TSLA · Tesla, Inc.

Auto - Manufacturers · mkt cap $1581.3B · calls: Q1 FY2026 vs Q4 FY2025

55.0 conviction

enthusiasm:36.0 · trend:0 · quantifies:12 · impact:4 · under_radar:0 · credibility:0 · confirmation:3

Enthusiasm latest 9 / prev 9 (flat)

AI is not a sidebar but the entire thesis at Tesla — FSD/autonomy, Robotaxi, Optimus, and in-house AI silicon (AI5/AI6/Dojo, Terafab) dominate both calls, with enthusiasm pinned near the maximum in each (flat, ~9/10). Credibility is reinforced by hard, AI-tied numbers — ~1.3M FSD paid customers (up from ~1.1M), ~10B fleet FSD miles, an early AI5 tape-out, Robotaxi expansion to Dallas/Houston with zero reported injuries, and FSD approvals in the Netherlands/China — and management has explicitly reframed the business so 'FSD [is] a product and vehicle as only the delivery mechanism.' The main caveat is that the largest payoffs (Optimus at 1M units/yr, material unsupervised-Robotaxi revenue, Terafab) remain forward guidance for 2027+, funded by a >$25B AI-heavy CapEx ramp and negative free cash flow.

GROUNDED IMPACT vs CONSENSUS

Grounded on actual base — revenue $94.8B · net income $3.8B · net margin 4.0% · diluted EPS 1.08

Aggregate est. rev uplift: 2.3% · EPS uplift: 2.0% · vs analysts: inline · priced in: high · confidence: 5/10

Claim	Figure	Arithmetic	Rev %	EPS %
FSD net adds ~180k/qtr (mgmt-confirmed) revenue	~180,000 net new paying users/quarter; 1.1M->1.3M cumulative	180k x 4 = ~720k adds in FY2026. Blended realized price ~ (30% upfront @$8k US + 70% subs @$99/mo, global pricing lower, mid-year timing) -> ~$2.0B incremental RECOGNIZED FSD revenue (range $1.5-2.5B). Rev: $2.0B/$94.827B = 2.1%. Software incr net margin 60% (gross ~90%, less compute/R&D, after 21% tax): $2.0B x 0.60 = $1.2B NI; EPS: $1.2B/$3.794B = 31.6%. (Big % = thin 4% margin base.)	2.1	31.6
FSD paid customers ~1.3M cumulative (Q1'26) engagement · soft	nearly 1.3M globally	Stock figure behind the flow above; revenue impact captured in the net-adds line to avoid double-counting.
FSD paid customers ~1.1M cumulative, ~70% upfront (Q4'25) engagement · soft	~1,100,000; ~70% upfront	Prior-quarter stock; informs upfront/subscription mix used above. Not separately sized.
FSD cumulative miles ~10B other · soft	~10 billion miles	Training/validation data metric; no direct revenue mapping. Supports autonomy roadmap, not FY2026 P&L.
Optimus capacity 1M units/yr (Fremont) revenue · soft	1,000,000 units/yr, long-term	Long-term capacity, NOT next-FY. FY2026 = 'starter production' late Jul/Aug, 'significant numbers next year' (FY2027). FY2026 Optimus revenue immaterial / largely internal -> ~$0 next-FY. Large optionality but unanchored for FY2026.
Robotaxi fleet >500 vehicles, doubling/month revenue · soft	>500 vehicles; mgmt: 'not super material this year'	Even at exponential growth to a few thousand vehicles by end-FY2026 @ ~$30-50k rev/vehicle/yr -> ~$0.2B. Rev: $0.2B/$94.827B = 0.2%. Capital-heavy early ramp, ~25% incr net margin -> $0.05B NI; EPS 1.3%. Mgmt says material in FY2027, not FY2026.	0.2	1.3
Robotaxi doubling every month (exponential) other · soft	~2x/month	Growth-rate claim, no revenue anchor; context for the >500-vehicle line.
Autonomy TAM: <5% (maybe 1%) miles human-driven other · soft	<5% / 1% long-term	Long-horizon TAM vision; no dated dollars. Not sizeable for next-FY.
CapEx >$25B FY2026 (AI/compute/Optimus/AI5-heavy) cost	>$25B (raised from >$20B)	Hard guide. Incremental ~$13-14B above ~$11B baseline is the AI build. Depreciation lags (assets in service progressively; fast compute depr ~3-4yr): incremental FY2026 D&A ~$1.5B pre-tax. After-tax drag = $1.5B x (1-0.21) = $1.185B; EPS: -$1.185B/$3.794B = -31.2%. Topline ~0 (capacity, not FY2026 revenue).	0	-31.2
CapEx >$20B (prior Q4'25 guide) cost · soft	>$20B	Superseded by the >$25B guide above; not separately counted to avoid double-counting the same envelope.
Terafab research chip fab ~$3B, few thousand wafers/mo cost · soft	~$3B, construction starting FY2026	Construction-phase capex (within/near the $25B envelope); no FY2026 revenue, minimal in-service depreciation while under construction. Strategic chip-supply hedge, immaterial to FY2026 P&L.
AI memory efficiency ~10x+ vs Grok other · soft	order-of-magnitude better	Capability/cost-efficiency claim, no dollar anchor; lowers future compute cost but unquantifiable for FY2026.

