$450M and a Penny Warrant

The next time you find yourself on EDGAR, consider using my SECSift reader instead!

It turns SEC filings into clean, readable documents. You can search every word, export anything, merge tables, highlight what matters and gray out the rest, and compare against previous filings.

Not familiar with the reader? Here’s a 3-minute interactive demo (viewable on your computer) and another video I made walking through the entire tool.

🌾 Welcome to StableBread’s Newsletter!

I wrote up The Elmet Group (Nasdaq: ELMT) on July 2, 2026, at $19.25/share, as my favorite of the seven tungsten names in that write-up.

The thesis was that Elmet, not the miners, was the safest way to own China’s tungsten squeeze. Elmet is the U.S.-owned maker of the finished tungsten and molybdenum parts that go into missiles, radar, and chips, and it has sourced 95%+ of both metals outside China for over a decade.

I added to ELMT just under $14/share in early August (which I told paid subs) when the stock broke below its $14.00 IPO price. On August 16, after a record $131.5M Q2 backlog (following earnings), I wrote that I wanted the first backlog print with the new defense appropriation cycle before adding more.

Then, on September 14, the U.S. Department of War (DoW) committed $450M to Elmet in redeemable preferred stock plus warrants that, once exercised, give the DoW 19.9% of Elmet’s shares, with $200M funded at closing.

The same morning Elmet disclosed a Defense Logistics Agency (DLA) stockpile contract with a $2B ceiling and $150M guaranteed, and a $150-175M plan to restart the Springer tungsten mine and APT plant in Nevada (APT is ammonium paratungstate, the traded form of tungsten).

None of this surprises me. Tungsten hasn’t been mined commercially in the U.S. since 2015, so a domestic supply chain has to start with a restarted mine and a processor to buy its concentrate.

ELMT jumped 32.8% on the news, from $16.19 the Friday before to $21.50 on Monday, on 13.9M shares traded (~46% of the shares outstanding). It’s now hovering ~$23/share, with nothing new on the deal from the company since Monday’s filings.

Currently, the stock sits ~19% above my July write-up price and ~64% above my August add.

At $23/share, Elmet is a $701M company (30.46M shares × $23) before the government’s warrants and an $845M one after them (36.7M diluted shares × $23).

Elmet has the $200M in cash today. The offset is a $200M preferred (senior to the common) whose balance grows 5.5% a year, paid in kind rather than in cash.

So is this the July thesis validated with the government’s capital, or did Elmet hand over a fifth of the company and a list of consent rights to fund a five-year buildout?

Below, I break down the preferred and warrant terms in detail, what the Pentagon controls now, where the $450M goes, and what $23/share already assumes. Then I conclude with whether I’m adding more.

Elmet sold the DoW $200M of Class A Redeemable Preferred Stock on September 14, 200,000 shares at $1,000 each, and issued two warrants alongside them at no extra cost to the government.

The preferred pays 5.5% a year, cumulative, compounding quarterly, and entirely in kind, so nothing goes out in cash and the balance grows to ~$263M by September 2031 ($200M × 1.01375²⁰).

Elmet can redeem at any time at the accreted value with no premium. The government can force redemption after 10 years, on a change of control, or if Elmet defaults on the terms. It ranks senior to the common and votes with the common, capped at 19.9%.

The warrants cover 7,567,341 shares, 24.8% of the 30.46M shares outstanding today and 19.9% of the count after exercise (7.57M / 38.03M). They come as two separate warrants:

Both are exercisable from Sept. 14, 2027 through Sept. 14, 2036, with no price-based anti-dilution. Elmet has to file a resale shelf for the shares by November 13, so the overhang starts in September 2027, not today.

At $15.92, the penny warrant was worth $90.3M the day the deal was signed (5,675,506 × $15.92). So the first $200M cost Elmet $90.3M of stock up front, 45% of the money raised ($90.3M / $200M), plus 5.5% a year in kind on the full $200M.

