Or you can have a Starship land on Mars, have a human stroll out of it, collect the samples, build an ISRU plant to generate fuel for return and fly back with the samples.
Both are audacious plans, but somehow SpaceX’s seems the less audacious.
Constructing a spacecraft capable of sustaining a human being for multiple years, landing them with all necessary support equipment, building an ISRU plant for which solid designs do not currently exist, and for which even the principles are still very much the subject of research, launching and returning is much, much more complicated than sending an unmanned spacecraft or two to recover a sample. This is really not too far off from what JAXA is already doing with Hayabusa 2.
Multiple rounds of aerobraking and landings in deep gravity wells is vastly different than Hayabusa 2.
And Mars is eminently survivable for humans, far better than the moon for example. Starship makes landing a few hundred tons of food, supplies and equipment before the first crews relatively easy.
ISRU experiments on earth have proven the concepts already. We can figure out ISRU details along the way, even if it takes a few extra synods and delivering more equipment. we will be getting far more science out of a human landing than robots can ever provide.
In fact, the first humans on Mars will explore more of it in a few weeks than 50 years of rovers have done. The first human to take a single shovel full of Mars will have have dug deeper than than the Insight orbit has done in a year of trying.
BTW: The budget for this robotic Mars dirt retrieval mission is $7B. Based on NASAs track record on large projects, it’s unlikely to be less than double that.
Elon Musk has estimated the cost of landing the first round of humans on Maes (probably over a hundred) at $10B total. Now Elon is a huge optimist on schedules, but on budgets he’s been pretty accurate.
A manned spacecraft re-landing on the moon is the natural prerequisite to a manned mission to mars. First the lift capacity of terrestrial objects into space must be improved to reduce costs.
NASA has many big, inefficient and impractibly expensive plans. What they typically lack is a sound gameplan for achieving them.
Landing on the moon would be an expensive diversion to any Mars missions. The moon is an entirely different, and far more hostile environment, you need entirely different suits and life support systems on the moon.
Mars requires lifting far less fuel to land on and return from its surface instead of the moon. Starship is optimized for Mars landings, and will be capable of landing up to 150 tons of supplies, equipment, and humans in a single ship.
It can land far less on the moon in the sane configuration.
First, Mars round trips typically take 2 years, not ten. Flying to Mars can be done in under two months.
Secondly, a fully fueled SpaceX starship in low earth orbit will likely land 100-150 tons on Mars. It won’t be able to land any payload on the moon.
Lastly, you don’t need to take fuel for the return trip to a Mars. It’s easy to make there.
Fuel will be virtually impossible to make on the moon for decades, requiring every drop to be shipped from Earth. Because of the tyranny of the rocket equation, that requires a massive amount of fuel to land a much smaller amount of fuel on the moon to be able to fly back a tiny payload to Earth.
The cost of food, water and fuel on the moon is probably 10x that on Mars. The moon is also a desert, full of razor sharp dust and nearly devoid of resources, as well as commercial and scientific value.
Probes travel to mars faster. They have lower mass to fuel ratios.
A manned flight with water shielding for radiation. Rotating crew quarters to generate artificial gravity. And other amenities necessary to support human life, will have far higher mass than a probe. It translates to travel time taking longer due to fuel consumption and storage being more limited.
Both are audacious plans, but somehow SpaceX’s seems the less audacious.