Assumptions: Next-FY = FY2026 (base is FY2025 actual). FSD: ~720k net adds (180k/qtr annualized, no acceleration credit), blended ~$2.0B incremental recognized revenue (price-assumption-dependent, range $1.5-2.5B), 60% incremental net margin (software/services, post-21% tax). Robotaxi: ~$0.2B revenue, 25% incremental net margin (capital-heavy early ramp), explicitly soft per mgmt 'not material this year'. CapEx: $25B hard guide, ~$13-14B AI-incremental, ~$1.5B first-year incremental depreciation at 21% tax. Optimus/Terafab/autonomy-TAM/miles/memory-efficiency = long-term, no FY2026 anchor -> null. Tax 21%. Only the latest CapEx guide and the net-adds flow are counted to avoid double-counting cumulative-count and superseded-guide claims.

Top line: Quantifiable AI topline for FY2026 is modest: FSD ~$2.0B (+2.1%) on ~720k net adds, plus a soft ~$0.2B (+0.2%) of early robotaxi = ~$2.2B, +2.3% of the $94.827B base. Optimus and material robotaxi revenue are FY2027+ events, not FY2026. So near-term AI revenue is real but small relative to the franchise.

Bottom line: The interesting result: AI is roughly EPS-NEUTRAL in FY2026. High-margin FSD adds ~$1.2B NI (+31.6% EPS on the thin $3.794B base) but the $25B AI-heavy CapEx introduces ~$1.5B incremental depreciation (~$1.185B after-tax, -31.2% EPS). The two nearly cancel -> net ~+$0.07B NI, ~+2% EPS. The 4% net margin means both swings look large in % terms; the offset is the real story.

Consensus FY2026 already implies revenue $102.75B = +$7.92B (+8.4%) and NI $4.62B = +$0.83B (+21.8%) over the FY2025 actual base. My AI-attributable topline (~$2.2B, +2.3%) is only ~28% of consensus revenue growth — it fits comfortably inside, so it is not incremental to the Street. On the bottom line my net AI EPS effect is ~flat (FSD +$1.2B offset by ~$1.2B capex depreciation), versus consensus NI growth of +$0.83B that must come from core auto/energy margin — i.e. consensus is NOT relying on AI to lift FY2026 EPS, and the quantifiable AI math doesn't point above it. The genuinely large items (Optimus 1M/yr, exponential robotaxi, <5%-human-driven TAM) are unpriced optionality but land in FY2027+ and cannot be sized for next-FY, so they don't move the near-term priced_in verdict.

QUANTIFICATIONS

FSD paid customers: nearly 1.3 million globally (growth led by subscriptions; upfront purchases +7%) (as of Q1 FY2026, topline)
“On the FSD adoption front, we continue to see improvement, reaching nearly 1.3 million paid customers globally.”

FSD paid customers: nearly 1,100,000 globally, ~70% upfront purchases (as of Q4 FY2025, topline)
“FSD adoption continued to improve in the quarter, reaching nearly 1,100,000 paid customers globally. Of these, nearly 70% were upfront purchases.”

