Abstract This paper focuses on ground investigation by deepwater cone penetration tests (CPT). All current offshore CPT systems deploy subtraction-type cone penetrometers. Their use now includes record-breaking deepwater applications. This paper presents results of geotechnical field and laboratory measurements that explore the operational limits of deepwater cone penetration tests in very soft soils. The review of field repeatability studies shows the feasibility of excellent performance of the subtraction penetrometer. The results of specific laboratory experiments indicate a relatively small influence of deepwater ambient pressures on its performance. A further consideration for deepwater applications is the zeroc-orrection to hydrostatic conditions prior to the start of a test. Seabed-based CPT systems allow zero-correction with an uncertainty approaching the resolution of the CPT system. Downhole CPT systems latch into the lower end of a drill pipe. The pressure conditions in the drill pipe may not be in full equilibrium with the surrounding ground water pressure and zero-correction will be subject to increased uncertainty. The main points are:currently available evidence indicates that a high-quality subtraction-type cone penetrometer is adequate for very soft soil characterisation to a water depth of 3000 metres and probably beyond,Class 1 accuracy results (according to ISSMGE) in very soft soil are likely to be feasible for deepwater seabed systems anddeepwater downhole systems are less accurate and will probably give Class 2 results. Introduction The drive for resource development of deepwater regions provides great challenges to the engineering profession. This also applies to deepwater geotechnical investigation practice. Operational limits, very soft soils and integration of the disciplines of oceanography, geology, geophysics and ground investigation are some of the specific challenges (Kolk and Campbell, 1997). The cone penetration test (CPT) has become the most widely used in-situ testing technique for offshore geotechnical investigations. Detection capabilities for distinct and gradual changes in stratigraphy are generally excellent. There now exists a comprehensive theoretical and empirical database world-wide (Lunne et al., 1997) that provides a high level of confidence in correlations between CPT cone resistance and laboratory measurements of undrained shear strength and other geotechnical parameters. Recent results of large-strain numerical modeling of cone penetration testing in undrained soil implicitly support the common empirical correlations (Fugro, 1998). This observation implies that the CPT has the ability to achieve a high level of accuracy, where accuracy is defined as "the closeness of the agreement between the result of a measurement and the true value of the measurand" (ISO, 1992). The use of the CPT in a deepwater environment is significant, as the role of CPTs in providing parameter values for analytical design models becomes increasingly important. Conventionally, laboratory testing provides the reference strengths. Laboratory testing of very soft soils requires highquality samples. This is an important consideration for deepwater sites, as the effects of sample disturbance increase with an increase in pressure release upon sampling. This is leading to a greater reliance on the results of in-situ tests.