It is not only LAMP and .NET, but other technologies like Phyton and Ruby that are being widely used. But I don't see it as getting the market share of other frameworks but getting specialized in one aspect. LAMP is lightweight and fast but still not a great idea for transactional applications. Phyton is ideal for information retrieval applications and Ruby is very maintainable. .NET is not lightweight nor yet trusted for transactional applications and suffers from scalability but if scalability is not your concern, it is great because of the ease of development. And Java is still the best candidate for enterprise and transactional applications like financial applications.
It is interesting that in the early years of SE, the programming languages like Cobol and Fortran were very specialized for certain usages. The idea of general programming language gave birth to Pascal, C and later, JAVA. But now again they are becoming specialized but not at syntax level, but at the level of solution stacks and frameworks.

This is not a good question, both of (Java and C#) and (Java EE and .NET)
have some strengths and weakness when compared to each other.
In some areas Java and Jave EE may win and in some C# and .NET.
I think that LAMP (Linux, Apache, MySQL, PHP) has nothing to say for future.
LAMP is good for only personal or very small web apps.

I think since Java object oriented language, and you are not limited only to one architecture in java, lots of 3'rd party tools and open source components are available for it, lots of them component based, it is much more easier to develop and maintain large scale applications in Java while it is not the same in PHP (and LAMP)

now with component based web frameworks like JSF and JavaFX, frameworks like Seam, EJB 3 component model, tools like Netbeans (with webpack) and JDeveloper, and the new introduced GPL license for Java, I think developing small applications on PHP is not much easier than developing Java applications.
also the learning curve can be shorten a lot using good tools.

so an open source Java Stack on Linux can not be treated by PHP in any way any more.
but as some people love Microsoft platform, there are some people who love PHP and will still use it.

.NET and C# will never be available on platforms other than windows seriously, I tried Mono port of .Net to linux and found out lots of packages specially those related enterprise development (COM+, DCOM) are missing those components are based windows core API and if some one wants to port them he needs to rewrite lots of windows functionality.
the other side is that except an experimental port of .net compiler to BSD rest of the works done are not official work of Microsoft and are not update with latest MS .net platform.
so .Net and C# are not also a threating technology for Java yet! also Microsoft is big company with lots of talented people and they have 40% of server side market and if they decide to go for the rest, they can do lots of great things.

Other have already given their comment and I agree with them. but to add, as much as Java replaced C and C++, others will replace Java.

There is no doubt that .NET is very successful in presentation tier and web components, and Java ee still has not pull itself together for presentation tier well, but the level of integration that exist in competitor frameworks (Java ee) is quite high as well. in today's industries the main intention of such frameworks (.net and Java ee) is to provide the best way to integrate and interoperate with other systems (either legacy or newly developed). neither are talking about replacing. Even when Microsoft goes to conferences (they are called architect / developer evangelists) to talk people into their new features their intention is never to convince people to replace their products because they already know it does not make sense.

Java has been evolved so much that it seems to be hard and unreasonable to leverage other platforms functionality to reach to this level. so far it seems to be waste of time and money to do so. Also as Arash has pointed out, some platforms e.g. Windows and Unix family have already shown which frameworks are more suitable.

After all these are all architecture decisions but from this perspective, it seems difficult to replace such competitors.

You state that, ".NET and C# will never be available on platforms other than windows seriously". Do you have any offical statistics about this opinion or you say just as your own experiences? Because Microsoft annouced that C# and .NET work in non-Windows platforms and Yager pulished in JavaWorld 2 Oct. 2000, as well.

I have try all available solutions my self to run one of my .net applications to other operating system for one of my customers.

there is one C# compiler and CLR plaform for BSD from Microsoft. they don't have any support for Linux.
it is completely an academic solution and only an small portion of .Net APIs is implemented.

then there is Mono from Novell
of course mono is far behind the MS implementation of .Net

the other obstacle I reached was some .Net components like "Enterprise services" were based on some services of windows operating system in this case COM+
so any port of .Net platform  should include a port of large portion of windows operating system!

the problem with .Net portability is that is designed to take advantage of other Microsoft technologies.

I agree with you,
the best platform  is the one that reduces the Total Cost of Owner ship of system.
and not every tool is good in every situation, but some times some tools target same market,
for example Foxpro and MS Access or OMG and DCOM or now JavaEE and .Net platforms.

As i know, the thing that is coming behind Microsoft's
announcement is sometimes much pompous and
opinionated. Microsoft has announced that CLR is Open
Source and is available to all third-party industry.
As we refer to the past, Microsoft's policy is
strictly tight with its support of existing customers
and businesses to enable well trading!. I mean holding
support of COM, DCOM and COM+ is a strategic policy
for Microsoft since its huge society of Customer
Businesses and Organizations needs that to stay
alive(Thanks for Mr. Rajaeeian has mentioned about
it). You can see such non academic developments in
other areas, like Virtual Machines when Microsoft's
operating systems switched from MS-DOS to GUI Windows
OS. Finally, Microsoft's notice on availability of CLR
as Open Source, has targeted third-party industry to
expand their support of CLR as what has been done to
JRE. How much Microsoft would be successed in real, we
should be waiting. Howevere Microsoft will accept any
offer in this area with shouts of welcome, indeed.