For a corporate leader who arguably ranks as entrepreneur of the century, Elon Musk sure has a strange way of running his companies. So bizarre are his frequent outbursts, instances of supporting antisemitism and fringe views, and general eschewing of corporate civility that if he’d been a regular CEO instead of an owner and dominant shareholder, he probably would have been axed long ago by an uptight board. Musk lives in the extremes: one moment boasting astronomical, unachievable future valuations for Tesla, while in another calling a near wipeout for the shareholders of X. Yet somehow, this combustible elixir has endured while smashing the tablets you’d find in a Harvard Business School case study. Below, a distillation of Musk’s management ethos.

1. Promote the vision

Musk is expert at selling a futuristic vision where he revolutionizes the traditional profitability model for carmakers. Put simply, he’s messaging not to watch Tesla’s current, plateauing numbers too closely, because they’re irrelevant alongside the coming takeoff. The ability to spin such exciting narratives creates a leap of faith for Tesla investors, and the funds and superrich backing his ventures from X to SpaceX. If the EV maker’s shareholders didn’t largely buy his outlook, the stock wouldn’t carry a nonautomotive valuation of $630 billion even after its recent sharp decline. During one 2023 quarterly earnings call he stated: “I see a path to a 5x value for the company, maybe 10x.” The low end of that prediction would make Tesla 50% more valuable than Apple is today.

2、Keep promising groundbreaking innovations are almost here

It’s a Musk mantra that he’s constantly on the cusp of introducing revolutionary new products, often on a mass scale. It’s his way of convincing investors that he’ll continue changing the world, and keeping their eye on a gauzy horizon of never-before-seen profits. But he’s always pushing the arrival dates, then doubles down by pushing them back again.

Musk said in 2019 that the world would see 1 million robo-taxis on the roads by the following year. He pledged to deliver the Cybertruck in 2021, then revised the intro to 2022. He now says major output will start early next year. But the Cybertruck—forged from flat plates of stainless steel and looking like it was dropped by an alien race—is extremely difficult and expensive to manufacture, casting doubt on whether Tesla can profitably make it at scale. His glorious predictions that the Boring Company would dig a 10-mile tunnel under L.A. or that Tesla Energy would produce 1,000 solar panel systems per week are long forgotten, perhaps even by Musk himself.

3. Control every aspect of the manufacturing process

For Musk, business is principally about invention and engineering. His genius: Once he’s achieved an innovative design for a car or rocket, he’s expert at creating a super-low-cost, ultra-high-volume production machine. In the words of his biographer Walter Isaacson, he prides himself as the architect who “builds the factories that build the products.”

When he took charge at Tesla in 2008, the engineers who drafted blueprints for the components worked in offices separated from the assembly lines. Musk moved them to cubicles right on the factory floor so they’d witness up close when a design problem caused a slowdown and, in many cases, remodel the part in short order. Musk is a hawk on what he calls the “Idiot Index,” which could also be called a conversion ratio. It’s the measure showing how much a finished component costs to make versus the expense of the raw materials that go into it. That enables him to isolate what the stamping and other manufacturing chains are adding and, if it’s a lot, find a cheaper process.

According to Isaacson, Musk demands that every production and design engineer at SpaceX and Tesla know the precise number for every part they’re responsible for making or designing. It’s their job to push down the costs—or face his wrath. SpaceX makes 70% of its components in-house, in contrast to the typical aerospace industry practice of outsourcing most parts’ production.

In going heavily with vertical integration, Musk also departs from the stance of his hero Steve Jobs. Like Musk, Jobs worked on every aspect of his products’ design, but outsourced manufacturing of both parts and the final iPhones and iPads. By contrast, Musk views doing everything in his own closed loop as essential to controlling his destiny.

4. Pump volumes at all costs

For years, Tesla’s great brand, its head start in the EV race, and Musk’s knack for orchestrating superefficient manufacturing gave Tesla sumptuous margins. But now, Musk is no longer putting profitability first. As he noted on Tesla’s second-quarter earnings call: “Short-term changes in margins and profitability are really minor in the long-term picture of autonomy, and will make these numbers look silly.”

To gun sales, Tesla has been imposing round after round of price reductions. In the U.S., it’s lowered the sticker for the Model Y Long Range from $65,990 to $49,000 since early January. The sharp discounting has shrunk operating margins from 17.2% in Q3 of 2022 to 7.6% in Q3 of this year. Musk’s wager that autonomy will garner software-like profits may be a long shot. And if it’s a long shot that misses, Tesla will look like your average automaker, rather than sporting a valuation that’s almost three times Toyota’s.

