High Stakes, Low Oversight: IM-2’s Tip-Over Tests the Limits of NASA’s CLPS Gamble

In the early hours of this week, Intuitive Machines’ second attempt to reach the lunar surface ended with a bittersweet result. The company’s Nova-C lander Athena successfully descended to the Moon near its south pole – a remarkable achievement – but soon after touchdown it toppled onto its side. For the Houston-based startup, it was déjà vu; their first lander did almost the same thing a year ago. The incident underscores just how unforgiving lunar landings remain, and it is raising tough questions about sensor glitches, engineering practices, and NASA’s hands-off approach to these commercial missions

Intuitive Machines’ Athena lander ultimately came to rest on its side in the dim interior of a crater, approximately 250 meters from its intended landing spot at Mons Mouton near the lunar south pole. With the spacecraft tilted, its solar panels could not properly face the sun, and frigid darkness quickly drained its batteries. Within about 12 hours of touchdown, the lander fell silent, effectively ending the mission as its systems lost power. Mission controllers managed to collect some data and even a few images – including the one above showing the half-lit Earth hanging in the sky above the stricken lander – but the ambitious science objectives (like drilling for water ice) were largely left unfulfilled.

The Altimeter That Couldn’t Find the Ground

Why did Athena tip over after a descent that otherwise seemed smooth? The culprit appears to be a glitch in the lander’s laser altimeter system – essentially its “eyes” for measuring altitude. Intuitive Machines executives revealed that during the final approach, the lander’s altimeters were plagued by “signal noise and distortion” that prevented accurate altitude readings. In other words, the spacecraft’s ranging sensor was seeing gibberish. Without a reliable gauge of its height and velocity above the lunar surface, Athena could not properly adjust its thrust and orientation for a gentle, upright landing.

Engineering details are still emerging, but the description points to some form of interference or sensor malfunction onboard. In small spacecraft, integrating multiple instruments in close quarters can lead to electromagnetic compatibility (EMC) issues – one system’s signals inadvertently jamming another. It’s possible the twin laser altimeters on Athena interfered with each other’s measurements or picked up spurious reflections. Regardless of the exact root cause, the outcome was that Athena likely misjudged its altitude by crucial meters. The result: it either cut off its engine a moment too soon or struck the surface with lateral motion, causing a “landing anomaly” (as the company delicately put it) that broke a leg and toppled the probe.

Intuitive Machines identified three key factors that contributed to the botched landing:

• Altimeter Signal Interference: Noise in the lander’s laser altimeters obscured accurate altitude data, leaving the guidance system partially “blind” during the final descent.
• Polar Lighting Conditions: The low-angle sunlight at the south pole created long, dark shadows that challenged the lander’s optical landing sensors. Key surface features were hidden in darkness, undermining precision landing algorithms.
• Terrain Masking: The lander’s navigation system relied on Lunar Reconnaissance Orbiter maps to recognize craters, but at low altitude the craters appeared different in the oblique light. Athena was descending into a 20-meter crater it hadn’t clearly identified, on terrain that wasn’t as flat as expected.

The combination of a sensor glitch and a tricky landscape proved lethal to the mission. By the time Athena realized the true nature of the surface below, it was too late to correct fully. To the lander’s credit, it did reach the Moon’s surface – in fact, it came down closer to the lunar south pole than any spacecraft before it, a location of great interest due to potential water ice deposits. But a landing that ends with hardware on its side is undeniably far from ideal.

Echoes of the First Mission

If all this sounds familiar, it’s because Intuitive Machines’ first lunar mission (IM-1) suffered a strikingly similar fate. In February 2024, the company’s inaugural lander (named Odysseus) also tipped over after touchdown. An investigation later revealed that on that mission the primary laser altimeter wasn’t providing data at all during descent – reportedly due to a human configuration error pre-launch – so the lander was essentially guessing its altitude. Odysseus hit the Moon while still moving sideways, snapped one of its landing legs, and came to rest askew on the surface. Sound familiar? The second attempt managed to avoid the exact same mistake (this time the altimeter was on, at least), but encountered a new flavor of failure with the noisy sensor data. Two attempts, two altimeter-related failures.

