AMD stock surged from $4.43 to $128, validating a contrarian bet

The €4,000 AMD bet no one wanted to hear

In January 2016, AMD shares traded at $1.86. Five months later, they hit $4.43. With that 2.3x gain under their belt, a retail investor put €4,000 on the table and published a thesis that sounded like a bad joke: Advanced Micro Devices would double or triple its value. They warned, debated, and endured the mockery. The forecast came true.

The discussion that peine those days wasn't about pretty charts, but a valuation anomaly. AMD had a market cap of $3.5 billion versus Nvidia's $25 billion and Intel's $150 billion. Two giants and a dwarf with similar technology in hand. The question was how long it would take the market to realize it.

Why would anyone believe AMD would double?

The argument had three pillars. First, the Radeon RX 480, launching for less than €200 with 5.5 TFLOPS against Nvidia's equivalent proposal's 6: two in CrossFire for €400 were said to outperform an $800 GTX 1080. Second, the arrival in October of the Zen processor, which for less than €300 matched an Intel 5960X costing €1,000. Third, control of the console market and the business of refreshing the current generation. There was even an external element: Google was going to use AMD Radeon cards in its cloud computing services. Such a contract isn't signed out of charity.

The skeptical response was equally forceful and also brought numbers. Teraflop accounting doesn't measure real gaming performance: the R9 390 delivered 5.1 TFLOPS and the 390X 5.9, but neither approached a GTX 980 Ti declaring 5.6. Less so when the circulating benchmark came from a title heavily optimized for the manufacturer. The technical debate remained there, unresolved.

Price moved before products

By July 2016, the stock already traded above $7. By December, it had doubled its value, exactly as predicted. Along the way came the usual noise: the effect of a capital increase that, according to one participant, diluted shareholders, doubts about whether new-generation consoles would arrive strongly, and a simultaneous 36% rise in Nvidia that dragged AMD up 9% in a single day. Some entered late, at $6.75, after others pointed to $4.80 as a key level to break.

From the $32 peak to $17

The second part of the story is what almost no one remembers. The stock surpassed $30 and there were staged exits: sales at $32, others at $29.31 with an average purchase price of $24.63. Someone warned that what goes up fast comes down fast, and they were right.

With the earnings report came the halt. Quarterly results missed expectations and the stock traded at $17.70. A participant summarized the situation with devastating fundamental analysis: a P/E ratio of 74 and EBITDA of $666 million in the twelve months until June 2018 do not support a skyrocketing market cap. The correction from $33 was 30% in one month. Multiplying profits by 21 to justify a rise to $60 was not a realistic scenario, the accounts said.

The business not in the script: servers and cryptocurrencies

One of the most lucid analyses pointed to another market: ARM processors and mobile devices, where real growth lay, while PC and console hardware had peaked. One participant argued that the end of the cryptocurrency fever, which had inflated graphics card sales, had hurt them. The server business, where the real money is, remained the big pending trump card.

And now?

AMD trades around $128. The company that seemed doomed to be the cheap rival changed its script with Lisa Su at the helm and a roadmap that, according to forum users, left Intel without an answer: its 7-nanometer lithography arrived before the competitor's 10 nm, burdened by rumors of insufficient yields. The Ryzen 7 3700X with 16 cores and 32 threads at 7 nm and a GPU with 20 TFLOPS and HBM2 memory are the arguments for the new phase.

But the disorienting data point isn't in AMD. Whoever bought Nvidia shares at $13 in 2016 has nearly $475 per share today. The winning bet was the one no one questioned.