Shai Agassi and the big batteries

TeslaShai Agassi, the California-based software superstar who wanted to run SAP but left the company in March when he didn’t get the top job, has come back into the spotlight as the CEO of an electric car start-up. The new company is funded by $200m of venture capital and investment bank money. This makes it one of the best-funded start-ups in history. Agassi does not intend to make electric cars. Wisely, he is leaving this to the auto industry. He is focusing on the batteries. He’ll lease them to anybody with an appropriate car and he’ll develop large networks of ‘filling stations’ where the driver can quickly take out a discharged battery and swap it for a fully charged version on long journeys. By 2010, he wants a hundred thousands electric cars on the roads of California and elsewhere.

The obstacles are huge. Although lithium-iron-phosphate battery technology is improving rapidly, and will continue to do so for decades, full-size car batteries now cost at least €7,000. Getting mainstream manufacturers to build large volumes of electric cars that will take his batteries is another formidable challenge. Third, he has to persuade retailers to install the equipment to swap batteries automatically.

But our weary European scepticism needs to be rested for a moment. The long-run economics favour this idea. My sums suggest that at current UK petrol prices it costs at least six times more to drive a mile on petrol than it does on electricity. Battery prices will fall and performance will improve. At some point it is going to be so much cheaper to power a car with electrons rather than octane that even the slothful auto industry will switch. When the market has tipped it won’t be long before passenger cars are all electric. Agassi may be too early, and his business model may require too much capital, but electric cars are coming soon.


Crushed EV1 carsGM designed an electric car in the 1990s. Several hundred EV1s were built and leased to generally enthusiastic owners. The cars were costly for GM to build, had a range of only 60 miles or so, and took a long time to recharge. The company abandoned the project and crushed all the cars.

Probably as a result of the perceived failure of the EV1, the idea of the all-electric car lost momentum. Attention switched to petrol-electric hybrids, such as the Prius, and the hugely eccentric idea of hydrogen fuelling. Conventional hybrids do not take electricity from the mains to power the car; they use petrol for much of the time. Energy is captured from the braking process and stored as electricity in an on-board battery to power the car when at low speeds.

More recently, engineers and car companies have started to investigate recharging the batteries of hybrid cars using mains electricity. These so-called ‘plug-in hybrids’ use electricity most of the time, particularly for town driving, and only have a petrol engine for long-distance travel or when recharging is impossible. The article on Hillary Clinton’s climate change plans in this edition of Carbon Commentary refers to her enthusiasm for plug-in hybrids. She also sees them as an important asset in smoothing the power needs of local electricity grids. When community demand is high and batteries are fully charged, the grid will start co-opting power from cars parked in driveways but connected to the mains.

Most major automobile manufacturers have development programmes for hybrid cars. Technological progress continues to be slightly disappointing. Light diesel cars now have CO2 emissions that match the Prius, although the Toyota car offers far greater comfort and is a more potent status symbol among the urban elite.

Small but increasing numbers of purely electric cars are being sold in the UK. The G-Wiz is increasingly visible on London streets, not least because it is taxed as a quadricycle, not a ‘proper’ car. The UK retailer says that it sells ‘one or two a day’ in London. 900 out of the 2,000 G-Wiz cars in the world are driven around the city, free of any congestion charge, and with electricity top-ups in some central London car parks. The range of the car is about 70 miles with a top speed of 45mph, though with commendable honesty the retailer stresses that it should probably be kept on congested, and slow-moving, urban roads. In fact, the average speed recorded by G-Wiz cars in London is a sedate 10mph.

Smart is just about to launch an electric version of its existing cars. Press comment suggests that the new Smart will manage a top speed of about 50mph and last for 70 miles between charges. At about 5 hours, charging time is still depressingly long.