New paying FSD users in quarter (analyst-stated, mgmt-confirmed): ~180,000 net adds vs ~100,000 NA cars sold (~implied 30-35% NA attach) (Q1 FY2026, topline)
“So you have 180,000 new users, paying users this quarter... [Ashok:] I think you're thinking about it the right way, Pierre.”

FSD fleet cumulative miles driven: close to 10 billion miles (within the next few weeks (Q1 FY2026), topline)
“the entire Tesla customer vehicle fleet, which is close to driving 10 billion miles on FSD in the next few weeks”

Optimus production capacity target: 1 million units per year (Fremont SX space) (long-term, topline)
“I'm confident that we'll get to a million units a year in Fremont of Optimus 3.”

Robotaxi fleet carrying paid customers: well over 500 vehicles (Bay Area + Austin) (as of Q4 FY2025, topline)
“In terms of taxi vehicles carrying paid customers, I think we're well over 500 at this point between the Bay Area and Austin.”

Robotaxi fleet growth rate: ~doubling every month (exponential) (near-term, Q4 FY2025, topline)
“This will probably double every month type of thing. It's on an exponential curve.”

Autonomous share of future miles (autonomy TAM): <5%, maybe as low as 1% human-driven (long-term/future, topline)
“probably less than 5% of miles driven will be where somebody's actually driving the car themselves in the future, maybe as low as 1%.”

CapEx (heavily AI/compute/Optimus/AI5-related): over $25 billion (FY2026, bottomline)
“our current expectation for 2025 -- 2026 is over $25 billion of CapEx... We're further increasing our investment in AI-related initiatives, including the AI infrastructure to support Robotaxi and the launch of Optimus.”

CapEx (heavily AI/compute/Optimus-related): in excess of $20 billion (FY2026 (as guided in Q4 FY2025), bottomline)
“we are expecting that CapEx would be in excess of $20 billion... we'll also be spending money for building our AI compute infrastructure.”

Research chip fab investment (AI chip supply): ~$3 billion, ~a few thousand wafers/month (construction starting FY2026, bottomline)
“This is something we expect to be probably a $3 billion-ish initiative and capable of maybe a few thousand wafers per month.”

Tesla AI memory efficiency ('intelligence density'): order of magnitude (10x+) better than large models like Grok (current (Q4 FY2025), both)
“Tesla's AI is, in terms of memory efficiency, I think more than an order of magnitude better.”

PAST (realized)

Q1 FY2026 — Tesla AI chip team taped out AI5 ('taping out AI5'), finished early after the team 'worked every weekend for 6 months straight.'
Q1 FY2026 — FSD v14.3, a 'major architectural update,' was released; Robotaxi in Austin/Dallas/Houston runs on '14.3 variants.'
Q1 FY2026 — FSD reached 'nearly 1.3 million paid customers globally,' growth led by subscriptions.
Q1 FY2026 — 'We've expanded Robotaxi to Dallas in Houston using the same software source in the Bay Area,' with 'not had a single one [injury] to date' / '0 incidents.'
Q1 FY2026 — Received FSD approvals in the Netherlands and in China.
Q4 FY2025 — 'We were able to do our first rides with no safety monitor in the car in Austin... paid rides' with no chase car.
Q4 FY2025 — FSD reached 'nearly 1,100,000 paid customers globally... nearly 70% were upfront purchases.'
Q4 FY2025 — 'We already use the AI4 chips in our data centers,' combined with NVIDIA hardware for training.
Q4 FY2025 — Robotaxi vehicles carrying paid customers 'well over 500... between the Bay Area and Austin'; Optimus did basic factory tasks in R&D.
Q4 FY2025 — Made the xAI investment; 'today, if you look at Tesla vehicles, we are using Grok in there.'

CURRENT (now)

Q1 FY2026 — 'we now emphasize FSD as a product and vehicle as only the delivery mechanism.'
Q1 FY2026 — Fleet 'close to driving 10 billion miles on FSD in the next few weeks'; scaling the QA fleet and growing the Robotaxi fleet quarter-over-quarter.
Q1 FY2026 — 'a lot of momentum for designing AI6' and beginning to 'discuss ideas for Dojo 3'; finalized plans for and placing orders for the Austin research chip fab.
Q1 FY2026 — Spend on AI-related initiatives (AI5 chip development, Optimus, Cybercab, Semi, Megablock) 'at elevated levels' expected to continue through 2026.
Q4 FY2025 — Scaling the supervised Robotaxi service to learn scaling issues while engineering 'laser-focused' on solving unsupervised FSD.
Q4 FY2025 — Transitioning FSD 'fully to a subscription-based model'; AI-related spend and AI compute infrastructure investment at elevated levels.
Q4 FY2025 — Spending most Saturdays/Tuesdays personally on AI5 chip design as 'the number one most critical thing to get done.'