The deck’s one-slide summary of the terms is below:

I’d say Elmet got a good deal here. Raising the same $200M in common at $15.92 would have meant 12.6M new shares ($200M / $15.92), 41% of today’s count, against 7.57M here (just 24.8%) plus a preferred that isn’t due until 2036. The other $250M comes with no warrants at all.

The preferred terms are lenient. No cash goes out for the coupon, nothing comes due before September 2036 unless Elmet sells itself or defaults, and Elmet can pay it off early at the accreted balance with no premium.

For a company with $31.8M of TTM adjusted EBITDA, $450M is 14x that ($450M / $31.8M), more than Elmet could’ve borrowed and more than the market would’ve given it in August, when the stock was below its $14 IPO price.

The remaining $250M is a commitment, not cash. Elmet can call up to five more tranches of $25-50M between March 2027 and March 2029, subject to progress at Springer, a fourth project the filing redacts, and Congress still appropriating the money.

No additional warrants attach to those tranches, and Elmet decides whether to draw the money.

One clause in the Certificate of Designations (the document that sets the preferred stock’s terms) links the preferred and the penny warrant, and it’s worth understanding.

When the DoW exercises the penny warrant, part of the warrant’s value is subtracted from what Elmet owes on the preferred. But only the value above $31.84/share counts, double the $15.92 the stock closed at when the deal was signed, and the price used is the 30-day VWAP (volume-weighted average price) at exercise.

It works like tax brackets. Nothing comes off for the value up to $31.84. Of the value between $31.84 and $39.80 (2x to 2.5x), 45% comes off. Each further half-turn of appreciation raises the rate 2.5 points, up to 65% of everything above $95.52 (6x).

The filing’s own schedule is below:

On the 5.68M penny warrant shares, that works out to $20.3M off at $39.80 ($7.96 × 45% × 5,675,506), $88.1M at $63.68, and $194.3M at $95.52.

The original $200M only comes off in full at $97/share (6.1x), and by September 2027, the first day the warrant can be exercised, the preferred will have grown to ~$211M ($200M × 1.01375⁴), which takes $100/share.

Below $31.84 there’s no reduction at all. Elmet still owes the full accreted balance and the warrant still dilutes. So the clause only helps common holders in the scenarios where the stock has already gone up 2-6x, and it only wipes out the preferred if ELMT goes up 6x.

Anywhere below $31.84, common holders pay for both the dilution and the full preferred.

Note: The filing’s own example in Appendix I runs the same 45-65% schedule on a $17.17 base price and fewer shares than the final 5,675,506, so its dollar totals don’t match mine. The schedule is identical; only the inputs differ.

The Investor Rights Agreement gives the DoW a say in decisions boards usually make alone, for as long as it holds any preferred or at least 25% of its warrant shares:

The September 14 8-K also discloses a cash bonus for Scott Knoll, Elmet’s EVP of Corporate Strategy and a director. He gets $1M if Elmet’s average market cap over any 10 trading days reaches $1.0B before April 23, 2027, $2M at $1.25B, and $3M at $1.5B.

On today’s count, those are $32.83, $41.04, and $49.25 per share, and the first tier sits almost exactly at the $31.84 warrant trigger.

Elmet’s September 14 investor presentation splits the $450M into these buckets:

Slide 9 lays out the same split with what each bucket is meant to buy:

In July I wrote that the one step Elmet didn’t own was turning concentrate into the oxide its powder line starts from, and that management called closing that gap an M&A target.

Springer is that step, and the government is paying for the plant.

Blue Moon Metals (Nasdaq: BMM) owns the Springer complex in Imlay, Nevada, a mine and APT plant last run by General Electric, with a 1,200-tonne-per-day mill and a resource that’s still a 1984 historical estimate of 10.7M tonnes at 0.45% WO₃ (tungsten trioxide) that Blue Moon says shouldn’t be relied on.