The engineering of all electric cars is moving fast. Several models are slated for launch in the States that can manage far higher top speeds and longer cycles between charging. The Tesla is an expensive ($100,000) sports car that claims acceleration better than a Porsche at the same time as fuel costs of less than 1p per mile. The manufacturer claims that the lithium ion batteries are fully charged in three and a half hours and deliver a range of 245 miles. The marketing of this beautiful car glorifies speed – a far different approach to the virtuous but slightly dull retailers of cars like the £8,000 G-Wiz.

So what is Shai Agassi trying to do? The detail on Agassi’s scheme is sparse. The idea is to set up a network of sites, powered by solar energy at which drivers can swap batteries. The driver never owns the battery, but leases the right to have one in his or her vehicle. Most of the time, the car user will not need to switch batteries. They will usually be charged overnight at home, or during the day at the workplace, through ordinary domestic power sockets.

He has raised $200m from venture capital sources and from Morgan Stanley. One of the key investors is Vantage Point Ventures, also a backer of Tesla cars. For a start-up company, an injection of $200m is a huge sum of money, but the capital needs of a business like this are going to be phenomenal. At today’s prices, each battery costs €7,000, and if the entire sum raised were used to buy batteries, it could only equip a few tens of thousand cars. But of course the money won’t all go to buying batteries. This task – essentially a conventional leasing activity – will be taken on by Morgan Stanley once the business model is proved. Agassi’s venture needs the money to begin to build a network of re-charging sites.

The wisdom of the electric vehicle industry has always been that the battery is part of the car. Agassi says that drivers need not own the battery. The battery is the gasoline of an electric car, and you buy gasoline when you need it, not when you first buy the car. This is the key insight – removing the cost of the battery from the purchase price will bring the vehicle cost down, eventually to a lower price than the equivalently specified car. An electric vehicle doesn’t need an expensive engine and a troublesome transmission system. There may be as few as a dozen moving parts.

Crucially, Agassi believes that the auto industry is about to tip away from gasoline and towards electric power. His theory is based on three elements:

  • The price of oil is going to go up. ‘Peak Oil’ is here.
  • The operating characteristics of batteries for electric cars are all moving favourably. Cost and weight are down. Power (think of this as the capability to accelerate a car) and storage capacity (the distance you can drive) are going up.
  • Third, the cost of renewable technologies to generate the electricity for the battery is also falling fast.

The price of oil is going up The world oil market can’t quite understand what is happening to the price of a barrel. Industry experts mutter crossly about an unwarranted speculative premium. Nevertheless, the $100 hurdle looks as though it may be crossed, taking the oil price to a level not seen in real terms since the 1970s. Part of the reason that Agassi’s plan attracted substantial attention in the US is that the country is rattled by the spike in the price of this most vital of American commodities. In the UK and elsewhere, our interest in the price of oil is muffled by the falling value of the dollar and the limited impact seen so far at the petrol station.

In the 1970s the very high price of oil (even more important to the economy then than now) pushed politicians and auto companies into improving fuel economy standards. The fuel economy performance since the late 1980s has barely changed in the US. The increase in the price of gasoline may push automakers and consumers into more efficient models. This would tend to mute the future interest in Agassi’s scheme but the reaction times of Detroit seem to exceed those of any other major industry in the world.

Batteries are getting better The early electric cars were powered by lead acid batteries, the technology still used for starter motors in most vehicles. They were superseded by nickel metal hydride batteries. Recently, the industry began to experiment with lithium ion, the same technology that powers laptops and mobile phones. The performance has improved markedly, but concerns persist over the safety of large lithium ion arrays. It is not long since we had an outbreak of exploding laptop batteries and conventional lithium iron cells can, in theory, release a huge amount of heat very quickly.

Agassi is going for the next step up – lithium iron phosphate batteries. Here the risk of catastrophic malfunction is lower, although the performance is slightly less good than conventional lithium ion. Companies like Lithium Technology Corp already make car-sized agglomerations of lithium iron phosphate batteries. This company claims it can produce a 25 kWh battery weighing 200kg that would take a car 160 miles.