FORWARD (guidance)

Q1 FY2026 — Whole pipeline of FSD improvements leading to unsupervised FSD 'anywhere in the world that it is legal'; v15 'a complete overhaul of the software architecture' on AI4 by end of this/early next year.
Q1 FY2026 — Unsupervised FSD to customer cars 'probably in the fourth quarter'; unsupervised FSD/Robotaxi in 'a dozen states or more later this year.'
Q1 FY2026 — Robotaxi revenue 'not be super material this year' but 'material probably in a significant way next year.'
Q1 FY2026 — AI5 will go into Optimus and the data center; AI4.1/AI4+ (32GB/SoC, ~10% more compute) into production ~mid next year.
Q1 FY2026 — Optimus starter production ~late July/August this year, ramping to 'significant numbers next year'; second Optimus factory at Giga Texas ~summer next year; 'Optimus will be... the biggest product ever.'
Q1 FY2026 — Terafab: research fab (~$3B) construction this year; SpaceX to do the initial scaled-up Terafab using Intel's 14A process.
Q4 FY2025 — 'fully autonomous vehicles in probably somewhere between a quarter and half of the United States by the end of the year'; 'dozens of major cities by the end of the year'; Robotaxi fleet to 'double every month.'
Q4 FY2025 — Goal of 'a million units a year in Fremont of Optimus 3'; Optimus to 'move the needle on US GDP significantly.'
Q4 FY2025 — AI5 to be used in data centers; AI6 'aspirationally... in under a year'; chip supply a 'limiting factor' in 3-4 years, prompting a Tesla Terafab (logic + memory + packaging) domestically.
Q4 FY2025 — Long term 'less than 5% of miles driven... maybe as low as 1%' will be human-driven; CyberCab to be 'several times more' units than all other vehicles combined.

PRICED-IN (REFINED)

UNKNOWN

Est. revisions unknown · Fwd P/E 254.5 · EV/Sales 16.1x

LLM synthesis failed: ask_json failed after 2 tries: claude exit 1:

COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS

RECENT AI CATALYSTS & NEWS

2026-06-02 · Tesla Stock and AI Race Heat Up as SpaceX IPO Targets $1.75 Trillion
2026-06-02 · Here's Why Tesla Stock Soared in May (And What the SpaceX IPO Means to Tesla Investors)
2026-06-01 · SpaceX's IPO is about far more than just rockets, satellites and AI
2026-06-01 · Tesla Stock Tests Short-Term Moving Averages Amid Surging Scandinavian Demand
2026-06-01 · Why Tesla stock is sinking over 4% today
2026-06-01 · 3 More Things I Learned from the SpaceX IPO Prospectus

OPTIONS / MARKET STRUCTURE

option liquidity: good

ATM IV

—

TYPICAL BID-ASK

—

OPEN INTEREST

—

proxy inputs — dollar-ADV $25.6B · beta 1.793 · px $421.83

source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.

CONFIRMATION — INSIDERS · 13F · LANGUAGE

Mixed — insiders selling, institutions adding, management language unknown.

INSIDERS selling 64 open-market sell(s) vs 0 buy(s) — net distribution

INSTITUTIONS (13F) adding as of 2026-03-31: 191 new / 381 closed positions; 2215 increased / 1455 reduced; institutional ownership -0.88pp; -196 net 13F holders

MGMT LANGUAGE n/a unknown language analysis failed: ask_json failed after 2 tries: claude exit 1:

VERBATIM AI QUOTES

“As you've heard me say a few times, I think Optimus will be our biggest product -- not just Tesla's biggest product ever, but probably the biggest product ever. And I remain convinced of that conclusion.”
— Elon Musk, Q1 FY2026

