Re: Sourceware budget and funding 2026/2027, Fri 12, 15:00 UTC

[Mark Wielaard <mark@klomp.org>]

Hi all,

On Thu, Jun 11, 2026 at 12:38:35PM +0200, Mark Wielaard wrote:
> Based on last months budget discussions and last years financial
> overview the Sourceware Project Leadership Committee wrote up an
> ambitious budget and funding proposal for 2026/2027, see below. Lets
> discuss if this matches community expectations and what the most
> effective way is to drive this fundraising campaign.

For those that couldn't attend here is a summary.

It looked liked emails with lots of dollar signs and asking for money
go into spam filters, so not everybody had seen the Sourceware PLC
financials overview. It is also here now:
https://sourceware.org/financials.html

The PLC document is based on the financials from 2025/2026 and the
looking forward/looking backward discussions last month:
https://inbox.sourceware.org/20260516203206.GR1219@gnu.wildebeest.org/
https://inbox.sourceware.org/20260510223754.GR1219@gnu.wildebeest.org/

Switching from metal to people is a good funding goal. Especially
dodging RAM prices right now :)

There were some worries that there aren't enough "tasks" listed for
corporations to fund explicitly. But ideally sponsorships are not
targetted, they are general funding for the project (to make managing
the money better and easier for the project). But if there was
specific contract work that would be relevant for companies, we could
consider that separately.

We currently have items listed per funding "source" and only list the
tasks we believe we can do this year. But it would be good to list all
the work we want to get done (over multiple years). We do list more
under https://sourceware.org/sourceware-security-vision.html#plans
Modernizing account processes, Release upload process improvements,
Hardware keys, Security audit. It might be useful to just list them
all.

Question about spelling out the financial assumptions. For those
forgejo additions we assume $60/hour (total budget for a week work),
for upgrade/migration/automation tickets, we assume $80/hour (total
budget 3 to 6 days of work), for the junior sysadmin (at the end) we
assume $40/hour (total budget for 2.5 days for 6 to 12 months). Many
of the SFC contractors are around $40/hour, we have to do special
approvals if it's over $100/hour. People in north america/australia
are often about $75/hour. It is normal for people to charge much lower
rates for nonprofits.

We might want to consider corporate sponsorship "levels". the
Reproducible Builds team (who is another SFC project) uses
that. https://reproducible-builds.org/donate/#benefits-of-sponsorship
https://reproducible-builds.org/who/sponsors/
Then a corporation would fund for a specific "level" and not for a
specific task/contract.

In general people seem to agree you have to start somewhere and the
current estimates are OK, but on the low end. We can be even more
ambitious, ask for more than what we currently believe we can spend in
a year and see if some individuals, corporatations or grant maker
surprises us.

We currently have a healthy reserve. We could operate for one or two
years without any additional income. And for our minimal baseline we
just depend on individual sponsors and some corporate in-kind
(hosting) support.

What is next?

- Create a new sponsors page that lists incentives, levels and "bonus"
  targets.
- Link it all together with the financials and donate pages.
- Find Sourceware Ambassadors to drive the fundraising campaign.