There are lots of flaws in his argument.
For starters, from your account, he appears to be assuming that 90% of the potential market for illegal drugs stays away from them because they are illegal. We have good reasons to assume that is flat out wrong, by looking at Portugal (where use appears to have dropped after decriminalisation) for example.
Secondly, there's reason to assume that a substantial proportion of deaths related to illegal drugs (and lets just take the numbers you cite at face value) are a direct result of criminalisation. Heroin, for example, is a very safe drug when it is pure, in predictable doses and safely administered. It becomes massively more dangerous when it is impure - mixed with who knows what (brick dust being one common example..), comes in unpredictable doses with according risk of overdoses, and is administered with potentially unclean syringes, in unsanitary conditions (clean syringes, or low enough prices for people to afford sufficient doses for oral use would both make a huge difference).
There is every reason to assume that number of deaths per 100,000 users for illegal drugs as a whole would plunge if well regulated legal sources were available, combined with less judgemental health care options for actual addicts. We have clear evidence of this from heroin treatment, where programs to provide clean syringes and safe places to shoot up on its own provides drastic harm reduction, even when addicts are still dealing with the dangers of massively impure drugs.
Moreover, many of these drugs are only used as more easily obtainable (cheaper, often) alternatives to safer, more harshly regulated illegal drugs. E.g. shooting up heroin is often the "low cost" alternative to obtaining prescription opiates that are safe enough for routine prescriptions on the black market.
As such, legalising the safest drugs - including "hard" ones like heroin that are quite safe in pure forms - and treating addiction to prescription drugs like oxycodone as a health problem rather than cutting people off and pushing them to illegal alternatives, may be assumed to reduce the number of deaths from some of the genuinely highly dangerous illegal drugs or mixes by shifting use away from the worse alternatives.
Another reason it doesn't make sense to extrapolate like that is that even if the number of deaths from now-illegal drugs would go up 10x, what matters would be the increased total harm, and for that to go up 10x there would need to be no replacement effect. That is, if 50% of those extra deaths are people who would otherwise be abusing, and dying from, alcohol, then the actual increase in mortality would be equivalent to 5x as many deaths as currently.
But it's even more complicated than that: Some of the illegal drugs are demonstrably substantially safer than drugs like alcohol. Depending on the size of replacement effects, and the effect of legalization on relative popularity of different drugs etc., actual effect on overall harm could be vastly different.
This before you even start looking at things like crime reduction.
Consider that when looking at crime reduction, it is not just theft to finance purchase, but also outright war-like conditions in many of the places the drugs are produced. As in, army units being used to try to contain manufacture, and failing, with associated huge number of deaths.