I need to tighten one premise from my previous post on senior engineers and capacity selling.
The public debate around Sweden’s new work-permit salary threshold has been much harder to follow than it should have been. Some of the language made it sound as if "engineers" or "tech" broadly were being exempted from the new salary requirement. That is not what the official material appears to say.
The broad software-development category -- IT architects, system developers and test leads -- was the largest named work-permit category in Sweden in 2025, with 2,181 granted permits out of 11,100 total granted work permits for workers outside the EU/EEA. It was larger than the named engineering categories in the same table. A separate IT operations/support/network category had 254 permits.
That largest software-related category is not listed as an occupational exemption under the new rules.
The IT occupations that are listed are much narrower: IT operations technicians, IT support technicians, system administrators, and network/system technicians. There are also separate exemption routes for certain former students and for employees at certain young tech or life-science companies. But those are applicant- or employer-based routes. They are not the same as exempting the broad software-development occupation.
That changes the policy premise.
I had framed part of the pressure on Swedish software engineers as if the broad SWE/system-development category was being placed into a large low-threshold exception. That appears too broad. The official material does not support the idea that ordinary software-development roles, in ordinary consultancies, incumbents, or large product companies, are broadly exempted as an occupation.
The Communication Failure
This should not have required reconstruction.
The official information exists, but it is spread across government communication, Migrationsverket’s implementation page, press conference material, statistics pages, and political commentary. In an already inflamed debate, that is a bad communication pattern.
One exchange on X captures the problem. Ludvig Aspling was asked whether the top occupations in the 2025 permit statistics were now under the new 90 percent salary requirement. His answer was: "Ja och nej." That may be technically defensible. It is also exactly the kind of answer that signals a communication failure.
The rule has multiple layers: a default 90 percent median-wage threshold, occupational exemptions at 75 percent, separate applicant groups, separate startup rules, excluded occupations, and statistics that use occupational groupings that do not map cleanly onto political shorthand such as "engineers" or international tech terms such as "software engineer."
But that is precisely why the communication had to be unusually clear.
Instead, the debate became predictable noise. "Engineers," "tech," "IT," "shortage occupations," "work permits," and "salary threshold" were used as if they referred to the same thing. They do not.
A country can have many communicators and still fail to communicate if nobody owns the explanation end-to-end. This looks like one of those cases. The fact that it took AI-assisted research across several official and semi-official sources to reconstruct the actual structure is itself evidence that the public explanation was inadequate.
The Policy Picture Looks Better Than I First Thought
After looking closer, my view of the new rules is more positive.
The largest software-related work-migration category appears to move under the stricter 90 percent median-wage requirement, unless an applicant qualifies through another route such as the student or startup exemption. That could reduce the ability to use low-wage global sourcing for ordinary software-development roles in Sweden.
That matters for the local market.
Swedish consultancy firms are already having a hard ride. Many resource-consulting firms are struggling to keep people off the bench. They do not need additional low-wage global competition in the same ordinary delivery market. If the new rule reduces that pressure for normal SWE/system-development roles, that is a positive development.
It may also be positive for juniors and mid-level developers in Sweden. If employers cannot as easily fill ordinary software roles below the new threshold, then the local market may get some breathing room. I would not overstate this without salary-distribution data for actual work-permit applicants, but directionally it looks better than the noisy debate suggested.
The strongest caveat is data.
Migrationsverket publishes permits by occupation and citizenship. I have not found a public salary distribution showing how many 2025 permits in the IT architects, system developers and test leads category were below or above 90 percent of Swedish median salary. Without that, we cannot quantify how many actual permits would have failed under the new rule.
So the honest conclusion is narrower:
The largest named software-related work-permit category is not occupationally exempted. The 11,100 figure is not a software-exemption figure. The 2,181 figure is the actual 2025 software-development-related permit figure. Some software workers may still qualify through student or startup routes, but that is a different mechanism.
That is a more favorable policy picture than I first assumed.
The Capacity Argument Still Stands
This correction does not change my underlying view on senior engineers selling capacity.
The strategic problem with capacity hiring is not dependent on one Swedish exemption list. The problem is that capacity markets are structurally bad places for senior judgment.
If the buyer asks for "a developer per hour," price becomes the comparison surface. Procurement systems, global sourcing, staffing intermediaries, and now AI all push that layer downward. The buyer is not really buying architectural judgment, integration responsibility, operational accountability, or problem ownership. The buyer is buying a unit of delivery capacity.
That remains a bad market for senior software engineers.
The better direction is still higher-value work: specialist consulting, integration-heavy delivery, product-adjacent services, FDE-like models, and responsibility for making software work inside the customer’s real environment.
But the policy correction matters.
The new rules may make the near future somewhat brighter for ordinary Swedish resource-consulting firms than I first suggested. If the largest software-related work-migration category is held to the 90 percent median-wage threshold, then the local market is not being opened to a broad low-threshold SWE exception. That reduces one source of pressure.
It does not make resource consulting a great long-term model for senior engineers. It does not remove the pressure from AI, procurement, and global sourcing. It does not change the argument that senior engineers should avoid being sold as interchangeable capacity.
But it does mean the migration-policy premise should be corrected.
The better reading is this: the government’s new rules appear to protect ordinary SWE/system-development roles more than the public noise suggested, while still preserving narrower exemptions for specific shortage occupations and for certain startup/student cases.
That is a better outcome than I first thought.
And it is also a reminder: when the public explanation of a policy is this fragmented, the debate will fill the gaps with heat instead of understanding.
Source notes
Official Migrationsverket page on the new work-permit rules and exempted occupations:
https://www.migrationsverket.se/nyhetsarkiv/nyhetsarkiv/2026-05-22-nya-regler-for-arbetskraftsinvandring---dessa-yrken-undantas.html
Migrationsverket 2025 work-permit statistics by occupation:
https://www.migrationsverket.se/Om-Migrationsverket/Statistik/Arbete.html
Government press release on the exemptions:
https://www.regeringen.se/pressmeddelanden/2026/05/regeringen-presenterar-undantag-fran-det-nya-lonekravet/
Migrationsverket page on the 90 percent salary requirement from 1 June 2026:
https://www.migrationsverket.se/nyheter/nyhetsarkiv/2026-04-17-nya-regler-for-arbetstillstand-fran-1-juni-2026.html
X exchange referenced in the text:
https://x.com/AsplingLudvig/status/2058612334583501286