Really? That's interesting, though not unheard of. I haven't seen their LP agreement of course.
There are actually two reasons why LPs typically don't want to agree to this. The first is that most funds focus on a specific stage (early, late etc); typically early stage funds want to invest through the entire cycle (so if firm X puts $5M into your A round they plan to put a total of, say, $20M through the liquidity event). But a later stage fund might want certain investment criteria (expansion, growth, whatever the partners come up with) and don't want it used for "familiar" deals that might not fit that thesis.
The main reason, the one I mentioned above, is propping deals up. the big driver for this is that the IRR numbers are completely made up based on judgement, which has to be the case as most of the fund will be tied up in illiquid investments. However the people making the decision is conflicted -- they are the fund GPs and of course want to look good to get new investors into their next funds. You shouldn't assume they are corrupt or malevolent: inherently they can't be dispassionate, otherwise they wouldn't be qualified to be supporting their portfolio companies through thick and thin. But this is why firms are unhappy about CALPERS releasing IRR numbers, or why Fidelity might legitimately (publicly, as they are required to) mark an investment down while the venture firm might just as legitimately consider it higher in value -- and different venture firms could even disagree.
So the poor incentive cross-fund investing provides is that a fund that is struggling, especially a zombie fund, might want to bridge some of its firms via an investment from a newer fund in the hopes that things might turn around. This incentive is even stronger during a negative macro event (i.e. a recession). An LP in the new fund doesn't want to see its good dollars following bad, it wants the money in new, promising investments.
Now if you are a big firm and can throw your weight around, or have some unicorns that everybody is clamoring to get into even at a nosebleed valuation, then you might be able to get the LPs to permit cross-fund investments. But you might not even want it (don't forget the GPs will likely be different in the different funds so their interests might not even align.
Ah, I think you misunderstand the structure. They have structured the raise as two interrelated related funds. The two funds presumably invest in the same set of portfolio companies. I assume they structured it this way as way of reducing fees and perhaps because some LPs were interested in different levels of exposure.
This is the same size as Andreessen Horowitz’s past two funds and, like each of those efforts,
includes a primary pool (which can do both early and late-stage deals), plus an overflow pool
for portfolio companies that require significantly more capital.
The breakdown this time is $1 billion for the main fund and $500 million for the parallel
fund―the latter of which only collects management fees once capital is committed.
If this parallel fund had been able to invest in earlier funds (remember each fund is a separate company with a different set of LPs and GPs) I am sure they would have mentioned it.
