Day10 is developing a software delivery model that it believes reflects changes in how engineering services may evolve. Co-founder and CEO Konstantin Tsybulko and founding investor Yury Yakubchyk outline their view that the sector is entering a period of transition.
On paper, their move may seem unexpected. Tsybulko previously helped scale a custom software firm from approximately $35 million to nearly $200 million in ARR. Yakubchyk is a founder and investor whose ventures have raised over $400 million from firms including Sequoia, SoftBank, Founders Fund, and Tiger Global, with one company reaching unicorn valuation.
Despite this, they chose to build a services-focused company in 2026 rather than pursue a product or AI-tooling approach.
Konstantin Tsybulko and Yury Yakubchyk see it differently: this is what you build when you understand what's actually breaking in software.
Their new company, Day10, launched in February 2026 with a pointed thesis. Tsybulko argues that AI is fundamentally misaligned with key aspects of the traditional software engineering services model. And the companies best positioned to see that - the dev shops themselves - are the least capable of doing anything about it.
The incentive trap
The traditional software services model is largely based on headcount and billable hours, with revenue often tied to team size and project duration.
AI introduces a shift by enabling faster code development, which can reduce the need for larger teams and extended timelines. For firms built around time-based billing, this may present structural challenges.
"You can't reconcile it," says Tsybulko, who served as Chief Revenue Officer at Vention, a global development agency where he managed go-to-market teams of over 100 people across the US, UK, Germany, and Austria. "AI wants fewer humans. Dev shops want more billable hours. The incentives are structurally opposed."
Tsybulko saw this tension play out across the industry. As agencies began introducing AI coding tools, adoption among engineers was often slower than expected. Not because the tools were bad. The opposite.
"For 15 years of my career, developers always pushed to buy tools that made them faster," he says. "But with AI, something flipped. Adoption of tools like Cursor and Claude Code was shockingly slow. I realized a lot of engineers weren't resisting bad tools – they were resisting a fundamental change to what their job actually is."
From client to founding investor
Day10 emerged from ongoing discussions between Yury Yakubchyk and Konstantin Tsybulko about how AI might reshape software delivery.
At the time, Yakubchyk was a client of the agency where Tsybulko worked, and their conversations highlighted a shared view that existing models were becoming increasingly misaligned with emerging technologies.
Tsybulko also observed internal challenges around how AI services were positioned, priced, and integrated into existing workflows, which contributed to his decision to move on.
Over time, these discussions evolved into a more defined direction, ultimately leading to the creation of Day10, which positions itself as focused on execution with active projects underway.
A different kind of engineering firm
Day10 operates through what it calls “AI Augmented Pods”—small teams of senior engineers supported by integrated AI tools and workflows. Instead of larger project teams, the model typically involves a product engineer and an AI architect working within a unified toolchain.
According to the company, internal benchmarks suggest higher productivity compared to traditional teams, although these figures have not been independently verified. Tsybulko draws a distinction between what he calls “AI-decorated” companies — firms that layer AI tools onto existing workflows — and companies built AI-first from the ground up. He argues that the difference is not in tooling, but in how teams are structured, how work is executed, and how incentives are aligned.
Day10 also uses an outcome-based, fixed-price model instead of hourly billing. “We price on outcomes, not activity,” Tsybulko says, noting that this structure is intended to align delivery timelines with client expectations.
The talent bottleneck – and the hiring advantage
Finding engineers who fit Day10’s model is one of its main challenges. The company screens extensively to identify candidates with several years of experience, strong academic backgrounds, and an openness to new ways of working.
“That combination is relatively rare,” Tsybulko says.
Once identified, however, recruitment tends to be more straightforward. Tsybulko notes that some engineers feel their current companies have been slow to adapt to AI-driven workflows, which can make alternative approaches more appealing.
Day10 primarily recruits from Latin America, with a smaller presence in Europe. The strategy reflects both access to skilled talent and the practical challenges of competing for engineers in highly competitive U.S. AI markets.
Early traction before the official launch
Before its website launched, Day10 had three paying clients and several others in advanced discussions, working with veterans of the AI space and AI-focused teams seeking faster execution.
The company is focused on sectors such as healthcare, drawing on Yakubchyk’s experience, as well as real estate, insurance, and legal, where complex problems and available budgets often intersect.
Tsybulko notes that these areas tend to offer both technical challenges and commercial viability.
For Yakubchyk, whose Cold Start incubator has backed multiple companies, Day10 addresses a recurring issue he encountered as an investor.
"When I invest in AI-first companies, I'm betting on execution as much as the idea," says Yakubchyk. "Slow or inefficient delivery increases burn and kills momentum. Day10 is how I make sure capital turns investment into AI products more quickly."
Why this shift is different
Tsybulko has navigated technology transitions before – mobile, SaaS, cloud, blockchain. Each wave created new categories of software and new customers, which meant more work for development agencies. But none of them changed how software was fundamentally built.
"Every previous shift brought more demand and more work," he says. "Mobile meant you needed Objective-C developers. SaaS meant you needed cloud architects. But the approach was the same: humans writing code, line by line. This time is different. You write less code. You read and edit more. You orchestrate agents. The entire process of creating software is changing, not just what you're building."
The comparison he keeps returning to: Boris Cherny, creator of Claude Code at Anthropic, described a project he once worked on at Meta that took 20 engineers and two years to complete—and estimated the same work could now be completed by five people in six months using agent-driven workflows. Tsybulko says Day10's own benchmarks tell a similar story, and that their workflow is inspired by the broader agent-orchestration approach described by Cherny.
"That's the real differentiator. It's not about which AI tool you're using. It's about whether your entire delivery system was designed for this era or bolted on top of the old one."
Building the machine that builds
Day10 is positioned not as another agency, but as an effort to rethink how software is built in the AI era. Its argument draws a comparison to earlier industry shifts: legacy models may still function, even as alternative approaches begin to emerge.
“The most interesting opportunity in 2026 isn’t another AI product,” Tsybulko says. “It’s building systems that allow smaller, AI-native teams to operate more efficiently than traditional ones.”
Whether that approach holds up at scale remains to be seen. But the underlying idea is clear: as coding becomes more commoditized, greater value may shift toward the systems that manage how software is developed.
Day10 reflects a broader view that traditional, labor-intensive models may give way to more streamlined, technology-driven approaches and smaller, AI-native teams.
VentureBeat newsroom and editorial staff were not involved in the creation of this content.