in a bear market
AMD beat Q3 estimates and still crashed 22% after hours because guidance disappointed. In a risk-off environment, the market prices forward earnings, not backward beats. When the macro backdrop is deteriorating, even good companies get repriced violently.
Hidden revenue dependencies create asymmetric risk
Crypto mining was 10% of AMD's Q1 revenue and collapsed to "negligible" by Q3. If a meaningful portion of a company's revenue comes from a volatile, unpredictable source, that's a risk that should be sized accordingly—not dismissed because the core business is strong.
Beta is not a free lunch
AMD outperformed the S&P 500 by 16.5 percentage points over this period—but delivered a total return of just 2.6% while subjecting the holder to a 48% drawdown. High-beta stocks can create the illusion of alpha when the real driver is just amplified market moves. Understand whether your returns come from stock selection or from leveraged exposure to market direction.
The best long-term thesis can still produce a terrible trade
AMD in 2018 was everything a growth investor wanted: superior products, gaining share, visionary CEO. The stock would eventually reach $160 by 2021. But from July to December 2018, the combination of crypto GPU collapse, Fed tightening, and trade war fears turned a correct long-term thesis into a round trip to nowhere. Being right about the company doesn't make you right about the trade.