Fujitsu CEO: Finishing Fast Kills Revenue in the AI Era
Fujitsu's CEO says the man-day billing model that sustained Japan's SIer industry for decades is collapsing under AI. The company is pivoting to outcome-based pricing — a move that signals trouble for every contractor-driven IT firm in Asia.
The billing model that built Japan’s digital infrastructure is breaking
When Fujitsu CEO Takahito Tokita stood before journalists at a May 28 strategy briefing, he didn’t mince words about the company’s core business. “You can’t expect further growth from billing based on man-days,” he said, referring to the person-month model that has governed Japan’s system integration industry for four decades.
The subtext was sharper than the quote. Tokita was acknowledging that AI is not just improving productivity inside software teams — it is dismantling the economic logic that those teams have been paid under since the 1980s.
How man-day billing actually works (and why it creates perverse incentives)
In the man-day model, aSIer charges a client based on the number of engineers deployed and the months they work. More hours billed, more revenue. Slower delivery, higher income. The math is brutal in its simplicity: if AI lets one engineer accomplish in two weeks what previously took a team of twenty over six months, Fujitsu’s revenue from that engagement drops by roughly two-thirds.
This is not a marginal efficiency gain. It is a structural contradiction at the heart of the SIer business model. The better Fujitsu gets at delivering software, the less it earns.
Tokita put it plainly: “Speed is the most important thing for any enterprise today. The shorter the month count, the better.” For the client, faster delivery is pure value. For the SIer billing by the hour, faster delivery is a revenue event.
Fujitsu’s answer: stop selling time, start selling outcomes
The company’s medium-to-long-term vision, framed through 2035, sketches a pivot away from time-based billing toward two new models.
The first is usage-based pricing for platform services. Instead of charging for the engineers who build and maintain a system, Fujitsu plans to bill based on data volume, computational workload, and resource consumption. This shifts risk from the SIer’s overhead to the customer’s actual usage patterns — a more aligned structure, but one that requires Fujitsu to own platforms rather than just deliver projects.
The second is revenue sharing on professional services. Here, Fujitsu would take a cut of the business results its solutions generate for the client, rather than a fixed fee for hours worked. “Actually, your company understands better than we do what the right compensation should be,” Tokita told reporters when asked about pricing in this model. That line is significant. It signals a willingness to let customers set the terms — a dramatic departure from the opaque, relationship-driven pricing that has characterized much of Japan’s SIer market.
The workforce implications are already visible
Fujitsu’s global headcount has fallen from 124,000 to 99,000 over recent years. The company attributes part of this to portfolio restructuring and natural attrition. But the direction of travel is unmistakable: the mass-deployment model of staffing dozens of mid-level engineers onto long-duration projects is shrinking.
Tokita went further in the Q&A, suggesting that in cutting-edge software development, a single elite engineer wielding AI could produce work equivalent to a hundred people. The implication for hiring practices is stark. Fujitsu has already abandoned its traditional annual mass-recruitment cycle in favor of hiring for specific skills. The era of bulk-coaching young graduates into generalist programmers — a cornerstone of the Japanese SIer labor pipeline — is ending.
Why this matters beyond Japan
Japan’s SIer industry is one of Asia’s largest employers of software engineers. Companies like Fujitsu, NEC, Hitachi, and NTT Data have built ecosystems around long-term, relationship-based contracts that prioritize job security and gradual skill development over speed and specialization. The man-day model is not just a pricing strategy; it is a social contract.
When AI collapses the link between effort and output, that contract frays. The question is what replaces it.
Fujitsu’s pivot toward outcome-based pricing mirrors trends already visible in Western IT services. Accenture, IBM, and Deloitte have all experimented with value-based and revenue-share models. But Japan has been notably slower to adopt them, partly because the relationship dynamics of the SIer market reward loyalty and longevity over innovation. A CEO publicly declaring that the old model is unsustainable is, in that context, a signal of real pressure.
Who wins, who loses
Clients that can negotiate revenue-share or outcome-based deals stand to gain the most. They get faster delivery and pay only when value materializes. Engineering firms that can genuinely automate development with multi-agent systems will capture margin that previously went to billing hours. Elite individual contributors who can wield AI at scale will command premiums that mid-level generalists have never seen.
The losers are clearer. Mid-tier SIers built on staffing arbitrage — companies that win bids by offering bodies at competitive rates — face a narrowing path. If AI compresses the number of engineers required per project, their entire cost structure becomes unviable. Small and mid-sized SIer contractors who lack the platform technology to offer outcome-based pricing will be squeezed from both sides: pressed on price by clients demanding faster delivery, and pressed on margins by the reduced headcount each project requires.
What happens next
Fujitsu announced its vision in late May. The company has not yet disclosed具体 pricing commitments or contract structures for the new models. Whether revenue sharing becomes common practice — or remains a marketing label — will depend on whether Fujitsu can build trust with clients willing to let it share in upside gains.
What is clear is the direction. The man-day model did not die from incompetence. It died from success: SIers spent decades getting better at delivering software, and now the tools exist to get even better, making the old billing logic self-defeating. Fujitsu is trying to race ahead of that contradiction. How far ahead, and how many competitors will follow, is the open question for Japan’s tech industry.
The era of billing by the month is closing. The era of billing by the result is beginning — and neither Fujitsu nor its clients fully know what the math looks like yet.