A company that wants to swap out foreign radiotherapy software just got ₩16B

Here is the deal: about one in three cancer patients gets radiation therapy. But the software that decides where the beam goes, how much of it, and at what angle — the "treatment planning system," or TPS — has been almost entirely foreign for decades. Companies like Varian (Eclipse) from the US, RaySearch (RayStation) and Elekta (Monaco) from Sweden have owned this market for years. Korean hospitals have been paying pricey license fees to run cancer treatment on somebody else's engine.

Then on July 2, 2026, a Korean startup that wants to flip that table loaded up on ammo. Oncosoft, an AI radiotherapy software company, announced it raised a ₩16 billion (~$10.4M–$11.6M) Series C from eight investors led by HB Investment. And the headline isn't just "startup got money." It's that Oncosoft pinned this cash to a specific mission: building a homegrown particle-therapy TPS to break the foreign lock on a market almost nobody else can enter.

The timing is almost too clean. Korea's government just tripled its 2026 AI budget to ₩9.9 trillion and declared this the "founding year of the AI G3 leap." With money flowing into deep-tech and AI startup funds, this deal lands right in the center of a "government pushes, market catches" picture. So let me walk through who this company is and what it is actually trying to do.

The players — a medical physicist's company, and eight backers

Oncosoft was founded by CEO Kim Jin-sung. He is not a generic developer-turned-founder — he is a medical physicist and a professor of radiation oncology at Yonsei University College of Medicine. A medical physicist is the person who takes physical and engineering responsibility for firing radiation into a human body safely and precisely. This is someone who has actually worked with both proton and heavy-ion therapy machines, and in 2019 he decided the software for this field could be built at home. A company built by someone who knows the clinical floor — that's the backbone of this story.

The round was led by HB Investment, joined by Devsisters Ventures, COMES Investment, Horizon Investment, WantedLab Partners, G.N.TECH Ventures, Timeworks Investment, and Korea Investment & Securities — eight backers in total. It is notable that the venture arm of game maker Devsisters showed up in medical AI, and it is meaningful that a securities-firm affiliate like Korea Investment & Securities is in the mix. When a brokerage arm joins, it usually smells like a late-growth-stage deal with an IPO in mind.

Oncosoft's flagship product is OncoStudio, an AI "auto-contouring" tool that automatically draws the boundaries between tumors and healthy organs on CT scans. This job used to require a doctor or therapist to hand-draw each slice, one at a time — tens of minutes to several hours per patient. OncoStudio got Korea's MFDS (drug safety ministry) approval in February 2022 and is now used for clinical and research work at about 50 hospitals and institutions nationwide. In other words, this is a company with a real, revenue-generating product, not a slide deck.

The funding history tells you how it grew. Roughly ₩3.5B in a Series A (August 2022), ₩11B in a Series B (September 2024), and now ₩16B in this July 2026 Series C. Each round scaled up sharply, pushing cumulative funding past ₩30 billion. That's the signal of a company moving from "early idea" to "scale it up" mode.

The core isn't contouring — it's localizing the particle-therapy TPS

Here's what you shouldn't misread: the heart of this raise is NOT the already-selling OncoStudio. The real battleground is the next product — the particle-therapy Treatment Planning System (TPS). Radiation therapy splits into two broad camps. One is the familiar X-ray (photon) therapy. The other is "particle therapy," which fires heavy particles like protons or carbon ions. Particle therapy can deposit energy precisely into the tumor while sparing healthy tissue, so it's celebrated as next-generation cancer care. But the software to plan it is far more complex than the X-ray kind, and only a handful of companies worldwide can build it.

Oncosoft says it will use this money to develop an "AI-based domestic particle-TPS core engine" through a government-industry joint commercialization project. It's participating in national research institutes' efforts to localize radiotherapy equipment, and its ultimate goal is an integrated TPS spanning X-ray, particle beam, and even FLASH therapy (ultra-high-dose-rate irradiation). CEO Kim Jin-sung said, "At a time when particle therapy is spreading rapidly worldwide, commercializing a particle-therapy TPS with our own domestic technology is a source of great pride."

