Maas Group Firmus Bet Faces Its Biggest Test as Managers Split on the AI Pivot

After selling its quarries for up to $1.7 billion to chase AI infrastructure work, Maas Group now needs Firmus to fund the orders it has won.

13 Oct 2026 · 9 min read

The investment case for Maas Group now rests on a single question: can Firmus Technologies fund the AI factories it has asked Maas to help build? In October 2026, after Firmus was forced to pull its $44 billion IPO when demand from international investors fell away, a string of fund managers quoted in an October article said they were standing aside, even as several acknowledged the stock had started to look cheap.

Hayden Beamish of Laser Beam Capital's global Laser Beam Fund said in that article that Maas Group's outlook was too reliant on the trajectory of Firmus. "The upside for Maas needs Firmus' growth plans to happen," Beamish said. "And the Firmus orders still need paying, which Firmus now has to fund privately."

SG Hiscock & Company portfolio manager Phillip Li framed the dilemma for value-minded investors. He said Maas was starting to look cheap, but added a caveat. "But cheap isn't a thesis on its own," Li said. "The big question is where the next leg of orders comes from, and that depends on Firmus funding its pipeline. Until we get those answers, we're not buyers."

The big question is where the next leg of orders comes from, and that depends on Firmus funding its pipeline.

— SG Hiscock & Company, Article, October 2026

Datt Capital chief investment officer Emanuel Datt said he needed further clarity on how Firmus planned to raise the money needed to develop its business, and pointed to the double-edged nature of the relationship. "Maas holds dual exposure as a Firmus equity holder and as a contractual counterparty. The sustainability of Firmus' continuing business growth is also yet to be proven," he said.

Even existing holders are cautious. Bell Asset Management portfolio manager James Nguyen said in the October article that Maas Group's shares were trading at fair value even if its Firmus stake and future Firmus-linked earnings were completely written off. He was nonetheless not planning to add to his holding soon, citing the risk of more "hate-selling". "We haven't stepped in yet because the sentiment is extremely bearish," Nguyen said, adding that Bell had already been lightening its position ahead of the Firmus IPO to take profits.

Tamim Funds Management had already left. In an October 2026 post, Tamim said its thesis was based on Firmus' continued ability to raise funds and ideally IPO. With that "off the table", Tamim wrote that Maas "might not crystalise that order book" and that its Firmus holding "might be worth less", so it exited to watch from the sidelines.

How a quarry business became an AI infrastructure supplier

The depth of the split makes more sense once it is clear how quickly Maas Group changed shape. When Akambo added to its position, its November 2025 Monthly Report described the company as a vertically integrated civil contractor and construction materials business. Akambo's October 2025 Monthly Report said it was exposed to several favourable trends: infrastructure build, electrical transmission growth and residential real estate. Akambo also noted that the company had given first-time FY26 underlying EBITDA guidance of $240-270m, compared with $219.4m in FY25.

The first sign of the new direction arrived in December 2025, when Maas Group's electrical infrastructure subsidiary JLE Group secured a contract worth approximately $200 million with Firmus Technologies. SG Hiscock & Company's December 2025 Monthly Report said the work related to Firmus' initial 100MW AI data centre cluster in Launceston, with works scheduled through calendar 2026, and that the announcement highlighted a potential staged rollout of a further 500MW+ of AI capacity in Tasmania by 2028. Wilson Asset Management's report that month said the project expanded Maas Group's addressable market into digital infrastructure and that, if the initial contract value was extrapolated across the remaining pipeline, "it implies a substantial runway exists with JLE Group."

QVG Capital, then a holder, offered a more measured read in its December 2025 Monthly Report. QVG acknowledged that Maas had traded on a modest rating for good reason, citing "an unfortunate habit of missing earnings expectations" and returns on invested capital that lagged management's own targets, but argued that "change is afoot", with the planned divestment of commercial property set to simplify the balance sheet and the Firmus win adding capital-lite earnings. "Execution still matters. But if returns lift, so will the discount," QVG wrote.

Execution still matters. But if returns lift, so will the discount.

— QVG Capital, Monthly Report, December 2025

The decisive move came in February 2026. Wilson Asset Management's February Monthly Report said Maas Group announced the divestment of its Construction Materials Division to Heidelberg Materials Australia, with gross proceeds of $1.7 billion to be used to reduce gearing and to be redeployed into growth investment, including electrification, digital infrastructure and industrial services. Alongside the sale, Maas announced a $100 million investment for approximately 1.7% of Firmus Technologies.

Ellerston Capital's February 2026 Monthly Report explained why the market reacted so sharply. It described the construction materials division as "the jewel in the crown" and one of the last remaining ways on the ASX to gain exposure to the construction materials thematic. Ellerston said the implied multiple of 11.8x FY26 EV/EBITDA compared favourably with peer transactions, but that the reaction was driven by "the absence of any explicit framework" for how roughly $550m in net proceeds would be redeployed or returned. Ellerston said it believed in the electrification thematic and considered the strategic direction sound, but kept its investment at a lower weight until management provided greater clarity on capital deployment, naming capital allocation announcements and contract news flow as the key re-rating catalysts.

