AMD Stock Forecast: AMD Plus Tesla Could Be A Championship Team For Self-Driving Cars

motek 1The article was written by Motek Moyen Research Seeking Alpha’s #1 Writer on Long Ideas and #2 in Technology  – Senior Analyst at I Know First

AMD Stock Forecast


  • Intel recently announced that it is working with Google’s Waymo to develop level 4/level 5 autonomous vehicles. Intel’s stock did not benefit from this official announcement.
  • CNBC reported that Waymo’s rival, Tesla is working with AMD to develop its custom AI processor. AMD’s stock rallies +6.4% during after-hours trading after CNBC’s report.
  • Nvidia is currently supplying the Tegra processor for Tesla’s self-driving car test unit. AMD probably secretly helping Tesla build its own Autopilot AI processor is a huge confidence booster.Unlike Nvidia, AMD can create AI processors using x86, ARM, and GPU technologies. AMD’s ex-employee Jim Keller is Tesla’s head of its Autopilot program.
  • Unlike Nvidia, AMD can create AI processors using x86, ARM, and GPU architectures. AMD’s ex-employee Jim  Keller is Tesla’s head of its Autopilot program.
  • Supplying artificial intelligence processors for self-driving car is one way AMD can benefit from the fast-growing AI chipsets market.

I wrote yesterday about how Intel (INTC) and Google’s (GOOG) Waymo subsidiary are collaborating to develop level 4/level 5 autonomous cars. CNBC reported today that Waymo’s rival self-driving car developer, Tesla (TSLA) is working with Intel’s rival, Advanced Micro Devices (AMD), to develop its own artificial intelligence [AI] processor. This report has fueled a+6.40% rally for AMD’s stock during the after-hours trading. AMD’s stock already closed with a +4.73% gain after Morgan Stanley echoed my bullish opinion last week over Amazon’s (AMZN) selection of AMD server GPUs.

TESLA logo

(Source: AMD/Tesla)

INTC did not benefit from Mr. Krzanich’s blog post two days ago announcing Intel’s partnership with Waymo. While an unverified CNBC report saying AMD is helping Tesla’s self-driving car efforts resulted in AMD’s stock receiving a huge boost during the after-hours trading. It is amusing that I profit more on my AMD bet than on my INTC investment.

AMD stock price

(Source: Google Finance)

Intel predicted a $7 trillion economy out of autonomous vehicles by 2050. It looks like AMD is going to be a part of it.

Why There Might Be Some Truth To CNBC’s Report

Some things are supposed to be trade secrets but there are people who love leaking classified information. Unless Dr. Su or Jim Keller calls/email me denying the story, I cannot dismiss outright the veracity of CNBC’s scoop. Further, I truly see Tesla and AMD both benefiting from this alliance.

The head of Tesla’s Autopilot program is Jim Keller. Jim Keller is a legendary CPU architect that used to work for Apple (AAPL) and AMD.  Keller left AMD in late 2015. After which, Tesla obviously hired Keller away from AMD in January 2016 to develop the processor or the CPU for its Autopilot program.  Keller is also likely responsible for Tesla allegedly poaching CPU engineers from AMD last year.

Before he joined Tesla, Keller and his team were responsible for developing AMD’s 14-nm Zen CPU products. AMD and EPYC processors (which are responsible for AMD’s 2017 rally) are the commercial implementations of Keller’s Zen CPU architecture designs. Deductive reasoning says Keller and his team would take inspiration from Zen CPUs for its future custom AI processors.

Further, Tesla and Keller’s problem right now is they don’t have the patents or the original IP to create their own processors. AMD or Intel will probably sue anybody creating patent-trampling processors without their approval.  Tesla/Keller obviously cannot go to Intel since it is already working with Waymo. Tesla and Google are fierce rivals in the race to develop the first commercial self-driving car. Sourcing out parts/technology from the same partner of Waymo will not help Tesla differentiate its autonomous car project from Google’s.

AMD, which touts x86 CPU, ARM SoC, and GPU patents/IP assets, is the obvious partner for Tesla’s in-house AI processor. No matter how brilliant Keller is, he still needs AMD’s patents and permission to deliver working AI processors for Tesla Autopilot-enabled cars.

What’s In It For AMD

There is little immediate impact to the topline/bottomline of AMD being picked by Tesla for its Autopilot AI chip.  It is only an investor confidence booster. The stock jumped after hours after CNBC reported this development. Tesla hiring AMD to help it develop its in-house processor while Tesla Autopilot is currently using Nvidia Drive PX 2 is a bit devious on Keller’s part. However, all is fair in love or war. Tesla is at war/competition with Intel/Waymo over the autonomous car industry.

