SatCamp 2025 — Landsat Next and the Art of Resilient Thinking
Last week I got to give the keynote at the third annual SatCamp in Boulder, Colorado. I had an immeasurably lovely time at the event, and was particularly pleased with how many attendees were at SatCamp for the first time.
The topic for the conference was resilience, and I was asked to set the stage thematically for the short few days that we spent together. Without fully rehashing my talk here, I want to touch on a few things that I leaned on in setting the tone for the conference.
The Landsat Next RFI: A test case in adapting to new constraints
Resilience can be found in many modes and at many zoom levels - ecosystem, community, personal, planetary. And while my talk ultimately centered on industry resilience, I also touched on how our individual mindsets contribute to our ability to solve community or industry-based problems.
Using the recent Landsat Next RFI, which proposes that the private sector fulfill the requirements of the Landsat Next mission while the government becomes a customer of the data, I dove into a case study.
About a month ago, the government put out a request for information asking private industry to propose plans for achieving the Landsat Next project at a fraction of past mission costs. The reaction from many folks in the remote sensing and earth observation industry was immediate and emotional.
Landsat SHOULD be a fully government supported program. The data SHOULD remain free and open. The government SHOULD fund this critical infrastructure.
Beware of your “shoulds”
For those of you who know me from a product management lens may be familiar with my “rules of the road” for PMing, one of which is “beware of your shoulds”.
I learned this lesson in my early days at Mapbox. We were transitioning from a skunkworks R&D mode into a more fully formed product suite. Our investors were scrutinizing our spending, and the free flowing cash for experimentation was coming to a close. I remember the first time I got asked to rigorously justify the compute and data spends of my team, the head count we were keeping, and the overall existence of the product. I was surprised. I had been lulled into a cushy development zone where everyone loved what we were building, and the team was small enough that we all knew why we were rowing in a certain direction. I didn’t understand how our leadership team went from seemingly understanding why what we were doing was cool, interesting and worth the money to not being willing to pay for it. I found myself saying in my head, “What? I don’t understand… they SHOULD get it.”
That mindset didn't serve me, my team, or my goals. And as I matured through my career and various jobs, I started to understand, of course, that adapting to new constraints is the norm in any industry. Whether that’s going through a maturation phase at a company, there being a cashflow change, or there being new regulatory frameworks that affect your day-to-day. And I was going to have to figure out a better way to manage myself through these changes.
So, I came to understand that my “shoulds” lead to emotional escalation and paralysis. "It SHOULD work this way" had become a way to try to bend reality to my will and to avoid dealing with the job in front of me. And in our current political moment - where POTUS just told the UN that climate change is a "con job" while Indonesia's president described building sea walls to protect Jakarta - we can't afford to stay stuck in our “shoulds”.
From "shoulds" to goal-setting and gap identification
I want to be very clear that in asking folks to check their “shoulds” I’m not in any way saying abandon your aspirations, your goals, your wants, your drive. Aspirations are a beautiful and necessary thing. But instead of saying “should”, say, “I’d like”. And start to study the gap between what is now and what you’d like to be.
This reframing may seem minor, but it allows you to pause, assess the situation, identify the gaps between where you are now and where you’d like to be, and plan out the actionable steps you need to take to get from here to there.
This may seem like a small shift, but for me it has been the difference between being paralyzed by frustration and being able to take action.
So let's apply this to Landsat Next. When we slow down and actually read the RFI, here's what we're working with:
Budget constraint: $70M/year initially, increasing to $130M/year by 2030, then holding steady
Constellation format: A shift we've actually known about since a 2020 RFI
Quality requirement: Stated commitment to maintaining gold standard data quality and longitudinal continuity
That last point is critical. Landsat is the gold standard for radiometric, spectral, and spatial calibration. Many downstream products rely on it. We can't lose that.
Okay. So now we understand the challenge. We can translate it into accepted constraints and clear goals. The question becomes: what tools and resources do we have to work with?
The 2010s gave us abundance. The 2020s gave us capability.
I came up through an incredible era in geospatial tech. I was first exposed to mapping technology in 2008, the same year that Landsat became free and openly licensed to the general public. As I moved from graduate school to the tech industry, I watched small sats launch into orbit for the first time, with Planet Labs and Skybox Imaging both putting up their first satellites in 2013. Cloud computing became a consumer-accessible resource with AWS (2008), Azure (2010), and GCP (2010) all coming on the scene. We got COGS and Zarr (2015), STAC (2017), Landsat ARD (2016), the DigitalGlobe (now Maxar) commercial archives moving to the cloud (2017).
