Unlock Editor’s Digest Lock for Free
FT editor Roula Khalaf will select your favorite stories in this weekly newsletter.
A year ago, generation AI enthusiasts cleaning Silicon Valley and Wall Street faced serious reality checks.
In a widely cited memo, Goldman Sachs Head of Equity Research Jim Cobello questioned whether companies planning to spend $1 billion on building a generated AI could see the profits of money. Meanwhile, partners at Venture Capital Firm Sequoia estimated that tech companies would need to generate $600 million in additional revenue in 2024 alone to justify additional capital expenditures.
This warning has helped spark the first real test of investment sentiment since the launch of ChatGpt electrified the industry. Revenues from end customers intended to benefit from this new technology were negligible. Where was the “killer app” for the generator AI? That has led to summer anxiety for high-tech investors.
A year later, the major AI strains go through another volatile swing. Shaking the worries that began with a low-cost AI model created by China’s DeepSeek, Nvidia rebounded this week to score a new record high. Microsoft has registered a market capitalization of $100 million within three months.
However, what is noteworthy is that since the warning a year ago, the broader revenue outlook for generation AI has changed little. Hope (and hype) is still strong as ever, but it is still difficult to see where returns come from, at least in the short term, to justify huge capital expenditures on AI.
In terms of cost, all the benefits of AI enthusiasts are clear. Four tech companies leading rates, including Alphabet, Amazon, Meta and Microsoft, increased capital expenditure by two-thirds ($95 billion) in 2024.
Some content could not be loaded. Please check your internet connection or browser settings.
It is expected to further escalate the rest of the decade. Bank of America Securities predicts that across the technology industry, data center spending will jump from $333 billion last year to around 1tn in 2030. By the end of the period, 83% of money will become AI-related investments.
Meanwhile, in terms of the revenue side of the equation, some AI leaders are beginning to see a major increase in their business, but the extra revenue counts in tens of millions rather than tens of billions.
Earlier this year, Microsoft said its annual return rate from AI rose 175% to $130 billion. This represents only about 5% of the total revenue expected to generate this year. Openai’s revenue from subscriptions, the main source of revenue, doubled since the end of last year, exceeding $10 billion. The rate of increase is noteworthy, but the absolute numbers are still pale and white compared to CAPEX.
Also, for a year ago, there have been signs of an explosion of chatbot use, turning Openai into an unlikely consumer technology company almost overnight. However, while many people are using AI chatbots now, the business remains small. According to a survey of 5,000 American adults by Menlo Ventures, only about 3% of AI services they use pays AI services, with annual revenues of around $12 billion.
Recommended
Meanwhile, the business world is still looking for the use of generated AI to justify serious spending. The technology has fueled at least a new generation of high-growth software startups, especially in coding, the first area of ”knowledge work”; However, it has not caused changes in revenue growth for the largest software companies. This is best positioned to bring AI into the business world in the form of a new generation of AI-powered apps.
The first wave of AI co-pilots and assistants rarely changed work life. Hope is now shifting to agents – a tool that can automate individual tasks, and even the entire work process. According to McKinsey, agents are committed to creating serious business value by automating complex and critical operations. However, the consultants also warn that this will require a rethinking of the entire business process.
Convincing customers to move beyond the many Genai pilot projects that are scattered across the corporate world would be a heavy lift for the tech industry. That doesn’t mean that in the long run, generative AI is not likely to bring about the kind of work life transformation boosters claim. However, for now, the cracks between capital expenditures and income show little signs of stricture.
richard.waters@ft.com
