NVIDIA Looks Expensive Until You Ask What Has To Ship
Consensus turns today’s multiple into a much lower one within two years, and by management’s own account the constraint on getting there is capacity rather than demand.
At about $217.56, NVIDIA (NVDA) trades at roughly 32.1 times its last twelve months of adjusted earnings. That multiple is what keeps some buyers out. The case for paying it turns on something less glamorous than the AI debate: whether the company can build and ship the AI data-center systems the forecast counts on, close to schedule.
The Multiple Consensus Says You Grow Into
Consensus earnings for the fiscal year now underway, 2027, put today’s price at about 24.2 times, and the fiscal 2028 estimate at about 17.1 times. Against today’s trailing 32.1 times, that is a 47% lower multiple with no move in the price at all. One caution: the trailing 32.1 times sits on adjusted earnings with stock-based compensation added back, a basis near the analysts’ own but not defined identically, so some of that drop is a change of yardstick rather than of earnings.
The Forecast Rests On How Much Gets Built
Getting to that lower multiple needs revenue up about 49.1% a year across the two years, with earnings growing more slowly, about 37.0% a year. Profit growing slower than sales is a forecast of thinner margins, and from a high base: 64.0% at the operating line over the last twelve months, against a three-year average of 56.6%. The revenue rate is not heroic against what the business has just done: revenue grew 70.7% over the last twelve months on $253.49 billion of revenue. Revenue growth of that order is one of the marks of quality the Trefis High Quality Portfolio favors.
On the company’s own account at its May results call, whether that revenue exists is a question of capacity, not orders. It raised total supply, including inventory purchase commitments and prepayments, to $145 billion, and in July agreed to prepay Amkor to expand U.S. advanced packaging capacity. Blackwell systems have been shipping; management expects to be supply constrained throughout the entire life of Vera Rubin, whose production shipments start in fiscal Q3 2027.
Where Thirty Analysts Stop Agreeing
The fiscal 2028 earnings estimate is a consensus of thirty analysts whose forecasts span $9.81 to $14.85 a share. Against the roughly $9.00 expected for fiscal 2027, the low end is a second year that adds about 9% to per-share earnings, while the high end is a different company.
About $263 Is A Scenario, Not A Target
Suppose the market pays about 20.6 times the fiscal 2028 consensus, roughly halfway between the fiscal 2027 and fiscal 2028 multiples: the stock would be worth about $263, or about 21% above today’s price. That is a scenario and not a target, because it requires the market to keep paying a higher multiple than the 17.1 times an unchanged price would leave, as the earnings arrive, and this stock has lost as much as 84% peak to trough in past market shocks. So you are buying a delivery schedule rather than a view on AI, and the schedule shows in the guide first: fiscal Q2 2027 revenue was guided to about $91 billion, up from the $78 billion guided for fiscal Q1 2027, with those results due on August 26. The case holds only while that outlook keeps being raised.
A Two-Year Shipping Forecast Is A Lot To Rest On
The valuation case here holds together, but only if two years of production land near plan. Worth part of your money, then, and not all of it. The compounding happens in a rules-based portfolio of quality businesses, not in one company’s build schedule. That portfolio has a track record of outpacing the three major indices – the S&P 500, S&P Mid-cap, and Russell 2000.