Amber Electric, the upstart electricity retailer that allows consumers to play the wholesale electricity market, recently sent out an email to its Tesla Powerwall customers, and it’s a mix of good and bad news.
SmartShift, Amber’s battery control software that automates home batteries to charge when the wholesale price is cheap and discharge when high, has a problem. With both Amber and the Powerwall’s popularity surging, Amber has been grappling with scaling issues that are beginning to affect users’ ability to control their Tesla batteries through the Amber app.
“Due to rapid uptake in Amber for Batteries, we are experiencing scaling issues that are affecting your ability to take control of your Tesla battery via the Amber app. As a result, you may have noticed that there are times when you request certain manual control actions via the app only to find that nothing happens. This is also starting to impact the ability of SmartShift to automatically command Tesla batteries to charge and discharge at optimal times.”
Why is this happening? It’s all about the growth rate. The overall fleet of SmartShift-controlled Tesla batteries has ballooned, leading to a surge in requests and commands sent to Tesla’s API within a given time frame. While the growth is a testament to Amber and Tesla’s success, the fixed number of calls that Amber can make to the APIs — a precaution against cyberattacks and resource abuse — has not changed to match this growth, leading to a bottleneck.
This situation affects users’ SmartShift experience, with some manual actions via the app failing to execute. Despite this, Amber reassures users that the automated SmartShift system will still make accurate decisions for charging and discharging their Powerwall most of the time.
Amber says their team is on the case, prioritising resolving these control issues. They’re rolling out updates to improve service in the short term and have halted onboarding new Tesla batteries onto SmartShift to prevent exacerbating the issue.
Also, Amber is asking users to ease off on manual controls to give everyone a fair shot at optimisation.
The takeaway from Amber’s message is clear: they’re committed to fixing the Powerwall problem as swiftly as possible1 and will keep users updated on their progress.
I think it’s kinda cool to see Amber hustle like this. Sure, there will be hiccups along the way, but it’s all about making our grid greener, faster. Go Amber!
Footnotes
- Hopefully, before it all turns to SmartShit 🤣 ↩
About Finn Peacock
RSS - Posts
Our experience with Amber left a lot to be desired, after trying it out for 4 months it became clear that while we were getting warm fuzzy feelings from doing our bit for the grid by exporting during the evening peak, it was actually costing us financially to do it, as well as cycling our battery deeply every day.
We have 10kW of solar, a Tesla Powerwall and an Ioniq 5 EV.
Despite doing a lot of manual adjustments as the “Smartshift” wasn’t very smart at times, it was costing us about $30 per month. We are now with Powershop on a 12c FIT and the last bill was -$5.
What is really annoying is that now we’ve left Amber there’s a problem with the Powerwall due to a setting that was changed remotely to enable Smartshift. Now, if the Powerwall gets down to its reserve level and the house load exceeds 5 kW (eg. charging the car at night, 7kW) the grid supply won’t exceed 5kW and the balance comes from the PW reserve. Once the PW is totally flat, the grid supply will increase.
Amber’s response to this has basically been “Sorry, not our problem, go and talk to Tesla.”. Tesla’s response is “We’ll get back to you in a month.”
Hi Brian,
So sorry to hear about your pain, but thank you for sharing. The story is a big help to “fence-sitters” who are still wondering about the benefits of signing up to become a VPP.
It was also great to hear that you could get a 12c FIT.
Hoping you are able to quickly resolve the current issue.
Andrew
I would suggest rolling your own – check this out https://www.smartmotion.life/2023/09/12/amber-electric-curtailment-with-home-assistant/
I’ve got a sonnen battery and Amber does t support this brand at all. Sonnen won’t allow Amber access to their API as they have their own VPP service so Amber is considered competition. So I implemented EMHASS in my Home Assistant system and developed an API to control the battery and other deferrable loads. Don’t need smartshift. I know others have done the same with EMHASS and Tesla batteries so that’s a better way to do it if smartshift is oversubscribed.