Why does this matter so much? Because the TPS is the "brain" of a radiotherapy machine. You can localize the hardware all you want, but if the planning software running on top is foreign, your "domestic" system is only half done. What Oncosoft is chasing is the last piece of the localization puzzle — software independence. The cash goes to three places: next-gen product (particle TPS) R&D, global market expansion, and recruiting key talent.

Item Detail
Round Series C
Size ₩16B (~$10.4M–$11.6M)
Announced July 2, 2026
Lead investor HB Investment
Other investors Devsisters Ventures, COMES Investment, Horizon Investment, WantedLab Partners, G.N.TECH Ventures, Timeworks Investment, Korea Investment & Securities
Flagship product OncoStudio (AI auto-contouring, MFDS-cleared Feb 2022)
Next battleground Homegrown particle-therapy TPS
Hospitals deployed ~50 in Korea
Cumulative funding Over ₩30B
Founded 2019, CEO Kim Jin-sung (Yonsei Univ. radiation oncology)

What each side gets

Oncosoft gets a clear win. It carries a revenue-generating cash cow in OncoStudio while securing ammo to move into the much larger particle-TPS market. For a startup to expand from one validated product into a next-gen lineup, it needs money, time, and people all at once — and ₩16B buys a runway of several years. Layered on top of a government localization project, Oncosoft isn't shouldering pure private risk; it's developing under the umbrella of a national program. That's a big deal.

The investors boarded a deal where three beats line up: medical AI, localization, and government backing. The presence of a brokerage arm like Korea Investment & Securities specifically reads as a signal that this company is seriously eyeing an IPO path, likely via Korea's tech-special listing track. With rounds scaling from A through C, late-stage backers see a deal where the exit is visible. Of course, that valuation only gets justified if particle-TPS commercialization actually succeeds.

Patients and hospitals stand to benefit long-term on cost and choice. Until now, Korean hospitals paid big annual license fees for foreign TPS and were tied to overseas HQ schedules for customization and support. A domestic alternative creates negotiating leverage and faster feature tuning for Korean clinical settings. That's contingent, though, on Oncosoft's products proving foreign-grade stability and accuracy.

The government gets a candidate success story that fits the "AI G3 leap" slogan perfectly. One justification for spraying ₩9.9 trillion of AI budget is "deep-tech localization," and radiotherapy software is a field with high regulatory walls and steep technical difficulty that not just anyone can enter. If a domestic champion emerges here, it becomes a symbol that validates the spending.

Past parallels — the wins and the flops

On the success side, think of Korea's medical-AI imaging companies. Firms like Lunit and Vuno cleared MFDS and FDA approvals for chest X-ray and pathology AI and made it all the way to public listings. They cleared both the clear value ("AI reads the image") and the trust threshold of regulatory clearance, then went global. Oncosoft already has MFDS clearance and a 50-hospital reference base with OncoStudio, so it has the potential to walk the same track.

On the flip side, there's a failure pattern to respect. Medical AI is famous for the "dazzling demo that never sticks in the clinic" trap. IBM's Watson for Oncology is the textbook case — it was sold to hospitals worldwide as a cancer-treatment-recommendation AI, but it couldn't clear the real-world complexity of clinical practice and regional variation in treatment norms, and the business was eventually wound down. The lesson: "is it used every day on the floor?" is the real test, not a flashy AI narrative.

Also, planning software like a TPS faces far higher regulatory and validation walls than diagnostic AI. Get a read wrong and you can re-check it; get a radiation plan wrong and the harm goes straight to the patient. That's why this field has extremely few new entrants and, once you're established, you're very hard to rip out — it's a "sticky" market. For Oncosoft that's a double-edged sword: entry is brutal, but breaking through builds an equally tough moat.