“For full self-driving and Robotaxi, version 14.3 was a major architectural update. And we have a whole pipeline of major improvements to full self-driving that, we believe, will lead to unsupervised full self-driving being available anywhere in the world that it is legal to do so.”
— Elon Musk, Q1 FY2026

“The -- congratulations to -- again to the Tesla AI chip team for taping out AI5. That's going to be a great chip. I think probably the best AI inference chip for edge compute that exists. And certainly, I think the best value for money.”
— Elon Musk, Q1 FY2026

“And we already have a lot of momentum for designing AI6, and we've begun to discuss ideas for Dojo 3.”
— Elon Musk, Q1 FY2026

“we have evolved our vehicle sales strategy, where we now emphasize FSD as a product and vehicle as only the delivery mechanism.”
— Vaibhav Taneja, Q1 FY2026

“On the FSD adoption front, we continue to see improvement, reaching nearly 1.3 million paid customers globally.”
— Vaibhav Taneja, Q1 FY2026

“we are simultaneously solving the long tail of safety by monitoring the metrics across the entire Tesla customer vehicle fleet, which is close to driving 10 billion miles on FSD in the next few weeks and also scaling up the amount of QA fleet that we have across the entire U.S. to accelerate our safety validation”
— Ashok Elluswamy, Q1 FY2026

“So I think you need kind of an orchestration AI, which Grok would be good for orchestration. And then for Optimus' voice, having a low-latency intelligent voice AI, Grok is actually very good for that.”
— Elon Musk, Q1 FY2026

“I think with the advent or continued growth of AI and robotics, I think we actually are headed to a future of universal income, not universal basic income, but universal high income.”
— Elon Musk, Q4 FY2025

“We were able to do our first rides with no safety monitor in the car in Austin. These are paid rides, randomly selected paid rides with no safety monitor.”
— Elon Musk, Q4 FY2025

“We expect to have fully autonomous vehicles in probably somewhere between a quarter and half of the United States by the end of the year, pending regulatory approval.”
— Elon Musk, Q4 FY2025

“I'm confident that we'll get to a million units a year in Fremont of Optimus 3. This Optimus really will be a general-purpose robot that can learn by observing human behavior.”
— Elon Musk, Q4 FY2025

“I think long-term Optimus will have a very significant impact on the US GDP. It will actually move the needle on US GDP significantly.”
— Elon Musk, Q4 FY2025

“Completing the AI5 chip design and having it be a great chip is arguably the number one most critical thing to get done, which is why I'm spending more time on that than currently anything else at Tesla.”
— Elon Musk, Q4 FY2025

“In autonomy, software will be the driver for growth from now. As we aim to maximize the global fleet, we have been laser-focused on COGS from our side.”
— Vaibhav Taneja, Q4 FY2025

“I actually think Tesla is ahead of the rest of the world in intelligence density of AI by an order of magnitude or more.”
— Elon Musk, Q4 FY2025

ANALYST QUESTIONS ON AI

Q (Q1 FY2026, Will Stein (Truist)): On the Terafab project, which party takes responsibility for funding, designing, building, operating and taking production — and what is Intel's involvement?
A: Elon: Tesla builds a ~$3B research fab on the Giga Texas campus (a few thousand wafers/month, to try out ideas and validate production processes); SpaceX handles the initial scaled-up Terafab; intercompany deals must clear both boards. On Intel: 'Intel is excited to partner with us on some of the core manufacturing technologies. So we plan to use Intel's 14A process, which is state-of-the-art.'

Q (Q1 FY2026, Pierre Ferragu (New Street)): You added ~180,000 new paying FSD users this quarter vs ~100,000 cars sold in North America — so most Hardware 4 owners in North America are already using FSD. Is that the right way to think about it?
A: Ashok: 'I think you're thinking about it the right way, Pierre.' Adds that churn of subscribers is coming down and customers are driving longer, reflecting that the product is getting better.

Q (Q1 FY2026, Pierre Ferragu (New Street)): On Optimus + xAI/Grok — will the intelligence sit onboard the robot, or will a fleet of ~1M Optimus drive significant data-center inference demand for 'system 2' thinking?
A: Elon: a lot of intelligence runs locally so the robot keeps working if disconnected; Grok serves as an 'orchestration AI' and low-latency voice; 'Optimus could probably work for several hours without any management oversight.'

