What's good for Røkke isn't necessarily good for Norway
Stargate Norway is Norway's largest KI venture. Without further action, it will mainly serve Røkke and Altman.

Ki-generated illustration from Sora.
Main moments
Aker and Nscale will build a giant artificial intelligence data center in northern Norway just outside Narvik with OpenAI as a customer. It offers enormous opportunities for Norwegian KI, but without clear action from the Norwegian authorities, the venture will be a lost opportunity for Norway.
The dimensions of the project are enormous. Stargate Norway will have 100,000 GPUs! For comparison, the new supercomputer Olivia has 300 GPUs. The cost of the first phase is NOK 10 billion for 20 MW, but those building out the maximum capacity can cost multiply.
The Norwegian KI community is celebrating. The hope is that this will give Norwegian academic communities and companies access to important computing power to build better AI. But without active measures, there is a greater chance that the computing power will go to those with the greatest ability to pay.
Stargate Norway is primarily intended to ensure OpenAI stable and affordable access to computing power in a key export region, not stimulate growth in Norwegian KI environments. OpenAIS dream is to make KI agents to replace workers. Those agents need a lot of computing power.
A data centre in northern Norway is wise in this respect, with cheap electricity, renewable energy, a cold climate that cools the machines and government that puts profit and the US ahead of other considerations.
Selling computing power to OpenAI and other KI giants is also a nice business model for Aker and a good source of income for Norwegian authorities. But if Norwegian AI environments are not granted access, this will be just another example of us building infrastructure for other people's innovation.
Can Norway take action to get more out of the big investment, which will, after all, seize large amounts of Norwegian power? Here we should learn from oil policy. The goal must be that the investment leads to the building of competence in Norway, the training of our own models that can be used for socially critical purposes and the development of our own IT industry in areas where we have an advantage.
The companies promise to secure access to Norwegian KI environments and boasts significant spillover impacts, especially in the North. But as long as Stargate is operated on commercial terms, it is difficult to see how these law-names will be realised as long as Norwegian environments lack financial muscle.
A concrete idea for the Norwegian authorities is to establish a Cern for KI in Norway. Stargate provides us with the data center capacity that enables the training and further development of own models, but we need to build up the environments that have the expertise and resources to do so.
A Cern for KI should be open to both research communities and start-up companies and can sell access at subsidized prices to anyone who wants to locate in Norway all or part of the year.
Only in this way can Stargate truly bring values to Norway and not just Røkke and Altman.