For Intuitive Machines, this repetition is humbling. Critical sensors like altimeters are the linchpin of any landing sequence – they feed the autopilot the information needed to throttle engines and time the final shutdown. Having the same type of instrument falter twice in a row hints at deeper issues in the lander’s design or testing regime. Perhaps the company underestimated the complexity of integrating a sensitive laser ranging system into a compact, electronics-packed vehicle. It’s one thing to test an altimeter in a lab or on a drone; it’s another to ensure it works reliably amid the vibrations, thruster plumes, and electromagnetic noise of a lunar landing. Larger, experienced space agencies rigorously test for EMC and sensor integration problems, running full-up simulations of landing sequences. A young startup on a tight timeline might not have caught every potential interference ahead of time.

To its credit, Intuitive Machines says it is learning from these glitches. Steve Altemus, the CEO, has outlined upgrades for the next mission (IM-3): the lander will carry redundant, dissimilar altimeters (to provide backup and avoid one system’s noise affecting both) and a new sensor that can measure velocity independent of lighting conditions​. The company is also expanding its terrain mapping database and incorporating imagery from the failed mission into its algorithms. These are smart fixes – essentially adding more eyes and smarter brains to the lander – but it’s notable that such measures weren’t in place from the start. In hindsight, having only a single type of altimeter with no true backup was a risky choice for IM-1 and IM-2. That’s the kind of design decision that robust aerospace engineering standards would normally flag. In the Wild West of commercial space startups, however, those standards are still being developed through trial and (often expensive) error.

Lessons from Peregrine’s Fate and Other Mishaps

Intuitive Machines is not alone in struggling with the first steps of private lunar exploration. Pittsburgh-based Astrobotic Technology saw its own lunar lander dreams falter last year. Peregrine, the first lander in NASA’s Commercial Lunar Payload Services program (and a direct competitor to Intuitive’s missions), never even got the chance to attempt a landing. It launched in January 2024 but suffered a mission-ending failure en route to the Moon. The culprit, revealed after months of analysis, was surprisingly mundane: a faulty pressure valve in the spacecraft’s propulsion system. Vibration during launch caused an internal component of a helium pressurization valve to come loose, which led to a propellant leak and left the vehicle unable to complete the lunar transfer. Peregrine ultimately tumbled back into Earth’s atmosphere, destroying a mission carrying 11 NASA experiments and several commercial payloads.

These early stumbles by Astrobotic and Intuitive Machines highlight different failure modes – one mechanical, one sensory – but a common theme emerges: small private landers are encountering the kinds of problems that decades of spaceflight experience have taught us can happen, especially on maiden voyages. A single misconfigured sensor or a single stuck valve is enough to doom an entire mission. In the Apollo days or even the big robotic missions like Viking or Curiosity, NASA would spend years double-checking such things (and building in redundancy so one failure wouldn’t be fatal). The new commercial landers don’t have that luxury of time or perhaps the depth of heritage in their hardware. They are learning on the fly. And unfortunately, learning on the fly sometimes means crashing (or at least landing on your side).

One positive sign: not every newcomer has face-planted. In March, just days before Intuitive’s Athena went sideways, Firefly Aerospace succeeded in landing its Blue Ghost lander in an upright, functioning state, achieving the first fully successful Moon landing by a private company. That mission, part of the same NASA initiative, suggests that with the right preparations and a bit of luck, these commercial landers can stick the landing. Firefly’s success will no doubt be dissected by others in the field for clues – what did they do differently? How did their sensors and systems avoid the pitfalls that caught Peregrine and Athena? In the end, each failure and success is adding to a collective knowledge base that all the players (and NASA itself) will draw upon.

