The right answer to this question comes out of a Manual J load calculation performed on your specific house, and there is no shortcut to it that works. Not square footage. Not the size of the unit currently sitting on the pad. Not the number a neighbor with a similar floor plan got quoted. Those shortcuts all exist, they all circulate freely, and in Florida they fail in the same direction: they tell you to buy a bigger system than you need.
Which is the expensive error, not the safe one.
Why too much capacity makes a Florida house worse
An air conditioner does two jobs in this climate, and they do not happen at the same speed. It lowers air temperature, which happens fast, and it removes water from the air, which happens slowly. Moisture only comes out while refrigerant is actually flowing and the coil is actually cold, and it takes several minutes of continuous operation before a coil gets cold enough and wet enough to condense water at any real rate. Sensible cooling starts almost immediately. Latent cooling has a warm-up period.
Now picture an oversized system on a July afternoon. It fires, drives the air temperature down hard, hits the setpoint in seven minutes, and shuts off. The coil never got fully wet. Very little water left the house. Twenty minutes later it does the same thing again, and again, all day, and at the end of that day the thermostat reads 74 while the house sits at 62 percent relative humidity and feels like a basement. The owner, reasonably, concludes the AC is not keeping up and drops the setpoint to 71. Now the bill is higher and the house still feels wrong, because the problem was never temperature.
Short cycling costs you mechanically as well. Startup is the hardest moment in a compressor's cycle, drawing several times running current, and a system that starts forty times a day instead of fifteen ages accordingly. Contactors pit faster. Capacitors see more thermal cycling.
Undersizing is real too. It is the failure people fear, and an undersized system genuinely cannot hold the setpoint through the hottest afternoons in August, running continuously and losing ground while it tries. The difference is that undersizing announces itself immediately and honestly, and it is comparatively rare, because nobody's instinct is ever to buy too small.
What Manual J actually accounts for
Manual J is the residential load calculation procedure published by the Air Conditioning Contractors of America, and it is the industry standard for a reason: it models heat gain into your house rather than guessing from its footprint. A proper one is done room by room and accounts for these things.
- Window area, orientation and glazing type. The largest single variable in most Tampa Bay houses. West-facing glass takes full sun during the hottest hours of the day, and the difference between clear double-pane and low-e coated glass on that same wall changes the load meaningfully. A calculation that does not ask which way your house faces is not a calculation.
- Insulation levels in walls, ceiling and floor, plus the assembly type. Block construction with interior furring behaves differently from frame.
- Infiltration. How much unconditioned outdoor air leaks into the house, which in Florida is a latent load, not just a sensible one, because that air arrives carrying moisture.
- Duct location, condition and leakage. Ductwork running through an attic that reaches 130 to 150 degrees loses capacity between the air handler and the register, and leakage on the return side actively pulls that superheated attic air into the system. See ductwork.
- Ceiling height and conditioned volume. Two 2,000 square foot houses, one with eight-foot ceilings and one with vaulted great room, are not the same house.
- Internal gains and occupancy. People, cooking, lighting, a home office full of equipment.
- Shading. Overhangs, mature trees, a neighboring two-story house on your west side. All of it counts.
- Local design conditions. The outdoor temperature and humidity the system is expected to handle, which for our area means designing around a high dew point, not merely a high temperature.
Notice what is not on that list. Square footage appears only as one input among many, and by itself it predicts almost nothing.
Why the rules of thumb keep circulating anyway
Somebody will tell you a number of square feet per ton. Politely ignore it.
Those figures came out of an era of leakier houses, single-pane windows, thinner insulation and different construction, and they were rough even then. Applied to a modern Florida house they oversize routinely, and applied to two houses of identical size with different orientation they give the same answer for both, which is proof enough that they are not measuring the thing that matters. The existing equipment is no better as evidence. If your current unit was sized by the same rule of thumb fifteen years ago, matching it just inherits somebody else's mistake, and matching it plus a half ton compounds it.
Sometimes capacity is not the constraint at all
Here is a pattern we see constantly. A homeowner is told the house needs a bigger unit because certain rooms never cool properly. The actual limitation is a return that is undersized for the airflow the system needs, or a duct system leaking a quarter of its air into the attic, or a supply run crushed behind a stored box.
Put a larger compressor upstream of that same restriction and nothing improves. Airflow is still the bottleneck, the extra capacity has nowhere to go, and now the system short-cycles against the restriction while costing more to run. Static pressure measurements tell you which situation you are in, and they take a technician a few minutes. Correcting airflow is often cheaper than buying capacity and produces the better result. See why one room is always hotter.
Do envelope work first, then calculate
Order of operations matters more than people expect. If you are also planning attic insulation, air sealing, or solar control film on the west windows, do that work first and run the load calculation on the improved house. A lower load can mean a smaller and less expensive system that dehumidifies better, since the same equipment matched to a smaller load runs longer cycles. Do it in the other order and you have paid for tonnage you then spend money making unnecessary.
Equipment type interacts with this too. Two-stage and variable-speed systems run extended low-capacity cycles that pull far more moisture out than a single-stage unit of the same nominal size, which is why they outperform their efficiency ratings in this climate. See what variable speed buys you.
What to ask the contractor quoting you
Ask whether they will perform a Manual J on your house, and then ask to see the output. A real one produces a document carrying your address, your window areas and orientations, your insulation values, and a calculated load in BTU per hour broken into sensible and latent. That last split is the part worth looking at in Florida. A quote that arrives without any of it, sized off the old unit or off floor area, is a guess wearing a suit.
Then ask two follow-ups. What is the calculated load, and what capacity are you proposing? Modest headroom above the calculated load is normal and expected. A jump of a full ton over it deserves a specific explanation, and "to be safe" is not one.
We run the calculation before quoting on every installation, and when the honest finding is that your existing system is sized correctly and the real problem is somewhere else, that is what we will tell you. We work across all four Tampa Bay counties.
