No, it is a good idea. But, Tesla won't sell it because it is very valuable and the value increases many fold if Tesla keeps it to itself.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
- Status
My understanding was that the Assembly line 2 was designed to transition to producing both S and X and this transition would free up the space of assembly line 1 to be used with Model 3. The much heavier use of robots on Assembly Line 2 is necessary to allow both S and X to be produced at suitable speed on the same line.
I administer an ESPP. Not that it matters for share count since they are outstanding as soon as the purchase is made, but selling prior to the 2 year holding period is rare and carries tax consequences (no LT cap gains and disqualified shares carry tax on the discounted amount). By its nature, a qualified ESPP just can't contribute much to issued shares. Employees can only buy up to $25k worth of company stock per year under the plan, per the IRC. Thus, at most, each employee can only buy $12,500 worth of TSLA (less discount) per 6 month period. Then you consider that most TSLA employees are not highly compensated and likely setting aside $100-200 per check (if anything) the number starts to not look very large.
But good news as Cal is recognizing we need or have many more all electric alternatives, as I'm sure you agree. They're making room for extending the courtesy to thousands of M3s.
Yep. So where do we disagree? Maybe MBLY didn't handle it very well but the bottom line is they're effectively forced to sell one of the key parts of their overall solution to a competitor. They probably are cursing themselves for not putting some of these clauses into the initial contract.
One can foresee the possibility of licensing the software, of course. And later a government call for industry standard as it prepares for mandatory equipment and software on all cars. Any estimate as to when either will happen?
We cannot count on our speculations to come true but we can consider their likelihood. (Pompous I know, "sometimes sounding like a sumptuous pounding of water in a mortar," as Trotsky described Kerensky's speeches.)
Last edited:
racer26
Active Member
I wonder how soon we'll see MBLY equipment no longer in new TSLA products, if we arent already. Any forum members got their shiny new P100D willing to look and see? I wouldn't be surprised if the EyeQ3 is gone, replaced by something else. TSLA has to know that someday soon the flow of EyeQ3's will dry up, and so I'm sure there's a replacement in the works, if not already implemented, and TSLA sitting on enough supply of EyeQ3 on hand already to carry them through implementation of the replacement.
racer26
Active Member
Agree. I don't think TSLA would ever sell the raw data set. Might sell licenses to the software, to a maker of a compelling EV with a compatible hardware set, but only during the period where TSLA can't keep up with the demand. Once they can, why give away your differentiating feature? For that reason, I won't be surprised if they never license it.
I would be surprised if they don't sell it to a consortium of developers for mandate by the government. It's not quite like seat belts which anyone can make given simple requirements with followup inspection and penalty.
My rationale is that Tesla is selling a lot of cars, and so it gains tonnes of data every day, in a compound way.
From here: Tesla Fleet Racking Up The Miles With & Without Autopilot
it seems that overall miles (AP/nonAP) are going exponential.
I've not yet to see a chart for AP miles alone, but:
So, Tesla (which is always in need o cash), could reach a tipping point
in which selling X mlns miles data is *almost* harmless,
because it will gain other ten of millions in few days.
I understand that the competitive advantage is crucial, so that of course the key:
selling data in a "harmless" way.
a.
From here: Tesla Fleet Racking Up The Miles With & Without Autopilot
it seems that overall miles (AP/nonAP) are going exponential.
I've not yet to see a chart for AP miles alone, but:
- EM recently said "1.5 mln AP miles per day" See f.e.: Tesla Motors, Inc. Autopilot Miles Are Soaring, Hitting "Roughly" 200 Million
- the rate of sold auto is increasing
- the "percentage of miles each driver using Autopilot is increasing as people learn to use the system."
So, Tesla (which is always in need o cash), could reach a tipping point
in which selling X mlns miles data is *almost* harmless,
because it will gain other ten of millions in few days.
I understand that the competitive advantage is crucial, so that of course the key:
selling data in a "harmless" way.
a.
Last edited:
Not if robots are accounted as tooling. In this case, it depreciates on a per car basis.