NASA’s High-Risk, Hands-Off Approach

All these lunar lander ups and downs are unfolding under the umbrella of NASA’s Commercial Lunar Payload Services (CLPS) program. CLPS is an experimental approach by NASA to fast-track lunar science and exploration by outsourcing the delivery of instruments to the Moon’s surface. Instead of NASA designing, building, and landing spacecraft (as it traditionally would), the agency contracts companies like Intuitive Machines, Astrobotic, Firefly, and others to handle the whole mission as a service. Unlike traditional NASA missions, these CLPS landers are fully company-built and operated with minimal NASA oversight. NASA basically says “Here’s what we want delivered and roughly where; the rest is up to you.” The space agency only specifies the payload and general landing region, and the companies are free to innovate (or, occasionally, invent new ways to fail).

This model comes with a “high risk, high reward” philosophy. NASA leaders have been very transparent that they expect some of these missions to fail. In fact, NASA deliberately chose to put mostly low-cost, relatively non-critical payloads on the first batch of CLPS landers, knowing that these early missions were stepping into unknown territory​. The rationale is that even if a lander crashes or tips over, the loss is not catastrophic to NASA’s overall science program – and the lessons learned will inform later missions. As if to underscore this point, after Intuitive Machines’ first lander ended up crooked in the lunar dust, NASA and the company still declared the mission an “unqualified success” in terms of what it taught them. And after the second lander’s rough landing, NASA officials again stressed that every CLPS mission is a success “regardless of the outcome because lessons are learned”​. It’s a remarkable reframing of failure: in this program, failures are not embarrassing setbacks but rather expected data points on the path to sustainable lunar access.

However, this adventurous approach also raises concerns. Two private landers in a row have not achieved a stable, long-duration lunar surface mission. Do these repeat mishaps signal gaps in NASA’s oversight or in the guidance provided to the companies? By design, NASA gives the CLPS contractors plenty of freedom – perhaps too much, some critics suggest. There are no detailed NASA human-rating style requirements or exhaustive checklists imposed on these lander designs. For instance, NASA didn’t dictate “Thou shalt have two independent altimeters” – that decision was left to Intuitive Machines (which initially went with a single-string altimeter system on IM-1/2). Nor did NASA vet every component of Peregrine’s propulsion system with the rigor it might apply to its own spacecraft. The lack of standardized guidance for these commercial landers means each company is drawing on its own expertise (or lack thereof) to meet broad performance goals. While innovation thrives in such an environment, so do the chances of oversight. A simple valve design error or sensor integration flaw might have been caught under a more stringent regime.

It’s worth noting that NASA isn’t completely hands-off – agency engineers do interface with the companies, and there are milestone reviews. But fundamentally, CLPS is about shifting responsibility (and risk) to the private sector. As a NASA Inspector General report put it, the contractors assume “increased responsibility, risk, and accountability with little Agency oversight” in exchange for a fixed payment. The upside for NASA is obvious: if a mission succeeds, NASA gets its experiments delivered for a fraction of what a in-house mission might cost. If it fails, NASA can still glean insight without having spent a fortune, and there’s usually another CLPS flight in the pipeline. The downside, as we’re seeing, is a string of very public failures (or partial failures) that can start to wear on the patience of scientists waiting for data – and on investors backing these firms.

Indeed, the commercial nature of these missions means market pressures are part of the equation. Intuitive Machines is a publicly traded company, and the stock market doesn’t grade on a curve for “learning experiences.” After the IM-2 news, Intuitive’s stock reportedly plunged by about 20–25% in a single day. Each high-profile failure can make it harder for these companies to raise funds or attract customers, putting the onus on them to fix problems fast. NASA’s forgiving stance on outcomes won’t necessarily comfort a shareholder or a private customer with a payload manifest. This dynamic – private capital at risk in an endeavor NASA historically shielded from market forces – is new in lunar exploration. It adds urgency for companies to demonstrate reliability sooner rather than later.

