Background: Surgery for cancer is often delayed due to variety of patient-, provider-, and health system-related factors. However, impact of delayed surgery is not clear, and may vary among cancer types. We aimed to determine the impact of the delay from cancer diagnosis to potentially curative surgery on survival. Methods: Cohort study based on representative sample of patients (n = 7,529) with colorectal, breast, lung and thyroid cancer with local or regional disease who underwent potentially curative surgery as their first therapeutic modality within 1 year of cancer diagnosis. They were diagnosed in 2006 and followed for mortality until April 2011, a median follow-up of 4.7 years. Results: For colorectal and breast cancers, the adjusted hazard ratios (95 % confidence intervals) for all-cause mortality comparing a surgical delay beyond 12 weeks to performing surgery within weeks 1-4 after diagnosis were 2.65 (1.50-4.70) and 1.91 (1.06-3.49), respectively. No clear pattern of increased risk was observed with delays between 4 and 12 weeks, or for any delay in lung and thyroid cancers. Concordance between the area of the patient's residence and the hospital performing surgery, and the patient's income status were associated with delayed surgery. Conclusions: Delays to curative surgery beyond 12 weeks were associated with increased mortality in colorectal and breast cancers, suggesting that health provision services should be organized to avoid unnecessary treatment delays. Health care systems should also aim to reduce socioeconomic and geographic disparities and to guarantee equitable access to high quality cancer care.