It makes sense that Mobileye would not want to be tied down to only Tesla for high definition crowd sourced mapping and fleet learning of driving.
At CES last year, Mobileye introduced their "Road Experience Management (REM)" effort:
Mobileye N.V. - Mobileye To Offer User-Generated Mapping For Autonomous Driving
They signed up GM, Volkswagen, and Nissan to join into this effort. I don't think anything actually shipped yet. Presumably, BMW joins this effort too.
At CES last year, Mobileye introduced their "Road Experience Management (REM)" effort:
Mobileye N.V. - Mobileye To Offer User-Generated Mapping For Autonomous Driving
They signed up GM, Volkswagen, and Nissan to join into this effort. I don't think anything actually shipped yet. Presumably, BMW joins this effort too.
So the semi-annual period purchases were made some time during August, under Tesla's plan?
As you previously noted with the quarterly vs semi-annual pricing, different corporations have different terms for their plans including different holding periods Once the holding period pipeline is full, quarterly/semi-annual sales offer both a low-risk, assured return and favorable tax treatment. At the end of last year, Tesla's headcount exceeded 13,000 and continues to hire aggressively. Anyone who thinks zeroing out stock based compensation (of which the ESPP is only part) in non-GAAP as a no expense accounting technicality needs to look more closely.
"Stock-Based Compensation
We use the fair value method of accounting for our stock options and restricted stock units (RSUs) granted to employees and our Employee Stock Purchase Plan (ESPP) to measure the cost of employee services received in exchange for the stock-based awards. The fair value of stock options and ESPP are estimated on the grant date and offering date using the Black-Scholes option-pricing model. The fair value of RSUs is measured on the grant date based on the closing fair market value of our common stock. The resulting cost is recognized over the period during which an employee is required to provide service in exchange for the awards, usually the vesting period which is generally four years for stock options and RSUs and six months for the ESPP. Stock-based compensation expense is recognized on a straight-line basis, net of estimated forfeitures.
The Black-Scholes option-pricing model requires inputs such as the risk-free interest rate, expected term and expected volatility. Further, the forfeiture rate also affects the amount of aggregate compensation. These inputs are subjective and generally require significant judgment.
We estimate our forfeiture rate based on an analysis of our actual forfeiture experience and will continue to evaluate the appropriateness of the forfeiture rate based on actual forfeiture experience, analysis of employee turnover behavior and other factors. Quarterly changes in the estimated forfeiture rate can have a significant effect on reported stock-based compensation expense, as the cumulative effect of adjusting the rate for all expense amortization is recognized in the period the forfeiture estimate is changed. If a revised forfeiture rate is higher than the previously estimated forfeiture rate, an adjustment is made that will result in a decrease to the stock-based compensation expense recognized in the consolidated financial statements. If a revised forfeiture rate is lower than the previously estimated forfeiture rate, an adjustment is made that will result in an increase to the stock-based compensation expense recognized in the consolidated financial statements.
As we accumulate additional employee stock-based awards data over time and as we incorporate market data related to our common stock, we may calculate significantly different volatilities, expected lives and forfeiture rates, which could materially impact the valuation of our stock-based awards and the stock-based compensation expense that we will recognize in future periods.
My understanding is that each dataset is required to be held independently, not commingled. Slows down a process in which the are already far behind. Tesla was an outlier in that they wouldn't give access to any data.
davecolene0606
Member
Imagine a genetic testing company (23&me) claiming they own all the data from their customers.... Oh yea they do and that is their "business model". More and more companies are claiming thing ownership of the data and we sign away that ownership all the time.
Fire Away!
3Victoria
Active Member
Might AP2.0 hardware starting after the October break at Nummi? That would fit well with the quarter end push and discounts. The later might assuage to pain of missing the new hardware.
One of Tesla's early "innovations", was that the robots would change the tools, so you could have one robot performing 3-5 different tasks, like welding or installing glass. But changing tools takes time, so I think they are now prepared to pay more to save the time needed to change tool heads.
- Status
Similar threads
- Locked