There was tremendous energy and capital flowing into our industry, and with it there was a huge push towards modernization and cloud-nativity. As many of the companies from this time of abundance matured, investors began to look for stability and returns, venture capital tightened, and we went through a global pandemic, this time of great cashflow waned in the late 2010s and early 2020s.
And as good as the early 2010s were, now is way better. We have the tools, the expertise, and the community to solve hard problems at constrained budgets. Even with a stricter spending ecosystem, we’ve seen a myriad of impressive and needed companies arise in our industry.
We’ve seen the launch of purpose-built satellites like MethaneSat (2024, RIP 2025 unfortunately), FireSat (2025). Muon Space (2021) has come on the scene to do build-to-spec manufacturing, launch and operations of satellites (FireSat is one of their projects). The open source ecosystem has continued to evolve with tools like Xarray (2021) and Icechunk (2024). New platforms for hosting, accessing, and analyzing large amounts of earth observation data have emerged in Whereobots (2022) and Earthmover (2022). We have a nascent landscape of earth foundation models, with the goal of reducing time-to-insights in the world of data science and analytical remote sensing.
Our constraints may have tightened over the last decade, but our industry is still evolving and meeting the moment with meaningful solutions. While times of abundance provide excellent acceleration pads, they are not the norm and we shouldn’t expect them to be.
What resilient thinking looks like
When we evaluate the Landsat Next challenge with our current ecosystem in mind, we're not in as daunting a position as one might think at first pass. We're in a position where we need to:
Accept the budget reality and work within it
When we look critically at the budget numbers, the total spend is not actually changing very much - it’s the mode of spending that is shifting. Historically, Landsat has been funded with massive upfront costs. Based on historic Landsat missions, we can estimate $1B upfront for a mission that lasts 10 - 25 years. The NASA RFI proposes a budget of $70M/year starting in 2026, growing to $130M/year in 2030 and stabilizing there.
$1B over the course of 10 years is $100M/year, and over 25 years it’s $40M/year. So, dividing up the cost of previous Landsat missions over the course of their lifetimes gets us in the range of the proposed Landsat Next annual budget. This is a different mental and financial model than previous Landsat missions, and it will certainly be work to achieve it. But, it’s not an intolerable change to the overall budget for the program.
Protect the non-negotiables of data quality, continuity, and calibration standards
The RFI explicitly states that the commercial production of Landsat Next would have to maintain the gold standard, scientific quality of previous Landsat missions – big sigh of relief. Our industry relies heavily on Landsat being the canon for tightly calibrated longitudinal radiometric continuity. With that sigh of relief, though, we should acknowledge a gap. The commercial earth observation data vendors have never had the market incentives to produce this kind of longitudinally calibrated data product. The scientific expertise, hardware and software infrastructures for achieving this level of data calibration is not well distributed across the commercial industry.
That’s not to say that we couldn’t close this gap – but, it’s worth noting that some evolution would need to occur for the commercial world to fill this need. As Aravind details in his TerraWatch Space post, one of the major existing players would need to fund this effort. Alternatively, I could see a new startup emerging to close this gap, or this becoming part of the public-private deal where the USGS maintains operation of the Cal/Val Center and provides calibration services to a commercial party.
Leverage what exists through our constellation experience, cloud infrastructure, and processing pipelines
I have faith in the commercial sector’s ability to build and launch a well spec’d satellite constellation for this mission. The EarthFire Alliance is already leveraging Muon Space’s manufacturing, launch and operation capabilities to build out a 50 satellite constellation. I don’t worry about the software infrastructure for handling and distributing this data, either.
Fill the gaps collaboratively, something our community has shown we’re capable of W
While distribution of this data is not a technical challenge, we should expect to need community effort around keeping Landsat a freely accessible and openly licensed data resource. The RFI suggests that NASA would become a customer of a commercial producer of this data. What does that mean for commercial consumers of this data product? Will the openly licensed and freely accessible version of this data only be available for scientific and academic use cases? If a private company is the arbiter of licenses, how do we protect the commercial operators that have benefitted massively from this free resource?
We will need to be active participants in this conversation as we move forward. We’ve seen the proliferation of an open ecosystem around OpenStreetMap - let’s work to keep access to Landsat open as well.
The gap between expectations and reality
For me, resilient thinking isn’t about having all of the answers immediately. It’s about acknowledging my “shoulds” and turning them into goals. That allows me to identify the gap between where I’m now and where my goals are, and assess my ecosystem for tools, people, and potential partners in achieving those goals.
Landsat Next is just one arena where we’ll need to abandon our “shoulds”, come together and get creative over the coming years. Much of life is managing the gap between our expectations and our reality. Right now there's a big gap between what we expected Landsat's future to look like and what we're being asked to deliver.
But our industry has built remarkable things under constraints in the past. And we have more tools to work with now than we've ever had before.