Incentives matter
Successive governments have relied on global consultancies that largely originate from America, and we've long committed to consuming their product rather than developing our own.
This was the prevailing wisdom from the economic miracle of globalisation - why waste your own resources, when someone else is doing it cheaper? It's become embedded in public service thinking, and reinforced by budget constraints. High capex, investing in infrastructure and the delays of implementing our own tools look like a bad choice when we can pay a small monthly fee to a global supplier. OK, so the small monthly fee always increases, and the lock in is real, and we end up paying for an entire product category when individual tools would do the job - but, still, there's a nice guy in a shiny suit ready to smile at any civil servant within 100 miles.
Private organisations like Nutanix are already looking at the economics of AI and coming to the conclusion that robust independence is key to adoption - open weight models, on prem services rather than reliance on the new shiny from corporate giants that might go pop at any minute. This isn't about committing to the bleeding edge, but about pedantic adoption and ensuring that actual value is delivered rather than just PR friendly high metrics.
Politically though, this government (and any of the opposition) would not dare walk away from 'our friends in America' - and certainly don't have enough spine to weather the admonishments they'd get if they dared to 'go rogue' with open weight models or independent development. We'll lose this race, because politicians are more concerned with being seen to wear the sponsors shoes than championing the plucky underdog.