Going forward, AMD could really benefit if it can compete with Nvidia and Intel on artificial intelligence processors for the self-driving vehicle market. AI chips for self-driving cars are part of the fast-growing niche market for artificial intelligence processors. As per the estimate of Markets&Markets, the AI chipsets industry is growing at CAGR of 62.9% from 2016 to 2022. Five years from now, the AI chips market will be worth $16.06 billion.

A Tesla Autopilot AI chip could a system-on-chip [SoC] design using a custom EPYC and Radeon Radeon Instinct Mi25. Tesla’s Autopilot add-on option can cost up to $8,000. AMD can charge Tesla $1,500/$2,000 each for its custom AI chip for Tesla cars and Keller won’t complain too much.

Supplying $2k AI custom chip to a car company that can sell 150k units/year (my guesstimate of Tesla’s future performance after Model 3 goes to full production) of Autopilot-enabled electric cars is a future tailwind for AMD. It can contribute $300 million of new annual revenue for AMD.

Why Tesla Needs Its Own AI Processor

After its ugly dispute with Mobileye last year, Tesla sourced out its self-driving car processor from Nvidia (NVDA). A Nvidia + Tesla partnership can still work. However, there is no guarantee that Tesla won’t incur future disputes against Nvidia. It is, therefore, safer for Tesla to just build its own in-house processor for its Autopilot cars.

However, a Tesla + AMD tandem can also compete against an Intel + Waymo tag team on autonomous vehicles. AMD has a good record of accomplishment of supplying custom chips for game consoles and other electronic products. It has the experience and foundry partners to quickly ramp up Tesla’s custom Autopilot AI processor.

Unlike Nvidia, AMD has the license to create x86-based, ARM-based, and GPU-based AI processors that can be used to process sensor data, autopilot management, connectivity, and infotainment compute needs of Tesla’s future level 4/5 Autopilot-enabled cars.

Tesla Can Do The Apple Trick

Tesla can keep using Nvidia Drive PX 2 for its experimental autonomous cars while simultaneously developing its own AI chip with the help of AMD. Apple (AAPL) did this clever trick of licensing and using mobile GPUs (Graphics Processing Units) of Imagination Technologies for many years. Once Apple learned enough knowledge, it decided earlier this year to just build its own mobile GPU.

I speculated before that it could have been AMD that gave the patent rights and assistance which allowed Apple to create its own mobile GPU. I’m also convinced that AMD could help Tesla with its Autopilot AI processor. Building an AI chip is more complicated than building an integrated GPU/CPU mobile SoC. But like I said earlier, AMD and Keller (and his Tesla team) can learn a lot from studying Nvidia’s Drive PX and Mobileye’s EyeQ 5 self-driving car processors.

There is no need to hasten things too. Government regulators are going to be strict over level 5 or fully-autonomous self-driving cars. It also took Apple almost 10 years to escape from its dependence on Imagination’s GPU patents/designs.

Final Words

Self-driving cars are also, in a sense, datacenter on wheels. Consequently, my fearless forecast is that the Infinity Fabric interconnect technology is AMD’s killer advantage against Intel and Nvidia. Faster inter-die data processing exchange between a car’s semiconductor components should improve autonomous cars.

I have faith in Jim Keller’s cognitive ability. Collaborating with AMD can obviously accelerate development of his AI processor for Tesla Autopilot. I don’t care if CNBC’s report was fake news or a golden leak of secret information. I’m just thankful it helped boost AMD’s above $14.60 today.

AMD needs new markets for its CPU and GPU products. Tesla’s need for Autopilot processors could someday become an important tailwind for AMD.

I am long AMD and I humbly rate this stock as a buy. My position is contrary to the negative algorithmic forecasts of I Know First for AMD. Sometimes when it comes to stock picking, I still do favour my gut feeling/God-given intelligence over artificial intelligence. You can wait for a cheaper entry point but AMD is a long-term winner to me right now.

amd stocks

Past I Know First Success With AMD

I Know First has been bullish on ADSK shares in past forecasts. On January 20, 2017, an I Know First Analyst released a bullish article for AMD. It was about the company’s positive trend, encouraging key statistic and how AMD’s latest products could result in increasing market share. Since the forecast’s release, AMD shares have gained 36.13%. See chart below.

AMD Stock

(Source: Google Finance: AMD)

This bullish forecast for AMD was sent to I Know First subscribers on January 20, 2017. To subscribe today click here.

I Know First Algorithm Heatmap Explanation

The sign of the signal tells in which direction the asset price is expected to go (positive = to go up = Long, negative = to drop = Short position), the signal strength is related to the magnitude of the expected return and is used for ranking purposes of the investment opportunities.

Predictability is the actual fitness function being optimized every day and can be simplified explained as the correlation based quality measure of the signal. This is a unique indicator of the I Know First algorithm. This allows users to separate and focus on the most predictable assets according to the algorithm. Ranging between -1 and 1, one should focus on predictability levels significantly above 0 in order to fill confident about/trust the signal.

autodesk stock forecast

autodesk stock forecast