This is a discounted cash flow pricing tool which uses projected earnings to help us calculate an implied current stock price based on the stated assumptions about discount rates. The same tool can be used in reverse, to calculate the future earnings implied by a given stock price.
The graph below shows actual earnings up to and including 2025, as well as the projected earnings for 2026 and the earnings growth beyond that implied by the current stock price. The reason we have included 2026 as a fixed number despite the fact that these earnings cannot be fully known, is that given the nature of NVIDIA's business, where orders are taken months in advance, we have much more information about these earnings than those for 2027 and beyond.
As you can see below, NVIDIA's earnings have been growing at a phenomenal rate since the take-off of AI in 2023. When doing a discounted cash flow model, which is what this is, you must assume that earnings will normalise at some point. If you fail to do so you will not get a sensible stock price, and also, it simply doesn't make sense – if a company's earnings keep growing at 50% a year forever, eventually they would outgrow the entire global economy.
You can use the tool yourself to see what the stock should be worth based on your own guesses for future earnings. To start with, enter the period of years you believe it will take for earnings to normalise to the long-term growth rate. You can then override earnings growth (or any of the other model inputs) manually, so the tool gives you full control!
We've also included the implied price based on our own estimates, and finally we have included a growth path based on the revenue guidance for 2026 and 2027 given by the company's CEO, Jensen Huang, at their recent annual conference.
Notes
- To start, we need to establish a reasonable discount factor for NVIDIA. It is a massive company with an established market lead in its area and clear visibility on orders months, if not years, ahead. However, it is a one-sector company that has only been successful for a short period of time. Therefore, it should not trade for the same kind of return as an ultra robust company that has stood the test of time like McDonalds or Coca Cola. These trade at prices implying roughly 6% discount rate. NVIDIA probably belongs closer to something like an investment bank, where prices typically imply a 10% discount rate.
- For GA's numbers we are working off the assumption that AI capex growth will slow to 30-50% in 2027 and 25% in 2028. After that we normalise to the long-term rate of 3 years. This, we believe, is reasonably conservative, and there is risk to the upside if the capital shortfall is solved. On the downside, we believe the major risk is to Nvidia’s market share or margins, both of which are very high and will almost certainly diminish in the long term, but we don’t see significant drop off in the foreseeable future as they keep pushing the frontier of their technology faster than the others can catch up.
- For Jensen’s numbers, we are referring to the quote “At this time last year, I mentioned that we saw $500 billion in high-confidence demand, covering Blackwell and Rubin through 2026. Now, right here and now, I see at least $1 trillion in demand by 2027.” We take this and apply an 8% uplift to convert to earnings, as this has been historically the relationship between revenue growth and earnings. For 2028 onwards, we again apply a 3 year normalisation period to the long-term earnings growth rate.