Navigating the Path Ahead

So where do we go from here? In the near term, Intuitive Machines is pressing ahead with its next mission, and so are other CLPS providers. The hope is that the hard-won lessons from these tumbles translate into steadier landings in the coming years. Intuitive Machines’ planned upgrades for IM-3 (better sensors, more testing, and not aiming for such a shadowy crater next time) give reason for cautious optimism. Astrobotic, for its part, has redesigned the faulty valve and is preparing a second try with Peregrine as well. NASA, too, is learning alongside its contractors – each failure highlights something NASA might choose to guide on more firmly in the future. For instance, NASA could decide to issue recommended practices for critical systems (like redundant landing sensors or more thorough pre-flight simulations) as part of the CLPS program’s evolving requirements. Without stifling innovation, there may be middle ground where NASA’s vast engineering experience can be fed into the design process of these private missions a bit more.

In a broader sense, the repeated moon mishaps are a reality check on the risks of rapid commercialization. The Moon may be Earth’s closest celestial neighbor, but it is no low-hanging fruit. Even with today’s advanced technology, a soft landing requires a symphony of systems performing in concert – guidance cameras, altimeters, thrusters, navigation algorithms, and more – all with zero margin for error. Any discord in that symphony (a misread sensor, a stuck valve, an overlooked software scenario) and the lander finds itself in a very expensive heap of scrap metal in the lunar dust. The CLPS program embraces this risk, wagering that the faster feedback loop will ultimately yield cheaper, better landers after a few crashes. It’s essentially “fail fast, fail forward” applied to lunar exploration.

If that sounds a bit brash, it is. But NASA’s alternative was to spend much more money and time to attempt near-flawless landings itself – an approach that, in today’s budget and schedule climate, might have meant few landings at all. By accepting some failures, NASA is buying many shots at the goal. The key question is how many failures are too many before the strategy should be adjusted. After two tipped-over landers, critics might argue it’s time for NASA to tighten the reins slightly, or at least ensure that obvious lessons (like “don’t rely on a single altimeter”) are disseminated across all teams. Proponents will counter that we’re still in the early innings – that this is what early aviation looked like too, with many crashes before the kinks were worked out.

Ultimately, the proof will be in the results over the next couple of years. Missions like IM-3, Astrobotic’s retry, and upcoming landers by Blue Origin and Draper will show whether the companies truly are learning and improving. If we start seeing routine successes – landers touching down gently, doing their two weeks of science, and beaming back priceless data – then the CLPS experiment will be vindicated. If failures continue at a high rate, pressure may build for NASA to provide more guidance or oversight to avoid spinning wheels.

For now, Intuitive Machines’ latest lunar saga offers a sobering but enlightening story. A private spacecraft reached the Moon’s forbidding south pole, proved out a lot of new technology, but fell just short of a fully successful landing due to a pesky sensor issue that might have been prevented with more rigorous integration testing. The electromagnetic interference that foiled the altimeter is a small gremlin with big consequences – and it’s now a known enemy to be designed against. The fact that two missions in a row were tripped up by the same subsystem shines a light on where these companies must focus their engineering rigor.

In the grand scheme, each “sideways landing” is a stepping stone. As frustrating as they are, they force improvements that will benefit not just one company but the whole cohort of lunar hopefuls. NASA’s commercial moon program is essentially learning to walk (or perhaps to land) via trial and error. It may not be pretty to watch at times, but it’s making progress inch by inch. And in a very real sense, every time a lander either succeeds or fails, we gain knowledge that brings the Moon a little closer within reach.

In the end, the Moon is teaching us the same lesson it taught the Apollo engineers half a century ago: space is hard, the details matter, and gravity never takes a day off. The new twist is that this time it’s not just NASA learning the lesson, but a mix of scrappy companies and government partners together. Intuitive Machines’ tipped-over lander is a visible reminder of those challenges. The hope is that, armed with these lessons about altimeters, interference, and integration, the next time a private lander descends to the lunar surface, we’ll see it standing tall – and not lying on its side – when the dust clears.