Y Combinator’s Summer 2024 Request for Startups, “A way to end cancer,” was not a promise that one startup could eliminate cancer. It was an invitation to pursue unusually ambitious, technically grounded ideas across the cancer-care pathway—from prevention and diagnosis to treatment, monitoring, and patient support.
The request remains useful in 2026 because it captures a difficult category for builders: cancer startups need scientific credibility, clinical evidence, regulatory discipline, and a business model that can survive long development cycles. A compelling pitch is not enough. Founders must show that the problem is real, the proposed intervention can work, and the route to adoption is practical.
What YC’s request was really asking for
The strongest interpretation of the brief is not “build another oncology app.” It is: find a high-impact bottleneck in cancer care and develop a defensible solution that can reach patients. That could involve:
- Earlier detection for cancers that are usually diagnosed late.
- Better treatment selection using molecular, imaging, or longitudinal patient data.
- Therapies that improve efficacy while reducing toxicity.
- Faster and cheaper drug discovery or clinical-trial operations.
- Tools that help oncologists manage complexity without replacing clinical judgment.
- Better access, adherence, navigation, and supportive care.
- Prevention strategies linked to measurable risk reduction.
Cancer is not one disease. A startup should identify a specific cancer type, patient group, clinical decision, or operational failure before describing a broad platform. “AI for cancer” is weak positioning; “a validated tool that helps Indian oncology centres identify treatment-resistant lung-cancer patients earlier” is testable and investable.
Where startup opportunities are strongest
Early detection and diagnosis
Late diagnosis is a major challenge in India, where access to specialists, pathology, imaging, and follow-up varies sharply by geography and income. Potential products include decision-support software, pathology assistance, risk stratification, and low-cost screening workflows.
The evidence bar is high. A model that performs well on a curated dataset may fail across hospitals because of differences in scanners, staining, protocols, language, or patient demographics. Founders should plan for external validation, prospective studies, calibration, false-positive management, and a clear clinical workflow.
Treatment discovery and precision oncology
Drug development, biomarker discovery, combination therapy, and patient stratification can create substantial value. Computational methods may reduce the cost of identifying candidates, but a prediction is not a therapy. Teams need access to wet-lab capabilities, domain experts, reproducible datasets, and a route from computational finding to biological validation.
Precision oncology products also need to explain what action follows a result. If a test cannot change treatment, trial eligibility, monitoring, or counselling, its clinical utility may be limited.
Clinical operations and patient support
Some of the most practical opportunities are operational: coordinating referrals, reducing missed appointments, summarising records, supporting trial recruitment, tracking symptoms, and helping patients navigate complex treatment plans. These products can reach the market sooner than novel therapeutics, but they still handle sensitive health information and must integrate with real clinical workflows.
For founders building AI-heavy products, an initial rapid AI prototyping approach can test workflow fit before the team invests in a large platform. Prototype quickly, but treat the prototype as a discovery tool—not clinical evidence.
What a credible application should demonstrate
A strong YC application should answer five questions clearly:
1. What specific problem are you solving? Name the patient, clinician, payer, or laboratory and quantify the cost of the problem.
2. Why now? Explain the enabling science, data availability, infrastructure, or regulatory change.
3. Why is your team suited to solve it? Clinical access, research expertise, engineering capability, and lived experience can all matter—but they should be concrete.
4. What evidence exists? Include experiments, retrospective results, customer interviews, pilots, publications, patents, or early revenue, while distinguishing evidence from ambition.
5. How will this become a business? Identify the buyer, purchasing process, pricing logic, and path to repeatable deployment.
Do not inflate preliminary findings. State the dataset size, comparator, endpoint, limitations, and next experiment. In healthcare, intellectual honesty is a competitive advantage because sophisticated reviewers will look for hidden leakage, weak controls, and unsupported clinical claims.
Regulatory and data requirements
A cancer product may fall under India’s medical-device rules, drug regulations, laboratory requirements, or a combination of them. Classification depends on the intended use and claims. Software that merely organises information is treated differently from software that makes diagnostic or treatment recommendations.
Before building, map:
- Intended use and clinical claims.
- Data ownership, consent, and permitted secondary use.
- Security controls, access logs, retention, and deletion.
- Validation design and performance thresholds.
- Human oversight and escalation procedures.
- Applicable Indian approvals, ethics review, and institutional permissions.
- Export, cross-border data, and international regulatory considerations.
India’s healthcare environment also demands attention to affordability and deployment. A solution requiring expensive hardware, uninterrupted connectivity, or highly specialised staff may struggle outside premium hospitals. Design for local constraints: multilingual interfaces, intermittent connectivity, varied data quality, and integration with existing systems.
If your product processes clinical notes or patient conversations, document how models are evaluated for privacy, hallucinations, and unsafe recommendations. General AI workflow automation for high-growth startups principles are useful, but oncology workflows need stronger review gates and auditability than ordinary business automation.
A practical 12-month build plan
A focused first year could look like this:
- Months 1–2: Interview oncologists, pathologists, patients, administrators, and payers. Select one narrow use case.
- Months 2–4: Secure lawful data access, define endpoints, establish a baseline, and build a limited prototype.
- Months 4–6: Run retrospective testing and failure analysis across more than one data source where possible.
- Months 6–9: Conduct a supervised pilot with explicit success metrics, human review, and incident reporting.
- Months 9–12: Convert pilot evidence into a regulatory, reimbursement, deployment, and fundraising plan.
For data-intensive products, a documented best tech stack for AI startups should prioritise reproducibility, access control, monitoring, and cost—not simply model performance. For multilingual patient-facing tools, India-specific language evaluation is essential; a multilingual chatbot design should include safe handoff to clinicians and avoid presenting general information as medical advice.
Funding and YC fit
YC funding terms and application windows can change, so applicants should verify current details on YC’s official website rather than rely on summaries of the Summer 2024 batch. More important than the headline cheque is whether the company can reach a decisive next milestone: validated performance, a clinical partner, regulatory clearance, a paying deployment, or a strong biological result.
Founders should separate grants, research funding, accelerator capital, and venture investment. A therapeutics company may need non-dilutive scientific funding and specialist investors well before commercial revenue. A clinical software company may reach revenue earlier but face lengthy procurement cycles.
The central test
The “A way to end cancer” request rewards ambition, but ambition must be paired with specificity. Choose a cancer problem where your team has unusual insight, define the measurable outcome, validate it in the right population, and build around the realities of Indian healthcare. The most credible companies will not claim to solve all of cancer. They will show a narrow intervention that works—and a believable path to scale its impact.