“Two coffees” sounds like a record. It is barely a description. Two small espressos and two large brewed coffees can belong to the same daily habit while representing very different amounts of caffeine.
Even an exact menu name may not resolve the number. When we checked Starbucks' US menu on November 18, 2026, a Grande Pike Place Roast listed 315 to 390 mg of caffeine. Reporting only 315 mg would discard information the company actually provides. Reporting the midpoint as a measured fact would invent precision.
We collected four standard menu configurations and calculated what their published values imply. The useful question is not which drink sounds stronger. It is what was ordered, how much was consumed, and how uncertain the estimate remains.
What the menu says today
These are US menu values, with no customizations and no particular cafe selected. Starbucks describes caffeine values as approximate. Its nutrition URLs can open at a default size, so match the size selector to the row before checking a number.
| Drink | Displayed size | Published caffeine | Calculated mg per US fl oz |
|---|---|---|---|
| Pike Place Roast | Tall, 12 fl oz | 235 to 290 mg | 19.6 to 24.2 |
| Pike Place Roast | Grande, 16 fl oz | 315 to 390 mg | 19.7 to 24.4 |
| Pike Place Roast | Venti, 20 fl oz | 390 to 490 mg | 19.5 to 24.5 |
| Signature Espresso | Doppio, 1.5 fl oz | 150 mg, approximate | 100.0 |
Download the four-row CSV or JSON with the calculations and order examples.
The espresso is more concentrated by this calculation, yet the Tall brewed coffee has more caffeine in the complete listed serving. Neither observation says how a particular person will feel. It does show why “stronger coffee” is an unhelpful substitute for the amount in the drink.
The range belongs in the record
A Grande's published interval spans 75 mg. That difference is not a measurement we made, and the menu does not establish a probability distribution within it. We cannot say that the midpoint is most likely or that the endpoints are laboratory guarantees.
What we can say is that a record of “Grande Pike Place, approximately 315 to 390 mg, US standard menu, checked November 18” contains more useful information than “coffee, 100 mg.” A neat single number should not outrank a better description.
The Doppio's single displayed figure needs the same restraint. The site calls it approximate too. In the download, it remains a point estimate with unknown numerical uncertainty; we do not give it a fabricated zero-width confidence interval.
Two possible orders, calculated
Suppose someone finishes a Tall Pike Place in the morning and a Grande later. Adding the lower endpoints gives 235 + 315 = 550 mg; adding the upper endpoints gives 290 + 390 = 680 mg. That is a calculated total of published ranges for this hypothetical order, not a measured daily intake or a confidence interval.
A Grande plus the standard Doppio produces 465 to 540 mg when the approximate 150 mg espresso figure is added to each endpoint. Its unquantified uncertainty still exists. The calculation does not make it disappear.
These examples are not intake recommendations. Their purpose is to make a record auditable: another reader can see which drinks, which servings and which assumptions produced the total. They also show how little a count of “two” tells us.
An order is not necessarily what you drank
If half a Grande is left in the cup, the full-serving menu value overstates consumption under a simple proportional model. Exactly half of a uniformly mixed serving would imply 157.5 to 195 mg from the published range. That is another calculation, not evidence that anyone drank exactly half or that a cafe prepared the drink to an exact composition.
Record the remaining amount as an estimate when that is what it is. If you added shots, chose another roast or changed the recipe, return to the matching menu configuration. Do not assume that a size-only adjustment captures every customization.
What a useful coffee log keeps
For later interpretation, keep the drink name, size, recipe changes, approximate amount consumed and time. Preserve the source of the caffeine value and whether it is a range, an approximate point or unknown. A blank value is preferable to a confidently invented default.
Those details can help you ask better questions about your routine. They do not establish that coffee caused a particular sleep score, nor can this four-item menu check identify a safe amount for an individual. We have not conducted a sleep experiment, caffeine assay or clinical assessment.
Our original work is the dated menu collection, unit comparison and reproducible interval arithmetic. It is deliberately small enough to check. Values apply to these configurations, not every Starbucks country, bottled product or home brew. Reuse our calculations with attribution and their limitations; check your current menu when ordering.
From the publisher: Vora connects nutrition, workouts and wearable context with proactive AI. We care about records that help people understand their days. The text logging guide covers the supported messaging setup; this reference does not claim a dedicated caffeine analysis feature.