Q (Q1 FY2026, Dan Levy (Barclays)): How much of Terafab is motivated by getting better economics/leverage on your mid-term chip purchases, and how long to ramp to yield parity?
A: Elon: 'Terafab is not some sort of mechanism to generate leverage over our chip suppliers... we don't see a path to having enough efficient quantity of AI chips down the road.' Believes Tesla has research ideas for radically better AI chips.

Q (Q1 FY2026, Mark Delaney (Goldman Sachs)): Given FSD's improvements (v14), has your approach to new vehicle models evolved (family/compact vehicle)?
A: Elon: Cybercab is the compact 2-seater; 'most of our production long term will be Cybercab because 90% of miles driven are with 1 or 2 people'; long term the whole lineup goes autonomous, with the only manually driven car being the new Roadster.

Q (Q1 FY2026, Colin Langan (Wells Fargo)): Now that you've removed the safety driver in Austin and are expanding to Dallas/Houston, what safety metrics give you confidence to expand (miles/intervention, /accident, /fatality)?
A: Ashok: 'We track basically all the metrics that you mentioned,' using a large QA fleet across the U.S. and neural-network simulators to assess interventions before expanding. Elon adds that most limits are 'convenience issues' (the car getting stuck/scared) rather than direct safety issues.

Q (Q1 FY2026, say.com (investor)): What milestones are you targeting for unsupervised FSD and Robotaxi expansion beyond Austin this year, and how will it drive recurring revenue?
A: Elon: hopes to have unsupervised FSD/Robotaxi in roughly a dozen states by year-end with a cautious rollout; 'unsupervised FSD or Robotaxi revenue would not be super material this year. But... it will be material probably in a significant way next year.'

Q (Q4 FY2025, Emmanuel Rosner (Wolfe Research)): On the >$20B CapEx increase — where are the investments going, is it one-time for 2026, and how will you finance the cash burn?
A: Vaibhav: six factories starting production, plus 'a lot more compute' for scaling Optimus and training; it's a multi-year investment phase (chip fab and solar are 'infrastructure plays'); funded from $44B cash plus bank debt against consistent robotaxi cash flow.

Q (Q4 FY2025, Andrew (Morgan Stanley)): On the xAI investment announced today — what does the collaboration look like and how is xAI's work leveraged at Tesla?
A: Vaibhav: it's 'furtherance of our master plan four'; Tesla already uses Grok in vehicles and xAI can accelerate progress. Elon: investors asked for it; Grok will manage/optimize a large autonomous fleet and act as the 'orchestra conductor' for Optimus robots.

Q (Q4 FY2025, Dan Levy (Barclays)): Given very tight memory supply, are there near-term constraints on procuring memory and how do you bridge the next few years?
A: Elon: Tesla's AI is highly memory-efficient ('intelligence per gigabyte... more than an order of magnitude better'); has a logic/memory solution for ~3 years, but beyond that will be supplier-limited, which is why a Terafab is needed — 'Optimus is completely useless without an AI chip.'

Q (Q4 FY2025, George Gianarikas (Canaccord)): With a surge of (mostly Chinese) humanoid startups, what are Tesla's long-term competitive advantages and how will Optimus differ?
A: Elon: China is by far the toughest competition (great at manufacturing and AI), but Optimus will be more capable in real-world AI, electromechanical dexterity and hand design; the three hardest problems are the hand, real-world AI, and scaling production — 'Tesla is the only company that actually has all three.'

Q (Q4 FY2025, Colin Langan (Oppenheimer)): On the ambitious tech program — how do you think about R&D spend and the synergies of purpose-built hardware (batteries, memory, system efficiency) across multiple end markets?
A: Elon: the goal is to scale autonomous vehicles and humanoid robots to high volume while addressing geopolitical risk ('only the paranoid survive'); ensure Tesla can keep building batteries, robots and AI chips no matter what. Vaibhav: 'all this comes out of necessity... we have no choice.'

Q (Q4 FY2025, say.com (investor)): What is the current bottleneck to increased Robotaxi deployment and unsupervised FSD — model safety/performance, monitors, or another blocker?
A: Ashok: scaled the supervised service to learn 'scaling problems' (charging, getting stuck) while engineering solved unsupervised FSD; started unsupervised public service in Austin; the v14 customer release saw 'a huge jump in performance.'