I did the sums and wasn’t sure I could make Amber (or any wholesale option) work for us, as our exports are too high every day. I think Amber even says on their site that people who export a lot are probably not going to find the service as useful. And as one of your other articles said here, the wholesale price wasn’t all that different to the FIT from the better retailers, so the difference was going to be marginal. But with their curtailment through the inverter option, I’m still curious.
Their auto-curtailment through direct control of inverters (to manage exposure to negative wholesale prices) only works for some inverters, Tesla’s being one of them, though with more being added over time. Does that mean if you have, like we do, Enphase microinverters on your panels, but a Powerwall 2 as part of the system, Amber’s system would work? They just need to control the Powerwall for the process to function?
Great to hear Powerwalls are selling well, are there any numbers available?
We were with Amber for about 6 months until February 2023. However once we upgraded our solar system to 11kw and added a 10kwh Solar Edge battery in January (2023) the low and often negative FiT became a problem. Solar Edge inverter can be automatically set to limit export during negative FiT via SmartShift. The average FiT we achieved was around 9c. Also SmartShift sometimes went crazy, draining our battery just before a high tariff time. It was actually stressful.
Switched to Powershop with a 13c FiT (now 12c) in February 2023. So far have paid PS once ($10.85 in June) and currently $125.64 in credit. No longer need to worry if SmartShift was draining the battery. Less stress and I think financially better off. I don’t think Amber makes sense for large solar systems with battery. Better off with conventional power company with highest FiT as we barely purchase any power these days.
Hi Alex, that mirrors our experience, hopefully Amber left you with your system working as it should when you switched suppliers. We’re still waiting (5 weeks later) to get our Powerwall restored to normal operation.
Hi Julian,
Thankfully the Solar Edge API recovered straightaway after disconnecting and switching to Powershop. Zero downtime. From your experience SmartShift apparently becomes deeply embedded in the Powerwall software. Sorry to hear that….
I still follow the Amber rates via Charge HQ (we solar charge our Tesla M3 and MY). Over the past few months I’ve noticed often negative FiT on offer at Amber… Glad I switched to PS.
I’d be reluctant to join a VPP in the future, as I would be prefer to be in control of the battery.
Alex
Sorry David , got Surname and first name mixed up
The key questions are still:
1) With SmartShift are people seeing payback periods under 5 years?
2) Why is home storage so small? 10 KWh for approx $13k is very expensive and for most people government bonds offer a safer and greater return on investment.
When cars have many times this capacity we are being asked to pay over $1K/KWh for homes with higher consumption? This equates to over $1 billion / GWh when modelling suggests 100 GWh of storage is require to balance renewables across the grid.
Why is home storage so expensive and such poor value? Is this unavoidable?
Very interesting thread, as I have 6 kw of solar, thinking of upgrading to 10 kw, then adding either Sungrow or Vaulta battery about 10- 12 kwh, and moving from Red energy ( 7c FIT) to Amber, and hoping I can make it pay for itself. So I think I’ll wait a little longer…. Unless I get a great offer on a battery….
This is frustrating but I’m in it for the long haul. If Amber can sort these teething problems I’m optimistic of long term gains. What I am struggling with is, smartshift appears to be able to charge to 100% during the day, but only discharge to approximately 50% in an evening. This 30% represents approx 15+ kWh of energy (4 × powerwalls = 54kwh) therefore we are losing earnings each day while we are asked to trust the automation. Yesterday we achieved $8.35 in earnings using smartshift whilst the day prior we earned over $40 using manual control. It’s a big ask Amber, to stand by so please hurry up with the fix for improved manual control!
You should try implementing EMHASS and manage the batteries on your local network. It will also manage your deferrable loads if you can implement some smart switches.
I don’t have a power wall, I have a 10kWh sonnen battery with only 3.3 kW peak power. But I seem to be able to keep it in profit and charge an EV at the same time.
