End of an Era: How Silicon Will Decide BIM’s Future

[PB]

A must read for the future of BIM software and, most pertinently, ArchiCAD.

Can Archicad be re-written for the multi-processor AI integrated future, or are we users all just treading water until a new player answers the challenge?

https://architosh.com/2025/12/end-of-an-era-how-silicon-will-decide-bims-future/

[jl_lt]

they are receiving suscription money now, so they should feel motivated to do it... or was it the other way around?

[Bricklyne Clarence]

We can't even get dual/multi-monitor capability in archicad in 2025 (....because......"reasons".......er..... I mean,........because,  "not technically possible at the current time",...therefore "ON HOLD"......).....and you think that this company is positioned to take advantage of coming advances in computer chip technology?

Or to position archicad in the future to do so?

Boy, are you the super-Optimist.

Besides which, if they do do this, then how will they have any resources left over to develop their long catalog of the next ten to 100 items and features that nobody ever asked for, nor needed, nor which ever appeared in the Wishlist request sections,....or the RoadMap (...until they belatedly add them in after-the-fact)?

[torben_wadlinger]

I’ve read that same article. The main point is the speed race is coming to an end because of physics. Intel couldn’t fulfill its promise for speed and efficiency so Apple was forced to develop its own processor. But the speed gains are beginning to drop either. The only way into the future is to rewrite the software to make use of multiple processors. 

This affects the main core of every CAD/BIM-Software at the market and opens a door for new competitors. For that reason it is the perfect moment to stop paying for archicad and wait what the future holds.

Five years ago I wrote on this forum that it would be better to re-develop archicad from the ground up just like Apple did with Final Cut. Now, five years later Archicad is a mess: it’s slow, has all the old bugs and new ones on top. The wishlist is growing and growing but the implementation of new functionality is mainly driven by marketing and not by user needs.

From that point of view the next five years will be interesting!

Happy new year 2026!

[Bricklyne Clarence]

---

@torben_wadlinger wrote:

I’ve read that same article. The main point is the speed race is coming to an end because of physics. Intel couldn’t fulfill its promise for speed and efficiency so Apple was forced to develop its own processor. But the speed gains are beginning to drop either. The only way into the future is to rewrite the software to make use of multiple processors. 

This affects the main core of every CAD/BIM-Software at the market and opens a door for new competitors. For that reason it is the perfect moment to stop paying for archicad and wait what the future holds.

Five years ago I wrote on this forum that it would be better to re-develop archicad from the ground up just like Apple did with Final Cut. Now, five years later Archicad is a mess: it’s slow, has all the old bugs and new ones on top. The wishlist is growing and growing but the implementation of new functionality is mainly driven by marketing and not by user needs.

From that point of view the next five years will be interesting!

Happy new year 2026!

---

Many people have made this point going back more than a decade now,....into the early 2010's,....and probably even earlier.

It's seems pretty obvious now that a large reason for the bottleneck to developing and implementing new features an tools - particularly tools and functionality that are increasingly getting requested by users - is the fact that they're trying to build everything now on outdated legacy code from the 1990's (and possibly even as far back as the 1980's).

Everything from the ability to have simultaneous and instaneous update of all open window views with any chang in any other view; the ability to have multiple floor plans open; the lack of a graphics/visual base custom parametric object creator and this mind-boggling stubborn refusal on their part to stop forcing people to have to learn how to code in GDL just to get what we need; even hardware stuff like lack of and inability for multi-monitor capability; lack of full exploitation of the powerful GPU graphics cards we have today - and not just for realtime rendering;....I could go on and on.....but the point is that most of those things I just listed and laid out are already standard features in ALL their competitor and rival software's features and toolsets.

Every. Single. One.

And they're still stuck in sometime between 1998 and 2012, in terms of how much they can stretch this software to do more.

A complete code re-write has been clamored for for as long as I can remember and they've resisted it, for only what I can imagine they would claim is limited resources and capability.

And yet with every new version you can see it all starting to come apart at the seams as they try to cram more into an old code that's not optimized to handle it - with the increased instability, the unresolved bugs and problems dating back at least 10-15 versions that have never been fully fixed, alongside new ones - all packaged in nice neat package of "new" tools that are not even fully or properly functional.

On the Windows side, Microsoft officially ceased all support for WIndows 10 this last fall - which is usually a sign that the next version of Windows (version 12) is probably around the corner in the next several years. 

With a new OS platform version they'll be new capabilities for software developers to take advantage of new features for their respective programs, and once  more Graphisoft along with, Archicad (certainly on the Windows side) will get left (even) farther behind.

And forget the Wishlist. Certainly they seem to have.

They've all but openly announced at this point that they don't really care what happens in that Wishlist section as they go about doing what THEY want that users aren't asking for.

Meanwhile the things we ask for, and that they tell us to vote on to help guide their development process to address our needs - the ones that get the most votes are increasingly falling under the "ON HOLD" banner. Meaning, they're never going to get around to them, nor is it in their current or non-immediate plans to ever fix them.

And in many of those cases it's because they legitimately can't overcome the technological limitations (read: legacy code bottlenecks and roadblocks) that they need to do in order to fix or implement them.

If they had taken your advice 5 years ago, or even that of another user suggesting the same 10 years earlier, and started a shadow project in the background aimed at upgrading Archicad's core code to a modernized version that takes advantage of modern OS features and capabilities while implementing new ones, we'd all be way ahead know - even in just knowing that a revamped Archicad is likely around the corner in the next several years.

But we know that's not happening.

Not now, nor anytime soon in the future.

I think most people are just tired of the antics of (both) the major players right now and are just waiting it out to see if another new upstart comes into the scene that can take advantage of all the disgruntlement that most of us are feeling and provide something truly worth adopting before they jump ship.

[PB]

Indeed, my take-away from the article is that all current BIM software needs to be rewritten from scratch with a thoroughly modern modelling kernel (nurbs, full multi-processor working throughout, multiple monitors, etc.).

As mentioned, it is also the ideal time to dig out the old (very extensive) wishlist and deal with all the core issues that have infuriated users for many years. No need to pay any attention to the recent 'Roadmap' which is devoid of ambition.

I suspect that, in reality, the task is too daunting for Graphisoft: If the kernel is re-written, then it is also the moment to re-invent the Tools, attributes, work environment, and much more to fully address the future...and that is before even getting on to the subject of AI integration !

If Nemetschek were to undertake such a venture (inevitable at some point?), then perhaps the logic is to consolidate their various software offerings into one coherent core programme...

On the subject of AI, rather than the 'cute' current integration of an AI facsimile, AI should be integrated to perform all the mundane time-consuming tasks that would liberate architects & designers for more constructive tasks (AI could use the tools to build as-existing models from point clouds, research & model site context, research & check for regulatory compliance, real-time building performance calculations, real-time cost estimation, Nano Banana quality rendering, integrate manufacturer data directly into tools - windows, doors, composites, etc. - and so much more).

[PB]

With further thought, I strongly suspect that all the big players are simply waiting for a start-up to make the breakthrough so that they can simply buy them up. The others will likely then reverse engineer and develop their own version.

[JeffH]

It's was always anticipated that no big changes would happen until or after version 30, to attempt to limit the number of people that stay perpetual and remain version locked.

It should be of no surprise that AC28 and AC29 have been so anticlimactic.

If AC30 is another dud.....the userbase will look significantly different and smaller.