Others walked away. QVG Capital's February 2026 Monthly Report said the sale came as a surprise and that, while the pivot "may prove to be an inspired move", it left the manager blindsided, so it sold. QVG portfolio manager Tony Waters reflected on the decision in an October 2026 article: "We always knew the Maas investment was largely a case of 'backing the man'. And respect Wes for what he's achieved thus far, but the journey stopped for us at that point." Pengana Capital Group reduced its holding in February 2026, saying the sale left it with "exposure to a business with more opaque future cash flows."

The bull case built on power cubes and a net cash balance sheet

Between the sale and the IPO setback, a group of managers built a detailed bull case. Cerutty Macro Fund's June 2026 Monthly Report described Maas Group as one of the most compelling ways on the ASX to gain exposure to the build-out of Australian data centre and AI infrastructure. In Cerutty's view, management crystallised full value for a mature, capital-intensive asset at the top of its cycle and redeployed into a vertical growing at multiples of GDP, with the balance sheet swinging from roughly $650m of net debt to approximately $650m of net cash after completion.

Central to that view is what Maas actually makes. Cerutty explained that Maas manufactures the Firmus "power cube", a modular electrical system connecting the grid to the data centre, and that it is Firmus' exclusive manufacturing and delivery partner for the electrical infrastructure underpinning its campuses. Cerutty argued the positioning avoids the riskiest parts of the AI trade. "MGH doesn't need to take compute demand risk or fund GPUs; it manufactures the critical electrical hardware every megawatt requires," Cerutty wrote.

MGH doesn't need to take compute demand risk or fund GPUs; it manufactures the critical electrical hardware every megawatt requires

— Cerutty Macro Fund, Monthly Report, June 2026

Cerutty estimated that Maas generates approximately $2m of revenue per MW of data centre build across its electrical scope. It noted that most estimates assume around 250MW a year of power cube delivery, equating to roughly $500m of revenue and about $75m of EBITDA at a ~15% margin, while management believes the business could manufacture 500MW to 1GW a year at limited additional capital, implying about $150m to $300m of annual EBITDA. Cerutty was explicit that it was not underwriting those outcomes and that only the initial 100MW was contracted at the time, presenting the exercise as an illustration of asymmetry.

Regal Funds Management took a similar line. A June 2026 article reported that Regal's Aldridge built a substantial position after the February sell-off and believes earnings could surge by up to 70 per cent. "Our view is this data centre stuff is real. Firmus is real, and there's incredible demand for Wes to win in his electrical contracting business. If Wes gets this pivot right, the market will eventually reward him," Aldridge said.

Execution updates supported the bulls through the middle of the year. Firetrail Investments' May 2026 Monthly Report said Maas reconfirmed FY2026 underlying EBITDA guidance of A$250-280 million and that the Firmus build in Launceston was 35% complete, with commissioning on track for calendar 2026. Firetrail added that the company anticipated further updates on the 3.3 gigawatt Firmus pipeline. Wilson Asset Management's May 2026 Monthly Report said it had added significantly after the sale announcement, describing Maas as a founder-led business with a strong track record of identifying and executing on growth opportunities, and said the data centre exposure "can support earnings growth over the coming years."

By July 2026, Wilson Asset Management argued in an article that the market had "significantly under-appreciated the strategic optionality" created by the sale. It estimated that Maas Group's Firmus stake was worth more than $1 per share, said the operating business traded at less than half of peer multiples, and pointed to management buying back stock aggressively. "We think the shares can double," the manager said.

The order book then took a large step forward. Wilson Asset Management's July 2026 Monthly Report noted that Maas received ACCC approval for the construction materials sale on 31 July and, on 4 August, announced an $855 million contract with Firmus to be delivered over the next 18 months. Bell Asset Management's Tim Johnston said in a September 2026 article that Maas had also spent $300 million to lift its Firmus stake to 3.2 per cent. "They've pivoted to focus almost exclusively on their electrical business, which is a decent-sized player in the data centre build-out," Johnston said.

Cerutty had also pointed to the strength of the counterparty, noting that Firmus raised US$505m in April at a US$5.5bn valuation with NVIDIA participating, announced a partnership with NVIDIA on a 360MW AI factory campus in Batam, Indonesia, and had appointed banks ahead of a potential 2026 IPO that would provide a public mark on Maas Group's stake. Maple-Brown Abbott added Maas to its watchlist in a May 2026 article, citing Australia's structurally strong data centre build-out and the potential for the stock to play catch-up with data centre-exposed contractor peers, while noting governance concerns.

What managers are watching next

The thesis has narrowed to one dependency. Cerutty Macro Fund had listed the catalysts as Firmus funding milestones beyond the initial 100MW, contract wins beyond Firmus, completion of the construction materials sale and deployment of surplus capital. With the IPO withdrawn, the first of those has become the swing factor for almost every manager quoted.

The October commentary shows managers converging on the same checklist from opposite directions. Bell Asset Management sees valuation support even with Firmus written off, while Laser Beam Capital, SG Hiscock & Company and Datt Capital want evidence that Firmus can fund its pipeline privately before committing capital. Ellerston Capital's earlier warning still frames the debate: the market is unlikely to reward optionality it cannot yet underwrite. What managers are waiting for is proof that the $855 million order book is paid for, and clarity on where the next leg of orders comes from.