Silicon Valley's AI-fuelled madness has echoes of the dotcom crash

Times

2 days ago

I returned to work last week after a brief sabbatical and Silicon Valley, in my absence, has gone and truly lost its mind. Tales of $100 million — even $200 million — pay packages for software developers ricochet through the Valley as AI-induced mania scales new heights. The value of chip giant Nvidia has just crossed the $4 trillion threshold, a 12-fold increase in just 30 months that has turned an estimated 80 per cent of its workforce into millionaires.
And then, of course, there are the start-ups. Thinking Machines Lab, a company started by OpenAI co-founder Mira Murati a mere five months ago, has raised $2 billion in funding at a valuation of $10 billion (£7.5 billion). This company has neither product nor revenue. What it does have is a collection of big brains who once worked at OpenAI. And that, apparently, is enough.
• OpenAI's former tech boss Mira Murati launches own start-up
Ilya Stuskever, another OpenAI refugee, also reeled in $2 billion for his company, Safe Superintelligence — but at a, checks notes, $32 billion valuation! It has yet to trouble the world with a product.
History does not repeat itself, as the trope goes, but it rhymes. Which is why the madness to which I have returned takes me back to December 1999. I had just graduated with a liberal arts degree and a surfeit of energetic naivety.
The internet boom was in full bloom and so I hatched a cunning, three-stage scheme: move to San Francisco, get a job at a dotcom (any dotcom would do), and then watch my share options explode in value, turning me into a millionaire by the age of 25.
You'll be surprised to learn that my watertight plan sprung a few leaks. I did move to San Francisco, and I did land a job at a buzzy dotcom that was so desperate for warm bodies, it looked past the rather underwhelming skills implied by my degree in international relations and Spanish. I lasted four months.
The dotcom bubble burst, my employer quickly realised it had no chance of ever raising money again, and deemed me — and much of the company — surplus to requirements. It went under not long after, amid thousands of others.
The experience was, in retrospect, a fabulous internship for the business reporter I would become. Seasoned business leaders, top bankers, the world's savviest investors … they had all drunk the Kool-Aid. Otherwise intelligent and canny operators had been swept up in a type of group hysteria, fuelled by a powerful cocktail of optimism, magical thinking and greed. Who needs profits when we are about to change the world?
Sound familiar?
So is this time 'different', as anyone with a PowerPoint deck and an AI idea to sell would have you believe? I think two things are true. AI will, indeed, change just about every industry and how we live life, in ways large and small. Yet it is also true that we are caught in the midst of a gigantic bubble that, when it pops, will lead to the disappearance of most AI companies and wipe out trillions of dollars of wealth in private and public markets.
The latter, after all, is not a big leap. Since the start of 2023 through last month, the 'Magnificent 7' tech stocks — Apple, Microsoft, Alphabet, Meta, Nvidia, Amazon and Tesla — have gained $9.7 trillion in market value. This bears repeating: that figure is the combined value increase, not the overall sum. Put another way, the rise in value of those seven companies over these past two and a half years is equivalent to recreating the entire listed universe of 2,000 companies on the London Stock Exchange — twice.
Downstream of those astronomical figures is the bubbling cauldron of Silicon Valley start-ups, which are taking advantage of the mania to raise eye-watering amounts of money at valuations that most have no hope of earning their way into. Indeed, last year alone, Californian start-ups raised nearly $50 billion across more than 850 financing deals, according to data from PitchBook.
The upshot is that money is sloshing around to a degree not seen since the internet turned the world upside down. And amid the madness, the biggest beneficiaries are, without question, software nerds.
What a time to be a coder!
The top keyboard jockeys are being courted with pay packages to rival Premier League footballers. OpenAI's Sam Altman made headlines recently when he said Meta's Mark Zuckerberg was offering $100 million bonuses to lure away his top talent. Andrew Bosworth, Meta's chief technology officer, poured cold water on the claim — sort of. 'Sam is just being dishonest here,' he reportedly told an internal meeting. 'He's suggesting that we're doing this for every single person … Look, you guys, the market's hot. It's not that hot.'
So, not every single person, but some?
Indeed, just look at what Meta has done in the past month. It paid $14.3 billion for 49 per cent of Scale AI, effectively 'acqu-hiring' the engineers at a start-up that labels training data for AI model developers.
Under the deal, 28-year-old founder Alexandr Wang joined Meta's 'SuperIntelligence Labs'. Zuck's bags of cash have lured a handful of senior OpenAI coders to the new unit. He also poached Daniel Gross, the 34-year-old chief executive of SSI, Sutskever's start-up, as well as Nat Friedman, the 47-year-old former chief of Github. Gross and Friedman ran an investment fund together, and it is understood that Meta bought out their stake for about $1 billion to bring them into the fold.
Just days after Bosworth's luke-warm denial, reports emerged of an even richer deal. Ruoming Pang, Apple's former head of AI models, was said to have joined after Zuckerberg offered him a multi-year deal worth $200 million.
One can understand why so many engineers are happily taking Zuck's money. The Silicon Valley machine is, by design, messy and unpredictable. OpenAI, for example, has jumped from a $1 billion valuation in 2019 to $300 billion today.
Anthropic, maker of the Claude chatbot, was founded just four years ago and has already seen its valuation soar to $60 billion. Sales, meanwhile, are on track to top $4 billion over the 12 months — from a standing start in 2022. OpenAI has forecast $13 billion in revenue for 2025, all from a suite of products that did not exist three years ago.
The mind boggles. And yet, soaring growth today does not guarantee success tomorrow. Remember when Yahoo was the master of all it surveyed? Remember when MySpace dominated social media? For every ten companies that get venture capital backing, roughly five will die, four will limp on, merge or get bought for a modest sum, and one, if its investors are lucky, will turn into a properly successful company. Perhaps it becomes a unicorn – until it is disrupted by the next white-hit startup. The sands are always shifting, so when offered life-changing, generational wealth, many are taking it.
And yes, Zuck has spent $15bn-plus in a matter of weeks for a battalion of sun-starved coders, but old hands see his moves as utterly logical.
Because if indeed AI is the generational technology that Silicon Valley believes it is, those who grab control of it stand to make hundreds of billions, even trillions, as it replaces humans and takes over more and more of the economically valuable work.
Reid Hoffman, the billionaire LinkedIn founder and early Meta investor, said last week: 'The talent race, to your average American, looks crazy — the amount of money you're paying individuals in order to do this.' But he added: 'If you invent the thing that essentially — for example, my own start-up, Manas AI, is trying to cure cancer — transforms industries, and if you think this individual is the one to do it, then it begins to get more economically rational.'
• Don't fear AI: used well, it can empower us all
There is, of course, another option — that AI sets off a more gradual and messy transformation, much like the internet just before the dotcom blowout, when most companies crashed and burned but a few generational start-ups emerged from the chaos.
That appears to be the view of Peter Thiel, another billionaire techie. 'My place holder is that it's roughly on the scale of the internet in the late '90s,' he said recently. 'It might be enough to create some great companies. And the internet added maybe a few percentage points to gross domestic product, maybe 1 per cent to GDP growth every year for 10, 15 years. It added some to productivity. That's roughly my place holder for AI.'
So, no revenue. No product. No problem.
Until, of course, it is.