The letter agreement with Blue Moon and EQ Resources (ASX: EQR), the miner in which Elmet already holds 23.7M shares, has four parts (per the 8-K):

Blue Moon’s April estimate had Springer producing 107,000-124,000 MTU a year (an MTU is 10 kg of contained WO₃) at a US$300-400/MTU cash cost, or 1,070-1,240 tonnes of contained WO₃.

A 4,000-tonne APT plant needs ~3,540 tonnes of contained WO₃ (4,000 × 88.5%), so Springer fills 30-35% of the feed (1,070-1,240 / 3,540), EQ Resources’ eight-year offtake covers 500 tonnes a year (4,000 tonnes in total), and Elmet Refining & Trading has to buy the remaining 1,800-1,970 tonnes a year, or 51-56% of the feed.

The plan runs on the $1,500/MTU long-term APT price Elmet assumes in its September 14 investor presentation, half of the $2,900-3,100/MTU quoted in Rotterdam on September 10.

Mine and mill production is targeted for Q4 2027 and APT for H2 2028, per the Springer press release, though the deck’s Springer slide says H1 2029 for the APT restart. Definitive agreements are due within 12 months.

Springer’s concentrate already has a customer lined up, and it’s the government itself.

On September 1, the DLA awarded Elmet a contract to supply tungsten ores, concentrates, and sodium tungstate to the National Defense Stockpile, with a $150M guaranteed minimum, a $2B ceiling at the DLA’s discretion, and a base period through August 30, 2031 (with an option to 2033). Pricing is set order-by-order.

The DoW’s release names Springer, Mt. Carbine, and Barruecopardo as the approved sources.

Spread evenly over the five-year base period, the guaranteed $150M is $30M a year ($150M / 5), close to what Elmet’s entire tungsten product line sold in FY2025 ($32M).

Elmet doesn’t mine, so Elmet Refining & Trading will be buying concentrate from Blue Moon and EQ Resources and reselling it to the stockpile.

Nothing obliges the DLA to order beyond the $150M, so I’d treat the other $1.85B of the ceiling as an upper bound, not as orders Elmet has in hand.

Elmet’s other deal this month, signed September 3, is the acquisition of OSRAM’s tungsten and molybdenum plant in Schwabmünchen, Germany, ~290,000 square feet that has made powders, rods, wire, and electrodes for lighting since 1961.

Elmet isn’t paying for the plant. OSRAM is paying Elmet €18M to take it off its hands, before adjustments for pensions and working capital, because the asset purchase agreement says the business has a “negative fair market value.”

OSRAM’s own orders are expected to dry up by the end of 2027, which is why the headcount falls from 157 to 120 by then at OSRAM’s cost, with a two-year contract manufacturing agreement covering what OSRAM still buys after closing.

Elmet leases the site for 10 years for next to nothing (a “nominal amount,” in the deck’s words), and closing is Q1 2027 after a German foreign investment review, with no DoW money involved.

What Elmet gets is a European plant for EU defense, semiconductor, and medical customers. The catch is that the plant loses its main customer, OSRAM itself, within a year of closing, so Elmet has to replace it with its own orders.

At $23/share, Elmet’s market cap is $701M (30.46M shares × $23). Count the penny warrant as shares and the strike warrant on a treasury basis, and the diluted count is 36.7M (30.46M + 5.68M + 0.58M) for an $845M diluted market cap.

Net cash after the deal is $60.5M ($71.0M at July 3, plus $200M, less $10.5M of debt and the $200M preferred), so enterprise value is $784M ($845M − $60.5M).

That’s 24.7x TTM adjusted EBITDA ($784M / $31.8M) and 3.4x TTM revenue ($784M / $228.5M), against 16.1x and 2.2x at $18.82 on August 14.

So the stock is up 22.2% since my August 16 update and the multiple is up 53.4%, because the warrants added 6M shares and the preferred is a $200M claim ahead of the common.

But does $23/share already pay for the whole buildout, or only a slice?