I think you misread the situation. The problem is that the current meager level of AMD success doesn't nearly ensure that healthy competition at all, not to mention ensuring it stays here in long term, with all those desirable effects that brings.
Look at the market: after year of riding Ryzen's improved competitiveness and mindshare, AMD has hardly any OEM sales, and their overall *DESKTOP* share (not to mention mobile/servers) has moved from 10 to paltry 12 %. And they had to cut prices on all the CPUs by 20-33% to achieve that. They are still just barely profitable (and that profit is tiny, not nearly enough to make any significant investments or debt reductions). And that is also thanks to the otherwise terrible crypto bubble (that better died swiftly...). Intel still eats all the profits and has 10x the revenue and what, 50x the profits? You only need to look at Intel being able to merrily sell its HEDT CPU for $2000 (2x more than two generations prior) even though there's Threadripper, which needs to be discounted heavily to keep selling, to see that the starting positions in the fight are still way too far from being equal.
As one of the other poster wrote above, AMD needs to get much stronger if the competition in the CPU market is to be ensured long-term. Guessing wildly, I'd say it needs at least double the revenue it will have in 2018 (6.5-7.0 billion), at least double the RD budget (800 millions/quarter, while it was ~400 in 2017), and then it needs to be bringing in about 1.5+ billions of yearly net profit and keep improving there, to keep the investors happy and the management from having to cut RD. They need to pay their debt on top of that. And they need to be able to field a new chip design every year instead of bi-annually (pinnacle ridge is the same design as summit ridge), if they are to be able to compete with Intel once it fixes its 10nm manufacturing. They also need to be able to design and pay mask costs for more chip variants, so that they can cover lowend/budget market under $100 - they are unable to do that now, you can clearly see how they are so cash strapped that they just can't afford to design chips for essential market segment. If not for anything else, this on its own tells you they are not in competitive enough position, because their size just doesn't allow them to field all the chip categories that the market needs. Result: some segments see no competition and the overall ability to keep Intel in check suffers.
So I am afraid that we really need Intel do tank for 2-3 more years so that AMD's position and healthy improves enough to make it really dangerous. Because now, it is not, it is still highly endangered instead. And that is without mentioning that it needs money and ability to press/market itself into big OEM and retail sales in the GPU segment too.
Even if Intel struggled for idunno, 5 years?, its position would hardly be endangered, you have to realize that. They wouldn't even lose that much of marketshare thanks to all those OEM deals and sheer economic strength. I'm not even convinces those 5 years of AMD lead would ensure the "long-term sustained competition" goal.