[Manuel Calavera]

Summary

-Nvidia hybrid ray tracing is a nice selling point, but performance is what's important - not $1,000 generation 1 tech. Don't drink the ray tracing Kool-Aid.

-Little competition in GPU means Nvidia can charge what ever it wants.

-Future Nvidia cards are experiencing some backlash due to prices. This presents AMD with an opportunity - IF - they can get a product out.

-AMD must step up its game in GPU (as it is far behind): Stop-gap products might just buy them the time to do so.

What can AMD (AMD) do to combat Nvidia (NVDA)? That is the question rolling around in my head. Clearly Nvidia's newest GPU RTX brings some unique features to consumers via hybrid ray tracing. Can AMD compete? Do they have a chance? What would/should/could AMD being doing to combat them? What problems might Nvidia face? Lots of questions. Let's speculate on answers and how we are playing it.

Ray Tracing

One of the strong points of the Nvidia RTX series is the hybrid ray tracing. Granted, this is generation 1 tech. Ray tracing reminds of me of when 3DFX introduced anti aliasing to gamers. At the time, the performance hit was immense and while it was neat - it was just not practical (if you wanted decent frame rates). Fast forward a few years and the tech has matured... the same might happen with ray tracing, but it is just not ready for - mainstream- prime time. Don't drink the Nvidia Kool-Aid... at least not until generation 2 tech hits.

Pricing

Even the highest Nvidia cards are only getting what passes as acceptable frame rates with ray tracing enabled (and some might even question that). However, ray tracing is a nice selling point and some gamers will indeed fork over premium prices for a premium experience. Obviously, the technology will improve with subsequent generations of development. AMD simply can't just allow this to pass. Yet... I do not think they have any plans to.

Future Tech

If we look at the DirectX and Vulkan, they both will support ray tracing and AMD is working on improving its Radeon Rays 2.0. Given the history of AMD, I suspect they are working on hardware-specific ray tracing but just how much energy they are putting behind it is the question. For that... we simply do not have an answer.

Energy

It comes down to a question of limited resources and energy. How many projects can you divide the GPU teams to work on? Currently, we know or suspect they are working on:
1. Vega 7nm for datacenter
2. New APUs
3. Navi for Sony Playstation (rumor)
4. Undisclosed GPU for Xbox successor (rumor)
5. Navi for consumer PC
6. Polar 500 series refresh (14nm to 12nm node update - once again a rumor) and a few I'm sure I'm forgetting.
The point being one only has so many resources to devote and AMD is putting those teams to work. On top of all of that, they have to do ray tracing, which brings us back to how much energy is going into it? Obviously, they have to devote some since DirectX and Vulcan are moving in this direction. Yet... does AMD really need it this generation? I do not think so. Honestly, I think it is a non-issue for this generation. It's a neat gimmick, but for the average guy it's simply not happening due to price considerations and performance. Which brings us to... what can AMD do to fight Nvidia?

Achilles Heel of Nvidia

The weak point of the Nvidia RTX series is simply the price points (from what we know). Per Digitrends:
"The Nvidia FE GPUs will launch at $1,200 for the RTX 2080 Ti, $800 for the RTX 2080, and $600 for the RTX 2070. Reference designs will be noticeably cheaper at $1,000, $700, and $500 respectively, but that is still much more expensive than previous generation cards have been. Equivalent 10-series GPUs debuted at $200-$300 less."
Since Nvidia arguably has no competition on the high end, they can charge what they want - they demand a pretty penny for the upper range of Nvidia cards. How AMD can combat this is to ignore the top $1,000 tier. If they focus on producing powerful, yet affordable, GPU solutions in the sub $200-$300 market they can capture market share with the vast majority of gamers.
The 500 series Polaris targets the $200-$300 range and is in dire need of a refresh. Moving from 14nm to 12nm could offer some performance boosts. Meanwhile, AMD should take its Vega 7nm 32 gig enterprise card and explore bringing it to the consumer market. While AMD will have to reduce the card from 32 gigs of HBM2 memory to a smaller amount, this might be a viable option for a stop -gap product as we wait for next-generation GPU Navi from AMD.
Navi is due sometime in the later half of 2019, but that's quite a ways out. These stop gap products will buy AMD time, but it's also a matter of cost from the prospective of AMD. Once again can AMD afford to tie up teams to work on various projects? So far, AMD has been silent on any stop gap plans.

How We Are Playing It

AMDs' run has been spectacular... parabolic comes to mind. With that said, we have shifted our money out of "in the money options" (taking profits) to "out of the money." We maintain long exposure via April $30 and $31 strike calls and have added February $30 calls. Come early October to mid October, we might see a pre-earnings accumulation. Right now, we are taking a cautious approach to AMD in the short term but the trend is your friend.
Long term, the story is intact. Rome and Ryzen 2 are due next summer. If AMD can fix its GPU division things will really start to take off.