Agassi’s venture is reliant on what he sees as the Moore’s Law of battery improvements: a 50% decline in cost per kWh every five years. It seems a reasonable assumption to make. The industry is bullish it can achieve better performance with lithium iron phosphate and large amounts of US government money are being ploughed into companies in the field.

The costs of renewables are falling Agassi’s venture is vague on the third leg of the business case. It seems that he believes that the company will need to invest in large numbers of sites that recharge batteries. Drivers straying from their home or workplace will need access to fully charged cells. In California, his company can use solar electricity for the task of recharging. I am sure Agassi is right that the cost of solar photo-voltaic installations are likely to fall at least as fast as the cost of batteries. Within five years we are likely to see very substantial improvements in the price per watt as advances in silicon nanotechnology allow PV panels to be made from paper-thin silicon, rather than the expensive blocks of today.

Why does he need to recharge the batteries using solar energy? What is wrong with simply plugging into the conventional electricity distribution system? The answer seems to be that he is concerned that the Californian power grid is simply not robust enough to provide even the very limited amount of power that is likely to be needed. Frequent summer ‘brown-outs’ might make it impossible to obtain enough power.

The big problem is going to be the size of the solar installations necessary to recharge a bank of tens of batteries. Very approximately, current technology needs a square metre to collect a kilowatt. A battery needing 25 kWh of charge over a period of five hours would be taking 2.5 kW. Even at the middle of the day, when the panels will be working at full capacity, one battery will need almost 30 square metres of PV panels. Charge fifty batteries at the same time and you need the whole roof of a large commercial building. (Advances in solar PV technology will lower the cost, but may not decrease the surface area needed to generate electricity.)

In my view, the real reason why solar energy is part of the current plan is that Agassi needed a software component to the business plan to show why he, an SAP veteran, should run an automobile venture. The press statements make play of the software needed to run the ‘grid’ that he plans to link the recharging sites. I think that this is nonsense – the sooner he drops the idea of solar PV on every recharging point the better. It adds to the complexity and difficulty of the business. If solar energy makes financial sense in five years, then he can build a huge central facility to ‘offset’ the amount used to charge the batteries. To build solar-powered charging stations now is unlikely to make sense, and significantly raises his overall capital need – never a good idea in a start-up.

Does the idea make sense? If a 25 kWh battery can propel a car for 160 miles, then at current UK retail electricity prices, the fuel cost is 1.5p per mile. For a typical family car, the petrol cost may be close to ten times this level.

The average UK car does about 9, 000 miles a year. (The figure is much higher in the States.) The petrol cost is about £1,000, compared to less than £150 for a car whose battery is always charged at home at standard UK electricity prices. The fuel cost advantage is clear, at least in countries with high-price petrol.

What about the cost of the battery? At €7,000 (Agassi’s figure), the pay-back period would be about six years, if the cost of the battery were entirely additional. If, instead, the cost of the car were reduced because it no longer needs an engine or transmission, the economics may become really compelling. They are likely to become more so as time passes.

There is thus only one real question: can the famously persuasive Shai Agassi get volume car makers to build cars with no internal combustion engine, but with an electric motor and a hole where the battery goes? The chances of Detroit starting to eat its own lunch seem remote, but Asian car-makers may grasp the opportunity gratefully.

What is the effect on emissions of moving from petrol to an electric car? A typical UK car produces about 180g of CO2 per kilometre. (Remember that an internal combustion engine is only about 25% efficient at turning chemical energy into motion.) An electric car with an advanced battery running on a charge from the UK grid will have responsibility for average emissions of about 35g per kilometre, or about one fifth as much.

Why not the UK, rather than the US? Many UK cars never travel on long trips. Petrol prices are almost double the US figure. When charged at night, some drivers could benefit from the extremely low overnight power prices on dual-rate tariffs. Electric cars are exempt from London’s congestion charge (£8 a day). And, eventually, wind power can be used to charge the batteries at the refuelling stations on major roads. If Agassi’s idea makes sense in California, it has very compelling economics here.