I’m in the same boat mate. Powershift seems to plan well and then keeps delaying the dispatch. When it finally starts it exports for a 10 min block and then self consumes again planning to dispatch but often delaying again. In a 90 min period it only dispatches half of what it planned and often misses the highest peak periods. With manual controls you get the start correct but even that stops and says it’s self consuming in the app despite saying you’re in control still. To my way of thinking Amber has a problem with smartshift. Once started why stop dispatch? Or does it default to self consume and Amber is required to reiterate dispatch every 10 min and if it can’t communicate with Tesla it reflects that in the app with “self consume”?
I think you are seeing the exact problem Amber have emailed about – it’s a lag between providing self management (or smartshift) instructions and the Tesla acting on them. This is due to a backlog of API calls.
It used to be very responsive, it’s now ~5-15 min lag. This means that smartshift decisions to despatch energy to make money are often late to start and late to finish, which is super frustrating.
My experience up until about 2 weeks ago was positive – it’s not perfect, it’s a bucketload better than the Tesla energy plan, and while I like the idea of EMHASS and other options to improve the sometimes dumb decisions of smartshift, I don’t have the bandwidth to dive into EMHASS right now.
An interesting week overall. Yes, I find the platform unstable, but when it works I’m chuffed. Fully on line since July, average monthly credit is $550. We peaked at $77 earnings for 1 day, but evidence of instability this week was $23, $45, $37, $8, $2. The $2 day the system totally crashed with not a single dispatch. I think the communications this week were good, a lot more transparent. I just want honesty and what is being done to resolve. My research says there is no better option so if Amber finally gets it right there will be bigger benefits to come. What I’m unclear of is what part Tesla plays in the delays and disruptions. Yesterday was a spike, and batteries fully dispatched so perhaps the latest fix has worked. I’m waiting on the Daily report to check earnings. I’ll keep researching for best option, but I want to stick it out for a full year and then start making month for month comparisons. After 60k invested, on track for a 6 year payback at the moment.
I ran into a similar problem Friday night with a 20kW AlphaESS battery.
SmartShift lost the plot, I had very limited manual control, import cost was high, battery was full but not covering household load (would only discharge ~150Wh). So it was costing me during the evening peak. Reported the issue but I only a basic auto reply “we’re busy onboarding new customers, please be patient”…have been with them just on 2 weeks. Appreciate all systems have teething problems and are not fool-proof (just ask Optus). Issue self corrected in about 24 hours, but still waiting on any proper investigation and human response.
Going to hang in there to see how it goes
Hi David, keen to hear how things go for you over next week to get a sense of any system improvements. It seems like you are paying close attention to your systems performance. I’m working on improved averages, with 3 month avg for FiT at 41.1c. If anyone is able to demonstrate a better way I’m open to ideas and will research, like the suggestion of EMHASS, I’ll see what that is about.
Keen to hear the channels thoughts on Amber Smartshift with the following setup.
Just over 15 kWh of panels with 15kWh 3-phase SMA Tripower S inverter and a Tesla Powerwall 2.
We also have 2x EV’s – a Tesla Model 3 and a MG ZS. We have a Smappee Smart Charger for charging those which is configured to only charge the car when the solar is exporting and there is a surplus of at least 3kWh.
“Bills” with EnergyAustralia have been $390 credit and $295 credit the last two quarters (last ending 2nd Jan).
It looks like Amber is REPS eligible here in SA with the rebate for a Tesla battery being $780 and we are also a CBA customer which gives us a $180 credit with Amber used up over six months.
We export between 1,200 and 2,200 kWh every month with our import being 200 – 500kWh the last couple months with the aircon running most nights. Before summer we barely imported anything.
We’re pretty diligent already in that the dishwasher only runs during daytime hours and the dryer is connected to a smart GPO which cuts the power to it between 9PM and 8AM.
We’ve also got gas hot water and our gas underfloor heating (which tends to in the dead of winter run overnight for a couple hours and maybe for an hour in the late arvo). Thinking about switching these to electric but I want 12 months of solar data to inform that (system was installed in July ’23).
What does the community think? Do you think I will make more money switching to Amber? Will it be “set and forget” or will I need to be tweaking loads all the time?
Cheers