San Francisco Fell Nine Points. Its Watch Status Never Moved.
On Friday the platform published a nine point drop for San Francisco, from 62 to 53. It was the largest move in the index this month, and it dropped San Francisco from third to sixth of 25.
Nothing on the ground changed. Five of the six pillars that make up the score read exactly what they read in June. The sixth, momentum, fell from its ceiling of 14 to 5.4.
Momentum counts how much scored signal a market generated in the last ninety days. In June that window still held the coverage of Joby's March flight across San Francisco Bay and around the Golden Gate. It does not anymore.
The watch layer, which reads categorical events like a bill signed or a zoning vote, has said DEVELOPING and STABLE since April 10. It still does. Both readings are right. The 62 was the flyover. The 53 is San Francisco.
What the model is seeing
The AirIndex score is a 0 to 100 reading of how ready a market is for commercial air taxi service. Six pillars build it: regulatory authority, trip demand, landing capacity, operator commitment, weather operability, and momentum. Five of them are stock. They measure what a market holds, and they move when a real event lands, a law enacted, an operator committed, a facility approved. Momentum is the one pillar that is a clock. It counts the market's classified signals over a rolling ninety day window, excluding operator expansion coverage, and scores that count against the top decile of the index. When the count falls, the pillar falls, and no fact on the ground has to change.
San Francisco is the cleanest illustration of that clock the index has produced. In the June reading its momentum pillar sat at the cap, 14 of 14, and its score was 62, third in the country. Every one of the other five pillars reads today exactly what it read then: regulatory 15.3 of 18, demand 8.5 of 18, landing 7.4 of 18, operator 7.2 of 18, operability 9.3 of 14. Only momentum moved, from 14.0 to 5.4, and the score followed it down to 53. Nine points, one pillar, zero events.
What was in the window in June was one flight. On March 13 and 14 Joby flew a piloted aircraft across San Francisco Bay and around the Golden Gate to open its Electric Skies Tour, and the platform classified 60 San Francisco items in March and 45 in April, the overwhelming majority of them coverage, and repeated coverage, of that flight. Thirty one of the items in the June window carried a certification milestone tag alone. Then the feed went quiet: nine items in May, ten in June, eight in July. By Friday the March and April coverage had all crossed out of the ninety day window, and the pool that momentum counts had gone from 45 signals to 23. Against a top decile benchmark of 62, that is 5.4 of 14.
Two markets hit the ceiling in June. Only one is still there.
Tampa was the other market pegged at 14 of 14 in June, and this brief covered it in Issue 10 as the loudest market no operator had committed to. Its momentum pillar still reads 14 today, because Tampa's feed kept producing, 64 non operator signals in the current window against San Francisco's 23. On Friday, in the same batch of published changes that took San Francisco down nine, Tampa moved up three, from 48 to 51, and crossed the MODERATE line. Same pillar, same ceiling in June, opposite outcomes in August. The difference is not what either city built. It is which one kept generating signal after the cameras left.
That is worth saying plainly because San Francisco's velocity, the raw count of everything the classifier tags to the market, still looks strong. Twenty four signals in 30 days, sixth in the country, accelerating. But twelve of those twenty four are operator expansion coverage, and the momentum pillar excludes that category on purpose, because operator interest is priced elsewhere, in the operator pillar's interest floor, and that floor is already maxed here. Most of the rest of the August volume is regional or national rather than local: Archer's Midnight flights between Salinas and Monterey on August 4, which the classifier associates with the Bay Area market, and a Federal Register cluster on August 12 covering powered lift pilot certification, beyond visual line of sight operations, and airworthiness criteria for Archer's and BETA's aircraft, which touches every market at once. Volume is back. Local signal is not.
The 53 is what the market actually holds
Strip the flyover out and San Francisco is a solid MODERATE market with a specific shape. Regulatory authority is the carrying pillar at 15.3 of 18: California's enacted AAM statute and vertiport zoning credit, with posture scored neutral. Weather operability reads 9.3 of 14, the marine layer is a real structural drag and it will not legislate away. Landing capacity is 7.4 of 18, roughly four in ten of the market's assessable heliports converting on the platform's feasibility read.