5. Ignore conventional corporate financial metrics

Musk doesn’t talk about how he’ll grow the measures that create shareholder value, and that investors want to hear about. He never discusses how he’ll raise returns on equity or on invested capital, or set targets for those bedrock gauges. In fact, his approach to financial management can be downright irresponsible. In early 2021, he overruled his CFO at Tesla to purchase $1.5 billion in Bitcoin, most of which he dumped in 2022 at about breakeven, and SpaceX took a $373 million loss when exiting its position in the signature cryptocurrency early this year.

According to Fortune’s reporting, the banks holding $13 billion in loans to X are deeply frustrated that Musk is supplying little data on its operations. Musk fired X’s CFO shortly after taking charge, and so far hasn’t deemed the role important enough to find a replacement.

6. Spin ‘situational’ narratives to different constituencies

“We may fail, as many predicted,” Musk said recently about his adventures with X. He is constantly portraying X’s financial condition in the worst possible light. Why, you might ask? Simple: To purchase X, Musk borrowed a gigantic $13 billion from a consortium of seven banks. The big interest burden from the deal’s heavy leverage is crippling X and denying Musk the profits needed to build his envisioned “everything app.” But now the banks are stuck with below-market-rate loans they’ve been unable to sell off. The worse X looks to the banks, the better Musk’s chances of either buying the debt himself at a big discount, or getting creditors to greatly lower the principal amounts.

But Musk also has a list of two dozen coinvestors that he’d like to keep happy, a gilt-edged group that includes Prince Alwaleed bin Talal of Saudi Arabia and Marc Andreessen. So at the same time he’s bad-mouthing X’s short-term outlook, he’s singing its eventual promise. In May, he stated that X “could be someday worth $250 million” in a memo to employees that got lots of press coverage, and went one better in July, posting that “X will become the most valuable brand on earth.”

7. Make Jack Welch look timid when it comes to turnover

Musk reckons that a workforce that isn’t constantly churning and ejecting all but the most competent and hardworking employees isn’t maximizing efficiency. Surprisingly, his approach resembles that of 1980s management hero, former GE chief Jack Welch, who made a policy of terminating the lowest-performing 10% of his workforce each year (a tactic that has since been somewhat discredited). “Elon would pitch that it makes sense to let 10% of the people go every year,” recalls Branden Spikes, who worked closely with Musk as a SpaceX engineer from 2002 to 2012. “He’d say that if you don’t fire 10% and replace them with great people, you’re not running at peak potential. It was all math to Elon.”

Spikes recalls working for Musk as both brutal and inspirational. “He never fired someone without cause, and the cause didn’t have to be a big one,” recalls Spikes. “You could make one mistake a year and be forgiven, but not two.”

Spikes also notes that Musk tended to trust engineers far more than people with other backgrounds. “At SpaceX, almost 100% of the managers were engineers, including the HR and finance people,” he says.

8. Dispense with public relations

Musk dissolved Tesla’s PR department in 2019, and as of today, it’s the only public corporation remotely its size operating without one. None of his other holdings has a comms staff either. In an interview at Morgan Stanley in March, Musk mocked the discipline, quipping, “Maybe we should have a VP of propaganda or a VP of witchcraft, that would be a great one!” Musk has made himself both the source of all news about X and one of the loudest voices on the site, recently drawing a huge amount of anger and advertiser fallout after he supported an antisemitic post on X.

9. Get paid based on short-term stock price, not long-term performance

At the close of 2017, the Tesla board awarded Musk a gigantic, 10-year stock options package called the “2018 CEO Performance Award.” It consisted of 12 tranches, each vesting in steps as the market cap rose from the starting point of around $50 billion to $650 billion. In addition, he had to hit rising benchmarks along the way for either sales or profits to clinch each grant.

A blueprint designed to create a super-long-term incentive didn’t turn out that way. Tesla’s valuation zoomed so fast that Musk, by also hitting the majority of the combined sales and Ebitda goals, received all dozen pieces by mid-2021, just three and a half years into the program. Today, Musk still holds an astounding 21% of Tesla’s shares, worth around $137 billion, all harvested from the original plan, and he sold tens of billions in the EV maker’s stock to finance the Twitter purchase and fund his other ventures.

The rub: Musk no longer gets any options or restricted stock, a condition of the 2018 pay deal. So his incentive to improve Tesla’s operating performance is much lower than when he was reaching for the next grant. Indeed, Tesla’s profitability in the first nine months of 2023 is running well below the levels over the same span last year.

10. Fund a big vision with a side venture that makes money

At SpaceX, Musk’s great goal is the quest to take his Starship to Mars. But his rocket business loses money. So Musk found a way to subsidize his passion by launching the Starlink satellite business that has created a space-based internet network of 4,400 satellites. The system enabled Ukraine to communicate with the world despite the destruction wrought by Russian missiles.

Musk aims to expand the service eightfold to 30,000 satellites. Will he get there? Maybe, or maybe not. As long as he can keep followers enticed by his newest shiny object, it’s doubtful anyone will even remember he pledged to in the first place.