Past localization attempts carry a lesson too. Medical-device localization has repeatedly hit the wall of "the tech works, but hospitals won't switch." Hospitals are reluctant to drop validated, familiar foreign tools. The fact that Oncosoft has already crossed the hospital threshold with OncoStudio gives it an edge over a pure greenfield localization challenger.

How rivals counter-play

The biggest opponents are the global TPS heavyweights. Varian (part of Siemens Healthineers), RaySearch, and Elekta carry decades of clinical data, worldwide reference sites, and tight integration with their own treatment machines. If Oncosoft goes head-on with an X-ray TPS, they'll defend with price, bundling, whatever it takes. That's exactly why Oncosoft aiming at "particle-therapy TPS" — a niche with fewer players and less-settled standards — is a smart detour rather than a frontal assault.

Those global players are also rapidly bolting AI auto-contouring and auto-planning onto their own products. Meaning OncoStudio's "AI automation" edge could get diluted over time. So contouring alone isn't enough; the strategy is to bundle an end-to-end platform — from contouring to planning, workflow management, and radiopharmaceutical therapy (RPT) — to raise switching costs and make customers stick.

The domestic competitive picture matters too. Korean firms specialized in radiation-oncology software are still countable on one hand, so Oncosoft brands itself "Korea's only integrated radiotherapy software company." But big medical-AI names like Lunit and Vuno could expand from adjacent areas into radiotherapy, and universities or research institutes could ship similar tech via open-source or national projects. The government-project umbrella is sturdy, but competitors can shelter under that same umbrella.

FLASH therapy and other ultra-next-gen techniques are wild cards too. Oncosoft's plan to preempt an integrated TPS across X-ray, particle, and FLASH is a bet to plant a flag first on future tech that has no standard yet. Lead here and you can make the global giants play catch-up; get too far ahead of a market that never opens, and you just burn development cash.

So what actually changes

For cancer patients and families, nothing changes you can feel right now. OncoStudio is software running behind the hospital scenes, so patients never touch it directly, and the particle TPS is still in development. But long-term, the more homegrown software takes root, the faster treatment planning gets (thanks to automation) and the lower the hospital's software cost burden — which can indirectly help treatment access.

For Korean hospitals and clinicians, the core change is one more option on the table. They've been effectively locked into foreign TPS; a mature domestic alternative brings licensing leverage and faster customization for Korean clinical settings. Auto-contouring already cuts repetitive labor — something 50 hospitals feel today.

For investors and the industry, this is a signal that "medical-AI localization" is a theme where real money moves alongside government budgets. With a brokerage arm in a Series C, it's worth watching Oncosoft as a future tech-special-listing candidate. Just remember: until the hard goal of particle-TPS commercialization is actually met, the valuation rests on expectation.

For the startup ecosystem, this deal sketches a blueprint. Build revenue and references with a validated product (OncoStudio) → use that trust to attack a much bigger market (particle TPS) → spread risk with a national project. It's close to the textbook route for a deep-tech startup surviving in a regulated industry.

🥄 Three Things You are Probably Wondering

— So what does this mean for me? Not much right now. But if you ever need radiation therapy, whether the software planning it is domestic or foreign could affect treatment speed and your hospital's cost structure. Think of this as the early moment when that board starts to shift.

— Why did this happen now? It rode the wave of the government tripling its 2026 AI budget to ₩9.9 trillion and declaring an "AI G3 leap." With money pouring into deep-tech and medical-AI localization, Oncosoft — which already has a product and hospital references — got its next round of ammo. The timing is no accident.

— Is it actually ahead of the foreign giants? Too early to say. The particle-therapy TPS Oncosoft is chasing is still in development, and the clinical data and machine integration of Varian, RaySearch, and Elekta remain overwhelming. But by aiming squarely at less-settled areas like particle and FLASH therapy, it has at least claimed the strategic position of "digging into the gap first."

References

Numbers and criteria are as of announcement and may change. Investment calls are yours to make!