The operator pillar at 7.2 of 18 is the one that will surprise Bay Area readers, because both of the country's leading eVTOL manufacturers are headquartered in the region, Joby in Santa Cruz and Archer in San Jose. The register carries one scored presence in San Francisco: Joby at testing stage, set March 31 on the strength of the demonstration flights. Archer's June 2024 plan for a five location Bay Area network, South San Francisco at Kilroy Oyster Point, Napa, San Jose, Oakland and Livermore, with a first site targeted for as early as the end of 2025, has produced no scored operational event and does not appear in the register as a San Francisco presence. If it were entered at announced stage, the pillar would not move: the operator interest floor, driven by 91 cumulative expansion signals against a cap of 50, already outweighs the curated depth. San Francisco is where the aircraft are built and demonstrated. It is not yet where anyone has committed to fly them for money.
Signal Event
On Friday afternoon, August 14, the platform published eight score changes in one batch after the audit gate review. San Francisco's was the largest: 62 to 53, a nine point drop, all of it in the momentum pillar. In the same batch Orlando rose seven to 64, Tampa crossed EARLY to MODERATE at 51, and Phoenix and New York both dropped from MODERATE to EARLY at 49. San Francisco's ranking moved from third to sixth of 25, tied with Houston at 53.
Why it matters
The publish rule for market scores requires a move of two points or more, a tier crossing, or a triggering event before a change is published, which is why the platform's daily reading drifted from 62 down through the fifties for ten weeks before Friday's number went out. Nine points is not a verdict on San Francisco. It is the platform stating, in public and on the record, that its June reading was inflated by one event's coverage cycle and that the resting level is 53. Anyone who read the 62 as a ranking of the market's readiness was reading a news wave. The score is now correcting for it, and the correction is the point.
Why the score fell and the watch did not
The score and the watch layer measure different things, and San Francisco is where they visibly disagree. The score is accumulated position plus a momentum modulator. The watch layer is a ledger of categorical events in the last 30 days: a bill enacted or failed, a zoning approval or rejection, a scored change on a stock pillar. DEVELOPING means the market is generating elevated activity with no categorical event to give it a direction. That has been true of San Francisco since April 10 and it is still true today. No bill was signed, no bill failed, nothing was approved or denied. The watch layer had nothing to react to, so it did not.
That is by design, and this issue is the reason the design matters. If the watch layer flagged momentum decay as deterioration, every market that hosted a demonstration flight would be marked DETERIORATING ninety days later. Instead the momentum pillar carries the news cycle and the watch layer carries the facts. When both point the same way, a market is genuinely moving. When they split, as they do here, the split is the information: the market did not get worse, the coverage got quieter, and the score has stopped paying for coverage.
What to Watch
Nothing in the pipeline projects points for San Francisco. Four things decide whether the 53 holds, drifts, or gets a direction.
1. AB 431 on the Senate floor before August 31
California's Advanced Air Mobility Infrastructure Act cleared its final Senate committee on August 13, seven to zero, and now waits for a floor vote before the session's last day to pass bills, August 31, then the governor's desk. It would direct Caltrans to plan vertiports into the statewide aviation plan and preempt local AAM ordinances. San Francisco already holds enacted legislation credit, so the score would not move. The watch layer would: an enactment is a categorical positive event, and it would be the first directional signal San Francisco has produced since April.
2. An operator commitment that names a Bay Area site
Nineteen operator expansion events are tracked, eight at high confidence, all validation of a presence the register already scores. What would move the operator pillar is a stage change: Archer breaking ground or naming an opening date at Oyster Point, or Joby naming a Bay Area launch route rather than a demonstration. Either would lift curated depth above the interest floor for the first time and start paying the pillar on facts instead of coverage.
3. The FAA corridor filings resolving in October
Five FAA corridor filing events are tracked, four at high confidence, with determination windows opening October 11 and running to November 10. A determination adds no points on its own, but a Bay Area facility determination would be the first locally originated primary document on any local factor in San Francisco's ledger, which currently reads zero across operator, vertiport, pilot, zoning, and weather.
4. Whether momentum keeps draining
The June signals leave the ninety day window between September 1 and 28. Absent new signal the pillar gives back another point or two by October, and San Francisco at 51 or 52 sits within reach of the MODERATE line at 50. Absent is the operative word: one local event of any weight resets the clock. The market has three points of cushion and about six weeks to earn more.
Final Take
Read San Francisco's 53 as its resting level, not as a fall. The 62 was one flight, covered a hundred times, sitting inside a ninety day window. The window closed and the score said so, in public, nine points at once.
The market that remains is the one where the aircraft are built and nobody has yet committed to fly them for money. Watch the operator pillar and the Senate calendar. One of them will give San Francisco a direction. Coverage will not.
