I haven't spent much time modelling VIEs as I don't think it's a significant issue, but my understanding broadly is that c.50-75% of cash raised from VIE partners is repaid in year one with tax credit proceeds. The rest is repaid out of solar lease cash flows over 5-30 years, but should only be 20-30% of lease cash flows maximum in any year. So with solar lease cash flows at around $400m, this should only be around $100m obligation. Additionally for the redeemable interests ($551m), i presume under some circumstances and after a certain time the partners can ask Tesla to buy out their stake early.
Tesla has $6.3bn net solar lease asset book value on balance sheet, of which $5.1bn is in VIEs. Contracted cash flows over 30 years is likely closer to $10bn and then Tesla will also expect significant continued cash flows after end of lease. NPV of this varies significantly with your discount rate assumptions and estimates for post lease cash flows.
Of this $6.3bn book value, $1.3bn is the equity share owned by third parties. Against the $6.3bn book value Tesla has $2.2bn of non recourse debt.
So while VIEs could be a drain of cash, particularly the redeemable ones, Tesla still has c.$1.2bn of solar assets outside of VIEs and continued new solar sales, both of which it can continue to raise money from. Tesla also still has a significant equity value in its solar portfolio (in the $bns) and its solar portfolio could be a significant source of cash if Tesla chose to sell the entire portfolio (which would also remove the $1.3bn VIEs liabilities and $2.2bn non recourse debt). I think Tesla may do this once it has cross-sold Powerwalls to a significant % of its roof portfolio customers.
Many of Cunningham's numbers and calculations are obviously wrong, but I haven't spent time to fully understand Tesla's numbers.