QE, inflation, technology, and energy

Inflation numbers, this week, came in below expectations again. They are still below the FED’s 2% target.

I have previously discussed my views on how technology and energy trends contribute to deflation. How’s that? Technology is advancing exponentially and this contributes to make things more efficient and cheaper. Just look at NVDA’s past product roadmap. Four years ago, NVDA’s GPUs were 10x better than CPUs. In the past 4 years, NVDA’s GPUs computing power has advanced by 100x which is more than a 3x improvement per year. That’s 1000x improvement over CPUs from 4 years ago!!! Computing power, therefore, is now advancing at a higher rate than Moore’s Law. A GPU computer (DGX-1 which costs $149,000) with 8 GPUs can now compute a petaflop. What does this mean? A petaflop is 1 quadrillion operations per second… that’s 1,000,000,000,000,000. There are only 20 supercomputers in the world that have more computing power and I bet those computers cost many millions of dollars and suck a ton of energy to run. NVDA’s new V100 GPUs which go into the DGX-1 but are also sold as one-offs are largely run in data centers. So this product will run many of these world’s computing operations and contribute to make computing much cheaper. It will also give computing more superpowers in the area of AI. This enables companies to make sense of data and make better business decisions which will increase the efficiency of economic output; this is also deflationary. Competition between the various cloud providers (e.g. Microsoft and Amazon) will ensure that these lower compute costs will be transferred to the end users. When prices drop, adoption generally rises so we can expect continued and increasing utilization of computing power. Companies around the world will use computing power for more and more applications…applications that were previously too expensive will now be economically viable. In addition, AI applications will be increasing applied. Speaking of AI technology, it has also be doubling every year. NVDA’s CEO yesterday spoke of this “double exponential” but I wonder if people realize how world changing this is and will continue to be.

Technology is also contributing to making energy production less costly. Solar energy cost has been reducing steadily for 30 years. The sun is a free source so the cost of energy production is approaching zero or free. Each year the returns will diminish as you approach zero but the trend is still deflationary as energy is needed to do all work…it just boils down to physics.

So we have 2 deflationary trends that most people still underestimate. People look at the past and try to project the past trajectories into the future. But this becomes difficult when you are in an exponential trend that has not yet reach the hockey stick inflection point. I believe that this is happening with interest rate projections. People think that the Fed will continue to raise interest rates. People think that QE will cause super high inflation unless the Fed raises rates. Well, so far the Fed and the other world central banks can’t really raise rates. People assume that they are being cautious as to not induce a recession by raising too quickly. There may be some truth to that but I think most people miss the larger deflationary trend that technology advances and energy efficiency are having on economies. I will make another observation: when you drill down to the core enabler (or ground zero of technology advancement) of technology advancement it is computing speed increase coupled with computing cost reduction). NVDA is the one company that is currently contributing the most to technology advancements because computing speed (and now combined with AI advancement) is enabling unheard-of technology advancements by others.

So assuming that the above is true what does it mean?

  1. might be smart to be invested in ground zero of tech advancement enablement.

  2. interest rates might be low longer than current expectations; this is a headwind for banks although there are certainly other tailwinds for banks; bonds will likely to continue to underperform.

  3. a low interest rate environment encourages investment in projects so we can expect higher economic output.

  4. global economic output should continue to advance.

  5. even if the above are true there are occasional shocks that can cause disruptions to the trends discussed here; however, these shocks should be transitory and the general path should be higher but perhaps with some bumps along the way.


Absent massive government deficit spending, technology insures deflation. Stuff cost less. While there are few trends today equalling electrification or the ICE vehicle, there are more of them and the pace of adoption is constantly increasing.
Add in the near bank bust and the huge debt load of consumers (present purchases at cost of future purchases) .

Every experiment by meddling Federal Reserve has a down side. Those of QE have yet to be seen. They may show up soon or not for decades .Japan has shown me that economies can survive modest deflation high debt and low interest rates without collapse.

Ways to reduce debt

grow out of it
pay it off
inflate it away
destroy the system and start all over again (Russian Revolution)

Can the Fed actually do any of these? Do they want do any of them except growth ?

Traditional economists are convinced around 3% inflation is a necessity. Whether they are right does not matter as long as they have policies favoring that.
The Fed has no reason to to raise interest rates as long as inflation is controlled and employments high The only reason they are making some feeble attempts is so they will have some leeway in the future to lower again